Professional Documents
Culture Documents
16 April 2010
Space engineering
Threaded fasteners handbook
ECSS Secretariat
ESA-ESTEC
Requirements & Standards Division
Noordwijk, The Netherlands
ECSSEHB3223A
16April2010
Foreword
This Handbook is one document of the series of ECSS Documents intended to be used as supporting
material for ECSS Standards in space projects and applications. ECSS is a cooperative effort of the
EuropeanSpaceAgency,nationalspaceagenciesandEuropeanindustryassociationsforthepurposeof
developingandmaintainingcommonstandards.
This handbook has been prepared by the ECSSEHB3223 Working Group, reviewed by the ECSS
ExecutiveSecretariatandapprovedbytheECSSTechnicalAuthority.
Disclaimer
ECSSdoesnotprovideanywarrantywhatsoever,whetherexpressed,implied,orstatutory,including,but
notlimitedto,anywarrantyofmerchantabilityorfitnessforaparticularpurposeoranywarrantythatthe
contentsoftheitemareerrorfree.InnorespectshallECSSincuranyliabilityforanydamages,including,
butnotlimitedto,direct,indirect,special,orconsequentialdamagesarisingoutof,resultingfrom,orin
any way connected to the use of this document, whether or not based upon warranty, business
agreement,tort,orotherwise;whetherornotinjurywassustainedbypersonsorpropertyorotherwise;
andwhetherornotlosswassustainedfrom,oraroseoutof,theresultsof,theitem,oranyservicesthat
maybeprovidedbyECSS.
Publishedby: ESARequirementsandStandardsDivision
ESTEC,P.O.Box299,
2200AGNoordwijk
TheNetherlands
Copyright: 2010bytheEuropeanSpaceAgencyforthemembersofECSS
2
ECSSEHB3223A
16April2010
Table of contents
3
ECSSEHB3223A
16April2010
5.2.3
25H Eccentric Axially Loaded Joints ................................................................. 32 427H
5.2.4
26H Shear Loaded Joints.................................................................................. 32 428H
5.2.5
27H Combined Loaded Joints........................................................................... 32 429H
5.2.6
28H Low Duty Joints ......................................................................................... 33 430H
5.3
29H Joint Analysis Procedure ......................................................................................... 37 431H
5.3.1
30H Overview.................................................................................................... 37
432H
5.3.2
31H Margins of Safety....................................................................................... 37 433H
5.4
32H Joint Geometry ........................................................................................................ 39
434H
5.4.1
33H Fastener Geometry.................................................................................... 39 435H
5.4.2
34H Thread Geometry ...................................................................................... 40 436H
5.5
35H Safety Factors ......................................................................................................... 43
437H
5.5.1
36H Overview.................................................................................................... 43
438H
5.5.2
37H The Joint Fitting Factor.............................................................................. 43 439H
5.6
38H References .............................................................................................................. 44
440H
6.1
40H Overview ................................................................................................................. 45
442H
6.2
41H Coefficient of Utilisation........................................................................................... 45 443H
6.3
42H Theoretical Aspects................................................................................................. 46 444H
6.3.1
43H The Relation between Torque and Preload............................................... 46 445H
6.3.2
44H Determining the Design Torque Level ....................................................... 49 446H
6.3.3
45H Locking Devices and Prevailing Torque .................................................... 52 447H
6.3.4
46H Sources of Error ........................................................................................ 53 448H
6.3.5
47H Forces Induced by Thermal Fluctuation .................................................... 55 449H
6.4
48H Mechanisms of Preload Loss .................................................................................. 56 450H
6.4.1
49H Overview.................................................................................................... 56
451H
6.4.2
50H Embedding ................................................................................................ 57452H
6.4.3
51H Fastener Group Interaction........................................................................ 62 453H
6.4.4
52H Gaskets ..................................................................................................... 64
454H
6.4.5
53H Metallic Creep............................................................................................ 64455H
6.4.6
54H Washers .................................................................................................... 64
456H
6.4.7
55H Torsional Relaxation.................................................................................. 65 457H
6.5
56H Margin of Safety on Tightening ............................................................................... 66 458H
6.6
57H Worked Example ..................................................................................................... 67 459H
6.7
58H References .............................................................................................................. 67
460H
4
ECSSEHB3223A
16April2010
7 Concentric Axially Loaded Joints...................................................................... 68
59H 461H
7.1
60H Overview ................................................................................................................. 68 462H
7.2
61H Joint Stiffness .......................................................................................................... 68 463H
7.3
62H The Joint Diagram ................................................................................................... 69 464H
7.3.1
63H Overview.................................................................................................... 69 465H
7.3.2
64H Compressive Loading................................................................................ 72 466H
7.4
65H The Force Ratio....................................................................................................... 73 467H
7.5
66H The Compliance of the Fastener ............................................................................. 73 468H
7.6
67H The Compliance of the Clamped Parts ................................................................... 75 469H
7.6.1
68H Overview.................................................................................................... 75 470H
7.6.2
69H Compression Zone Configurations ............................................................ 76 471H
7.6.3
70H Determining the Compression Zone Configuration ................................... 79 472H
7.6.4
71H The Compression Zone Compliance ......................................................... 80 473H
7.6.5
72H Gasket Compliance ................................................................................... 81 474H
7.7
73H Calculation of the Loading Plane Factor ................................................................. 83 475H
7.7.1
74H Introduction................................................................................................ 83 476H
7.7.2
75H Simple Calculation of the Loading Plane Factor........................................ 87 477H
7.7.3
76H Analytical Calculation of the Loading Plane Factor ................................... 89 478H
7.8
77H Joint Separation ...................................................................................................... 91 479H
7.8.1
78H Introduction................................................................................................ 91 480H
7.8.2
79H The Margin of Safety for Joint Separation ................................................. 92 481H
7.9
80H Fastener Tensile Failure.......................................................................................... 92 482H
7.9.1
81H External Vs Overall Load........................................................................... 92 483H
7.9.2
82H Margin of Safety on Fastener Failure ........................................................ 93 484H
7.10.1 Introduction................................................................................................ 94
84H 486H
5
ECSSEHB3223A
16April2010
94H7.15 References ............................................................................................................ 108 496H
109H 8.3.3 Force Ratio for Circular Cantilever Flanges ............................................ 124 511H
110H 8.3.4 Simplified Joint Diagram for Eccentric Joints .......................................... 125 512H
111H 8.3.5 Joint Diagram for Eccentric NonLinear Joints ......................................... 126 513H
6
ECSSEHB3223A
16April2010
9.3.5
128H Shear-Out or Tear-Out Failure ................................................................ 150 530H
9.4.1
130H Overview.................................................................................................. 151
532H
9.4.2
131H Fastener Group Centroid......................................................................... 152 533H
9.4.3
132H Bearing Design of Eccentric Shear Joints ............................................... 152 534H
9.4.4
133H Friction Grip Design of Eccentric Shear Joints ........................................ 153 535H
9.5.1
135H Shear Loaded Joint Example .................................................................. 155 537H
9.5.2
136H Net Tension Section Failure Example ..................................................... 158 538H
9.5.3
137H Eccentric Shear Bearing Joint Example .................................................. 159 539H
7
ECSSEHB3223A
16April2010
11.2.3 Fatigue Design Principles........................................................................ 175
162H 564H
13 Corrosion......................................................................................................... 201
193H 595H
8
ECSSEHB3223A
16April2010
13.2.1 Introduction.............................................................................................. 203
196H 598H
9
ECSSEHB3223A
16April2010
16.2 Types of Testing.................................................................................................... 220
229H 631H
Figures
Figure 5-1 - Joint Categories (1 to 6).................................................................................... 34
247H 649H
Figure 6-1 Forces Present During Tightening (the Wedge Model) .................................... 47
254H 656H
10
ECSSEHB3223A
16April2010
Figure 6-7 - Effect of Fastener Stiffness on Preload Loss Due to Embedding ..................... 61
260H 662H
Figure 6-9 - Initial and Final Preload of a Typical Flange Joint for the First Tightening
262H
Pass.................................................................................................................... 63664H
Figure 6-10 - Initial and Final Preload Levels after Four Tightening Passes. ....................... 63
263H 665H
Figure 7-4 - Growth of the Joint Diagram Illustrating the Tightening Process ...................... 70
268H 670H
Figure 7-6 - Joint Diagram Showing the Effect of an External Axial Load ............................ 71
270H 672H
Figure 7-11 - The compression zone when multiple edge distances are present ................ 77
275H 677H
Figure 7-12- Approximation of the Compression Zone for Insert Joints ............................... 78
276H 678H
Figure 7-16- Joint Diagram for a Gasket Joint Showing Gasket Creep................................ 82
280H 682H
Figure 7-17 - Tension Joint Loading Planes and the Forces Acting within the Joint ............ 83
281H 683H
Figure 7-18 - Joint Diagram for Loading Planes within the Joint (n < 1)............................... 84
282H 684H
Figure 7-22 Joint Types According to the Points of Force Introduction ............................. 88
286H 688H
Figure 7-24- Moment Conduction Factor Variation with the Connector Body Position......... 90
288H 690H
Figure 7-26 Joint Diagram Showing Repeated Loading into the Plastic Region ............... 98
290H 692H
Figure 7-29 Typical S-N Curve (For a Given Mean Stress)............................................. 100
293H 695H
11
ECSSEHB3223A
16April2010
Figure 7-30 - Example problem of concentric axially loaded joint ...................................... 100
294H 696H
Figure 8-4 - Redistribution of interface pressure (with force FA1 not sufficient to cause
298H
gapping)............................................................................................................ 111
700H
Figure 8-6 - Redistribution of Interface Pressure (Force FA2 Causes Gapping) ................. 112
300H 702H
Figure 8-7 - Limitations of the Eccentric Joint Analysis Method ......................................... 113
301H 703H
Figure 8-8 - The Interface Area for Multiple Fastener Joints .............................................. 114
302H 704H
Figure 8-11 - The Joint Diagram for Eccentrically Loaded Joints showing the Non-Linearity
305H
Figure 8-13 - An Eccentrically Joint Loaded to the Interface Opening Limit with the
307H
Figure 8-14 - Definition of Flange Compliance due to Bending Loads at Different Points . 121
308H 710H
Figure 8-17 - Simplified Joint Diagram for Cantilevered Flanged Joints............................. 125
311H 713H
Figure 8-18 - The Joint Diagram for Non-Linear Gasket Joints .......................................... 126
312H 714H
Figure 8-19 - Idealised Linear Pressure Distribution across an Eccentrically Loaded Joints
313H
Interface............................................................................................................ 128
715H
Figure 9-2 - Load Deformation Curve for a Fastener in Direct Shear................................. 131
315H 717H
Figure 9-4 - The Joint Diagram for Each Fastener in Figure 6.3 ........................................ 133
317H 719H
Figure 9-5 - Typical Unsymmetrical Friction Grip Shear Joint ............................................ 133
318H 720H
Figure 9-6 - Fastener Bending and Tension in an Unsymmetrical Shear Joint .................. 134
319H 721H
Figure 9-7 - Simple Friction Grip Joint with Added Tension Loading ................................. 135
320H 722H
Figure 9-8 - Reduction of Shear Capacity with Increasing External Axial Load ................. 136
321H 723H
Figure 9-9 - Joint Design Features that Increase Slip Resistance ..................................... 137
322H 724H
Figure 9-12 - Graphical Determination of the Margin of Safety on Combined Loads ......... 142
325H 727H
12
ECSSEHB3223A
16April2010
Figure 9-13 - Fastener Group with hole clearance ............................................................. 143
326H 728H
Figure 9-14 - Shear load parameter versus hole clearance ............................................... 144
327H 729H
Figure 9-15 - Nomenclature for Net Tension Section Calculation ...................................... 145
328H 730H
Figure 9-16 - Ultimate Strength Reduction Factors for Net Tension Section Failure.......... 146
329H 731H
Figure 9-17 - Initial Bearing Contact Elastic Stresses and Deformations ........................... 147
330H 732H
Figure 9-18 - Elastic / Plastic Stresses and Deformations After Increased Loading .......... 148
331H 733H
Figure 9-22 - Nomenclature for Plate Shear Out Calculation ............................................. 150
335H 737H
Figure 9-24 - Load v Rotation Curve for a Typical Eccentric Loaded Shear Joint.............. 152
337H 739H
Figure 9-27 - Shear Joint Example to Show the Method of Net Tension Section
340H
Figure 11-3 - Typical Stress Ratio Mean Stress Diagram ............................................... 173
349H 751H
Figure 11-4 - Typical Preload Effects on Fatigue Life of Fasteners ................................... 174
350H 752H
Figure 11-5 - Joint Diagram Comparing Steel and Titanium Fasteners ............................. 175
351H 753H
Figure 11-9 - ESACRACK Model of a Fillet Crack under a Fastener Head........................ 181
355H 757H
13
ECSSEHB3223A
16April2010
Figure 12-4 - Prevailing torques during a series of loosing and retightening cycles of a LN
359H
29949 M6 fastener and with Helicoil LN9499 made of CuSn6 ......................... 198 761H
Figure 13-5 - Examples of Tensile Stresses in Short Transverse Direction Resulting from
364H
Tables
Table 5-1 Margins of Safety to be analysed for a threaded fastener.................................... 38
365H 767H
Table 5-4 - Minimum safety factors for analyses of threaded fasteners ............................... 43
368H 770H
Table 6-1 - Uncertainty factors of tightening methods (from reference 5.1) ......................... 51
369H 771H
Table 6-3 - Approximate values for plastic deformation caused by embedding ................... 59
371H 773H
Table 7-5: Preload in a Concentric Axially Loaded Joint, fastener material properties ..... 101
376H 778H
Table 7-6: Preload in a Concentric Axially Loaded Joint, Thread parameters.................... 101
377H 779H
Table 7-7: Preload in a Concentric Axially Loaded Joint, flange material properties.......... 102
378H 780H
Table 11-1 - Stress Intensity Factors for Thumbnail Crack in a Fastener Thread .............. 180
381H 783H
Table 11-2 - Stress Concentration Factors at a Fastener Minor Diameter ......................... 180
382H 784H
Table 12-1 - y V for a Range of Fastener Sizes and Typical Friction Conditions ......... 189
384H 786H
Table 12-2 - Typical Scatter of Ultrasonic Extensometer Preload Readings Using a Direct
385H
Table 12-4 - Prevailing Torques of Locking Devices after Several Loosening and
387H
14
ECSSEHB3223A
16April2010
Table 13-1 - Galvanic Series of Metals in an Aqueous Environment ................................. 205
388H 790H
Table 16-1 - Test Factors to be used for Given Probability of Failure ................................ 222
390H 792H
Table C-2 - Friction Coefficients for Common Joint Materials (from Reference 8.3) .......... 228
394H 796H
15
ECSSEHB3223A
16April2010
Introduction
0B
ABSTRACT
Theaimofthepresenthandbookistoassistthestructuraldesignengineersbypresentingthemina
single document with all the information relevant to the use of threaded fasteners in jointed
spacecraftcomponents.
BACKGROUND
Threaded fastener joints represent critical areas within structures and it is evident that a
standardisedmethodologyisneededfortheirdesign.
Thus,theaimsofthisdocumentare:
to bring together into one document a reliable and systematic design procedure to aid the
designerandimprovetheefficiencyofthedesignprocess,and
to provide an approved and reliable design procedure that will reduce the possibility of
designerrorsandsubsequentstructuralfailure.
This document addresses these aims with particular reference to Space Structures and related
hardware. This constraint is reflected in the range of joint geometries covered and structural
applications,andalsotherangeoffastenertypesandmaterialsforwhichdataispresented.
PURPOSEOFTHISISSUE
ESAPSS03208Guidelinesforthreadedfasteners,onwhichthisdocumentisbased,provideda
number of methodologies for analysis of threaded fastener joints in spacecraft structures. This
documentisanupdateofthePSStakingintoaccountmorerecentdevelopmentsinthefield.Itis
intended to standardise the analysis approach and corresponding documentation for threaded
fastenerjointsdevelopedinECSSprojects.Therefore,thisdocumentpresentsaMarginofSafety
foreachmodeoffailurethatshouldbeconsidered.Allrelevantmarginsofsafetyshouldbeshown
inverificationdocumentation.
GUIDELINELIMITATIONANDIDENTIFIEDPROBLEMAREA
Thishandbookdoesnotprovidetheoryforthefollowingtypesofanalysis:
Complianceofcantileveredflangejointswithnoncircularflanges
Frictiongripofeccentricshearloadedjoints
Nofailurecriteriaareprovidedforfastenerbending,whichsometimesshouldbeconsidered
(seeSubsection 9.3.2)
804H
GUIDELINESFORTHEUSEOFTERMSANDDEFINITIONS
Duetotheextensiveusedofthreadedfastenersinamultitudeofcountriesandindustries,thereisa
wide variety of terminology in use today. For the purposes of making a clear and coherent
handbook it was necessary to adopt a single consistent set of terminology. The terminology that
was chosen is presented in 3.2. To avoid confusion, it is recommended to refer to this section
805H
frequentlywhenreadingthedocument.
16
ECSSEHB3223A
16April2010
1
Scope
2B
Theusersofthisdocumentareengineersinvolvedindesign,analysisorverificationofjointsonstructures
used for space missions. It is a guidelines document; therefore it includes advisory information rather
thanrequirements.
Thisdocumentisintendedtobeapplicabletoanytypeofjointthatismechanicallyconnectedbythreaded
fasteners(e.g.bolts,screws,etc).Itiswrittenforjointsmadefrommetallicmaterials.However,subjectto
the engineering judgement of the user, many of the procedures presented herein may be applicable to
jointsmadefromcompositematerials.
17
ECSSEHB3223A
16April2010
2
References
3B
Referencesareprovidedattheendofeachsection.
18
ECSSEHB3223A
16April2010
3
Terms, definitions and abbreviated terms
4B
Forthepurposeofthisdocument,thetermsanddefinitionsfromECSSSST0001apply.
3.2.4 flange
clampedpartthattransmitsloadsfromathreadedfastenertootherregionsofthestructure
3.2.5 insert
partwithathreadedhole,whichisembeddedinaweakermaterialinordertoincreasetheloadsthatcan
betransmittedthroughthejoint
NOTE Itisnormallyametallicpart.
19
ECSSEHB3223A
16April2010
3.2.6 joint
regionofthestructureinthevicinityofthethreadedfastener
NOTE The dimensional extents of the joint depend on the engineering issues
underconsideration.
3.2.9 nut
devicewithaninternallythreadedholewiththeprimaryfunctionsbeingtheapplicationoftensiontothe
shaftofathreadedfastenerandcompressiontotheclampedparts
3.2.12 prying
changeinafastenersloadingcausedbyaleveringactioninflexibleflanges
NOTE Itoccursineccentricallyloadedjoints,seeSection 8.
806H
3.2.15 thread
helical groove on the external surface of a shaft (male thread) or the internal surface of a hole (female
thread)
20
ECSSEHB3223A
16April2010
3.2.16 threaded hole flange
flangewithathreadedholeintowhichathreadedfastenerisinsertedandtightened
Forthepurposeofthisdocument,theabbreviatedtermsfromECSSSST0001andthefollowingapply:
Abbreviation Meaning
ASME AmericanSocietyofMechanicalEngineers
BS BritishStandard
CNES CentreNationaldEtudesspatiales(Frenchspaceagency)
DIN DeutschesInstitutfrNormung(Germaninstitutefor
standardization)
ECSS EuropeanCooperationonSpaceStandardization
EDM ElectronDischargeMachining
ESA EuropeanSpaceAgency
HB BrinellHardness
ISO InternationalStandardsOrganization
IVD IonVapourDeposited
LEFM LinearElasticFractureMechanics
LN LuftfahrtNorm(Germanaviationstandard)
M Metrictypethread
MGSE MechanicalGroundSupportEquipment
MJ MetricJtypethread(forimprovedfatigueperformance)
MoS MarginofSafety
21
ECSSEHB3223A
16April2010
Abbreviation Meaning
NASA NationalAeronauticsandSpaceAdministration
PSS Procedures,StandardsandSpecifications
PTFE PolyTetraFluoroEthylene
SCC StressCorrosionCracking
UN Unifiedthread
UNC UnifiedCtypethread
UNJ UnifiedJtypethread(forimprovedfatigueperformance)
3.4 Variables
21B
Forthepurposeofthisdocument,thefollowinguppercaservariablesapply:
Variable Meaning
A Crosssectionareaofajointpart(intheplane
perpendiculartothefasteneraxis)
AS Effectivecrosssectionalareaofafastenerforstress
analysis
Asm Effectivecrosssectionalareaofafastenerforstiffness
calculations
Ath Areaofthe(cylindrical)surfaceassumedtofailduring
threadpullout
A0 Minimumcrosssectionalareaofafastenershank
A3 Crosssectionalareaofafastenerbasedonthethreadroot
diameter,d3
C Transversewidthofaneccentricjoint(seeFigure7.2.1)
Davail Availablediameterforcompressionzonespreading(see
Section 7.7)
807H
Dh Nominaldiameterofahole
Dlim Maximumconediameterofcompressionzonespreading
(seeSection 7.7)
808H
Duh,brg Outerdiameteroftheunderheadorundernutbearing
surface(seeSection 7.7)
809H
E Youngsmodulus
FA Axialforcetransmittedbyjoint
FQ Shearforcetransmittedbyjoint
22
ECSSEHB3223A
16April2010
Variable Meaning
FV Nominalpreload
FZ Maximumlossofpreloadduetorelaxation(includesall
sources,e.g.embedding,creep,etc)
Fb Tensileforcetransmittedbyafastener
Fc Compressionforcetransmittedbytheclampedparts
Fb , A Incrementinfastenertensioncausedbytheapplicationof
theexternalaxialload,FA
Fc , A Incrementinclampedpartscompressioncausedbythe
applicationoftheexternalaxialload,FA(negativefor
increasedcompression)
FK Clampingforceactingattheintersticebetweentwo
flanges
Fth Axialforcetransmittedbythethread
FT Maximum(mostpositive)incrementinafastenersaxial
loadduetothermoelasticeffects
FT Minimum(mostnegative)incrementinafastenersaxial
loadduetothermoelasticeffects
Gc Radiusofgyrationoftheeffectiveinterfaceareabetween
clampedparts
Ic Secondmomentofareaoftheclampedpartsinterface
area
K Stiffnessofajointpart(N/m)
KC Criticalstressintensityfactor
KISCC Criticalstressintensityfactorforstresscorrosioncracking
Lj Thejointlength.Thecombinedthicknessofallclamped
parts(measuredwithoutanyexternallypreloador
appliedload).
Ln Lengthofthenut
Lh Lengthofthefastenerhead
Lc Lengthoftheclampedparts
Lb Changeinlengthofthefastener
Lc Changeinlengthoftheclampedparts
Mapp Peaktorqueappliedtothefastenerornutduring
tightening(includestheprevailingtorque,MP)(Mappis
alsocalledinstallationtorque)
MB Bendingmomenttransmittedbyjoint(seeFigure4.3.1)
MP Prevailingtorque(alsocalledrunningtorque)
MT Torsionalmomenttransmittedbyjoint(seeFigure4.3.1)
23
ECSSEHB3223A
16April2010
Variable Meaning
Mth Torsionalmomenttransmittedbythethread(doesnot
includeprevailingtorque,MP)
Muh Torsionalmomenttransmittedbyfrictionalforcesatthe
underheadorundernutinterface
RS Shearstrengthratio
RZ Averagesurfaceroughness
Sr Slipresistanceofajoint
Tworking Theworkingtemperatureofthejoint(specifiedwith
upperandlowerbounds)
Treference Thereferencetemperature(generallytakenasthe
assemblytemperature)
T Changeintemperature
Wp Polarsectionmodulus
Yieldcorrectedpolarsectionmodulus
W p
Forthepurposeofthisdocument,thefollowinglowercasevariablesapply:
Variable Meaning
a Loadeccentricity(distancefromjointaxistoaxisof
appliedload)
ak Distancebetweentheedgeoftheunderheadbearing
surfaceandthepointofloadintroduction(usedfor
eccentricjoints,seeSection6.6)
al Edgedistance(measuredfromholeaxistoedgeofflange)
b Effectivelongitudinalwithofthejoint(seeFigure7.2.2)
d Nominaldiameterofafastener
dS Effectivediameterofafastenerforstresscalculations
dhead Diameteroffastenerhead
dsha Diameterofashanksegment
dsm Effectivediameteroffastenerusedforstiffness
calculations
duh Effectivediameteratwhichunderheadorundernut
frictionalforcesact
duh,brg Outerdiameteroftheunderheadorundernutbearing
surface
d0 Minimumdiameterofafastenershank
24
ECSSEHB3223A
16April2010
Variable Meaning
d2 Pitchdiameterofafastenerthread(assumeddiameterat
whichtorsionforcesareintroducedwithinathread)
d3 Diameterofafastenersthreadroot
fZ Plasticdeformationofthejointduetoembedding
f T Thermalelongationofanitem
h Thicknessofaflange(referredtoasalengthfor
compliancecalculations)
hk Thedistancebetweenthepointofloadintroductionand
theinterstice(usedforeccentricjoints,seeSection6.6)
n Loadingplanefactor
p Pitchofathread
s Clampingeccentricity(distancefromjointaxistofastener
axis)
sfy Safetyfactortobeappliedtoyieldanalysis
sfgap Safetyfactortobeappliedtogappinganalysis
sfult Safetyfactortobeappliedtoultimatestrengthanalysis
sw Sizeofwrench(measuredacrosstheflatsofaheador
nut)
x Numberoffayingsurfacesthatresisttheshearload
3.4.3 Symbols
105B
Forthepurposeofthisdocument,thefollowingsymbolsapply:
Symbol Meaning
Coefficientofthermalexpansionforamaterial
Uncertaintyfactorusedinthetypicalcoefficientmethod
(notapplicabletotheexperimentalcoefficientmethod)
Coefficientofutilization(ratioofpreloadstresstoyield
stress)
Underheadbearingangle(seeFigure4.4.3)
Torquewrenchaccuracyusedintheexperimental
coefficientmethod(expressedatavalue)
Compressionconehalfangle(seeSection6.6)
Threadhelixangle(seeFigure5.2.1)
Localfrictionangleatthreadinterface
Halfangleofthreadgrooves(seeFigure4.4.3)
25
ECSSEHB3223A
16April2010
Symbol Meaning
b Complianceofthefastener(seeSection6.5)
c Complianceoftheclampedparts(seeSection6.6)
V Axialstressinfastenerduetonominalpreload,FV
y Yieldstress
th Frictioncoefficientatthethreadinterface
uh Frictioncoefficientattheunderheadorundernut
interface
s Frictioncoefficientattheslippinginterface(theinterstice
betweentheloadedflanges)
br Bearingstress
v.m. Vonmisesstress
Shearstress
Basicforceratio(assumesn=1andaconcentricjoint)
n Forceratioofaconcentricjointincludingtheloading
planefactoreffect
e, n Forceratioofaneccentricjointincludingtheloading
planefactoreffect
C Forceratioofajointwithcantileverflanges
Inversioncompliance(usedforcircularflange
compliance,seeSubsection7.3.1)
26
ECSSEHB3223A
16April2010
3.4.4 Subscripts
106B
Forthepurposeofthisdocument,thefollowingsubscriptsapply:
Subscript Meaning
b Fastener(orbolt,screw,etc.)
c Clampedparts(referstothenetelasticpropertiesofall
theclampedpartsinajoint)
nom Nominal
max Maximum
min Minimum
free Usedforfastenerorclampedpartslengthstoindicate
thatthemeasureisintheabsenceofappliedloading
sub Substitutional
th Thread
y yield
ult Ultimate
eff Effective
eng Engaged
crit Critical
tot Total(oroverall)
sep Separation
comb Combined(loads)
27
ECSSEHB3223A
16April2010
4
How to use the guidelines
5B
4.1 Introduction
22B
This guidelines document contains a number of selfcontained sections covering specific joint
configurations and joint loading situations. Additionally, it contains a number of sections on ancillary
subjectsrelatedtojointdesign.
In general, the user of this document will be experienced in structural design and will make use of the
varioussectionsasneeded.However,toassistthefirsttimeuserandtoprovideachecklistofessential
featurestobeconsideredwhendesigningjoints,thissectionincludesaprocedureguidingtheusertothe
relevantsectionsofthisdocument.
Thevariousstepsarediscussedbelowandthensummarisedinaflowchart.Sincethepurposeofthispart
of the Guidelines is to direct the user through the document, in general, the logic criteria on which the
decisions(accept/reject,yes/no)aremadearenotstated,rathertheseareincludedineachoftherelevant
sections(referencedasappropriate).
4.2 Procedure
23B
4.2.1 Stage 1
107B
The precept on which the guideline is based is that the joint being analysed can be isolated from the
surrounding structure. The joint can then be represented by one of five main joint categories, which
dependonitsconfigurationandappliedforces.Therefore,thefirsttaskistodefinethephysicalextentof
thejointwithinthestructure,andidentifyitscategory(seeSection 5).Itisassumedthatananalysisofthe
810H
structural system in which joint exists has been carried out in sufficient detail to specify the loads
transmittedviathejoint.
Thecaseofcombinedaxialandbendingloadingistreatedasaneccentricaxiallyloadedjoint,anddealt
withinClause 8.AmethodisgiveninClause 9todealwithcombinedaxialandshearloadingwhichis
811H 812H
assumedtobethemostfrequentlyencountered.
28
ECSSEHB3223A
16April2010
4.2.1.3 Dynamic Loading
238B
Inallbutexceptionalcasestheworkingloadonthejointwillbedynamic.Thisisdealtwithinmoredetail
in each section, as appropriate. However, it is recommended that the engineer, at an early stage in the
design process, is aware of theinfluence of varyingexternal joint loads on the loads transmitted by the
fastener and at the clamped interface. Reference is therefore made to Section 7.4, which discusses the
813H
forceratio.
Analysis of preloaded joints operating in a varying temperature environment is needed to ensure that
they meet functional requirementsacross thefull temperature range.Section 6.4.1e coversthe effects of
814H
temperature change on the preload. Furthermore, certain joints have requirements for thermal
conductivity to be either above or below certain limits. Section 10 covers the analysis of thermal
815H
conductivity.
4.2.2 Stage II
108B
Depending on the category of joint, the user is directed to the appropriate section, which details the
equations needed to calculateMarginsof Safety(MoS)forallrelevant failuremodes.Worked examples
areprovidedtoillustratetheapplicationoftheproposeddesignprocedures.
Formanyjointdesignstheeffectsofembeddingandrelaxationareimportant.Theseeffectsarecoveredin
Section 6.4.1e.
816H
For a more detailed analysis of either fatigue or fracture mechanics analysis, the user is referred to
Section 11,whichtreatsthesesubjectsinmoredetail.
817H
If,withinthegeneraldesignconstraints,thejointmeetsneitherthebasicnorthedetailedfatiguecriteria,
itmaybenecessarytoreconsidertheinitialjointdesignconcept.
4.2.3.1 Corrosion
243B
Environmentaleffectsareidentifiedasanimportantdesignparameter.Section 13discussesanumberof
818H
corrosionmechanismsrelevanttojointcomponentsandprovidesachecklistagainstwhichtheproposed
design can be assessed and possible problem areas identified. Where a severe or intractable corrosion
problem is predicted it may be necessary to review the initial structural concept, which may lead to a
changeinthejointsmaterialsordetaileddesign.
29
ECSSEHB3223A
16April2010
4.2.3.2 Methods of Preloading & Lubrication
244B
Itisexpectedthatatsomestagethefastenerassemblymethodisspecified,togetherwiththeuseornotof
a lubricant. The method of preloading can be critical as this determines the accuracy with which the
fastenerpreloadcanbeachieved.
Section 12 covers a number of different methods of installing preloaded fasteners. Design data and
819H
estimatesofachievableaccuracyforthepreloadaregivenwhereappropriate.
Section 14containsadviceonlubricationaspectsandcontainsdetailsofthecharacteristicsofaselected
820H
rangeoffastenerlubricants.
Thecalculationproceduresneedthataparticularpreloadingmethodisspecifiedatthestartofthedesign
process. It is necessary to check later that the specified preloading method is capable of ensuring the
preloadaccuracytobeachieved,andalsothatitiscompatiblewithjointlocationandaccessconstraints.
4.2.4 Stage IV
110B
Section 15providestheengineerwithaguidetothefactorsthatneedtobecontrolledduringproduction
821H
toguaranteeadequatejointperformance.
Considerationneedstobegiventodemonstratingthatthejointdesignisadequate.Thismaybeachieved
either by testing the complete structure or by isolating and testing individual joints. These aspects are
coveredinSection 16. 822H
4.3 Annexes
24B
The following Annexes of the document contain additional technical data that may assist the engineer
undertakingdesignofthreadedfastenerjoints:
A:Recommendedstartingvaluesforseatingtorque
B:MeasuredFrictioncoefficientsoffasteners
C:Typicalfrictioncoefficientsforjointmaterials
D:Materialdataoffastenerstypicallyusedintheaerospaceindustry
30
ECSSEHB3223A
16April2010
5
General Fastener Analysis Guidelines
6B
5.1 Introduction
25B
For the purposes of this guidelines document the different types of joints have been divided into the
followingfivemaincategories;
a. ConcentricAxiallyLoadedJoints,
b. EccentricAxiallyLoadedJoint,
c. ShearLoadedJoints,
d. CombinedLoadedJoints,and
e. LowDutyJoints.
These are defined by the geometry and system of loading. Due to the different verification procedures
ShearLoadedJointsandtheCombinedLoadedJointsarefurthersubdividedinto;
BearingJoints,and
FrictionGripJoints.
Thecategorisationofjointsinthismannerisreflectedinthestructureandformatofthedocument.Within
eachofthemaincategories,subcategoriescanbeidentifieddependingonspecificgeometricalorloading
attributes.Detailsofthemaincategoriestogetherwithexamplesoftypicalsubcategoriesareillustratedin
thissection.Itisintendedthatthiswillbothindicatetherangeofjointtypescoveredintheguidelineand
operateasanindextodirecttheusertotherelevantpartsofthedocument.
5.2.1 Overview
111B
The definition for a particular joint configuration depends on the geometry of the clamped parts and
fasteners,andtheeffectiveloadingappliedatthefastener(orfastenergroup).AsstatedinSection1itis
assumed that the user has knowledge of the system of loads acting on the joint in the vicinity of the
fastener(s);thesebeingderivedbyanalysisoftheoverallstructurepriortodetailedconsiderationofjoint
design.Adequatedefinitionoftheappliedloadingsystemisessentialasthismaydeterminetheparticular
categorytowhichajointbelongsandhencethemethodofanalysisadopted.Thisisparticularlycritical
when distinguishing between concentric and eccentric axially loaded joints. The selection of the
31
ECSSEHB3223A
16April2010
appropriate analysis for the combined loading case is also dependent on adequate specification of the
loadingsystem.Morecomplexsituationsarisewherecombinedloadingoccurs,typicallyaxial,shearand
bending. However, where a single load is dominant in a combined loading case and other load(s) are
small it may be possible to assume the joint falls into one of the first three categories outlined below,
thereby simplifying the analysis without significant reduction in accuracy. Criteria are included to
determinewhensuchanassumptionisapplicable.
Thedistinguishingfeatureofthisjointcategoryisthatthelineofactionoftheappliedloadingonthejoint
isparalleltoandcoincidentwiththelongitudinalaxisofthefastener.Therefore,anycombinationofjoint
geometryandsystemofappliedloadswhichconformtothisdefinitionmaybeanalysedbythemethods
andproceduresspecifiedforthiscategory.
Examplesofjointsforthiscategoryareillustratedin Figure51to Figure53.Forclarityasinglefastener
823H 824H
isshown,althoughinmanycasesthesameanalysismaybeappliedtoagroupoffasteneriftheyareina
symmetric pattern. However, it should be noted that the effect of flange flexibility can lead to eccentric
fastenerloadsinjointsthatwithsymmetricfastenerpattern(see Figure53,Joint17).
825H
For this joint category the line of action of the applied loading on the joint, whilst being parallel to the
fastenerlongitudinalaxis,isnotcoincident,butoffset.Theresultofthisisthataprying(orprising)
action occurs between the clamped parts of the joint such that bending loads are introduced under the
boltheadandintheshank.Threefundamentalloadingcasevariantscanbeidentifieddependingonthe
relativepositionoftheboltaxis,lineofactionoftheappliedload,andthejointcentroid.
This category of jointandits variants representa large proportion of the joints encountered in practice.
The main examples of these are illustrated in Figure 51 to Figure 53. For clarity a single fastener is
826H 827H
shown,althoughinpracticethismaybeagroup.
Theprinciplefeatureofthisjointcategoryisthatthelineofactionoftheappliedloadingonthejointsisin
theplaneoftheclampedpartsimmediatelyadjacenttothefastener,andthereforenormaltothefastener
longitudinalaxis.Theforcesneednotbecoplanarifthejointisofthenonsymmetricalorsinglelapshear
type.Jointsofthistypemaybefurthersubdividedintofrictiongriporbearingcategories,dependingon
whetherloadistransferredthroughthejointbyfrictionatthefayingsurfacesorbytransverseshearinthe
fastener(s). Examples of joints in this category are illustrated in Figure 51 to Figure 53. For clarity a
828H 829H
singlefastenerisshown,althoughinpracticethismaybeagroupifthejointhasmultiplefasteners.
Theessentialfeatureofthisjointcategoryisthatmorethanonesystemofloadsactonthejointrelativeto
the fastener axis. In the most general case features of all the preceding categories will be combined.
Examples of joints for this category are illustrated in Figure 51 to Figure 53. For simplicity a single
830H 831H
fastenercanoftenbeconsidered,althoughinpracticethesamemaybeagroupifthejointhasmultiple
fasteners.
32
ECSSEHB3223A
16April2010
Duringthedesignprocess,jointsinitiallybeingplacedinthe(mostgeneral)categoryofcombinedloaded
jointsmaybereclassifiedintoasubcategory,forthepurposeofanalysis.Thisguidelinedoesnotprovide
specificcriteriafordeterminingwhenthissimplificationcanbeassumed.Rather,theappropriateanalysis
method should be determined using engineering judgment considering the relative magnitudes of the
shearandaxialloads,theconfigurationofthejointandanyotherrelevantattributesofthedesign.
Joints in this category have loadings and configurations that fit into one of the preceding categories,
howevertheyformauniquecategorysincetheyhavesmallexternalloadingwithrespecttothefastener
strength.Inmanycasesthiswillbereadilyapparentduetotheparticularapplication,e.g.holddowns,
accesspanelattachments,etc.Someexamplediagramsofjointsthatfrequentlyfallintothiscategoryare
shownin Figure51to Figure53.
832H 833H
33
ECSSEHB3223A
16April2010
Figure51JointCategories(1to6)
34
ECSSEHB3223A
16April2010
Figure52JointCategories(7to12)
35
ECSSEHB3223A
16April2010
Figure53JointCategories(13to17)
36
ECSSEHB3223A
16April2010
5.3.1 Overview
117B
The calculation of a mechanically fastened joint begins with the definition of the external operational
loadingenvironment,whichincludestheaxialloadFA,theshearloadFQ,thebendingmomentMBandthe
torsional moment MT (if present). For the fastener analysis it is assumed, that these external forces and
momentsareknown. Figure54showsajointwiththedimensionsindicated.
834H
MT
FA
FQ MB
FQ
MB
FA
MT
Figure54DefinitionsofForcesandMomentsActingonaJoint
Based on the applied loads and expected operating environment the joints configuration can be
determined including details such as the preload in the fastener, flange dimensions and the materials
used.Followingthis,thefunctionalityandstrengthofthejointsystemshouldbeverifiedbycalculating
allrelevantMarginsofSafety(MoS)accordingtheEquationspresentedinthisdocument(seebelow).
PositivevaluesshouldbedemonstratedforallMoSthatareapplicabletothespecifictypeofjoint.
ThegenericequationforaMarginofSafetyis,
Allowable Load
MoS 1 [5.3.1]
Limit Load sf
37
ECSSEHB3223A
16April2010
wheretheAllowableLoadistheloadatwhichamodeoffailureispredictedtooccurandtheLimitLoadis
the determined by analysis of the service loads expected during the joints lifetime. The sf is the safety
factorcorrespondingtothespecificmodeoffailurebeingconsidered.
For each of the joint modes of failure covered in this document, equations derived for the MoS are
provided. Table51showswhereeachoftheMoSderivationsisgiveninthishandbook.
835H
Table51MarginsofSafetytobeanalysedforathreadedfastener
MoS Relevantfailuremode Relevantjoint Referto
variable Description categories handbook
name Friction bearing setion
grip
MoSti,y& Combinedstressesinfastener Yes Yes 6.5
836H
MoSti,ult duringtightening
(yieldandultimate)
MoSsep Separation(orgapping)of Yes Yes 7.9
837H
flanges
MoSA,y& AXIAL STRESS IN FASTENER FOR Yes Yes 7.10
838H
MoStot,ult preloadandexternalload(yield
andultimate)
MoSth Shearpulloutofthread(either Yes Yes 7.11
840H
nutorinsert)
MoScrush Crushingofflanges Yes Yes 7.12
841H
MoSQ,ult (yieldandultimate)
MoScomb,y& Combinedloadsonfastener No Yes 9.3.2
844H
MoScomb,ult (yieldandultimate)
MoSnet Nettensionsectionfailureof No Yes 9.3.3
845H
flange
MoSbr,y& Bearingstressonflangehole No Yes 9.3.4
846H
MoSbr,ult (yieldandultimate)
MoSso Shearoutofflange No Yes 9.3.5
847H
1.Notincludedinthisversionofthehandbook,howeverthisfailuremodeshouldbeassessed
38
ECSSEHB3223A
16April2010
Forafasteneranalysisseveraldimensionsareimportant.ThesearedepictedinFigure55and Figure56.
848H
duh,brg
Lh
Dh t1
dsha
t2
Ln
sw
Figure55FastenerDimensions
=180
=100
Figure56Definitionoftheunderheadbearingangle
39
ECSSEHB3223A
16April2010
Table 52 shows the definitions of the physically measurable fastener variables used in this handbook.
849H
Someoftheseareomittedfrom Figure55forclarity.
850H
Table52PhysicalDimensionsofFasteners
Variable Definition Notes
p Pitchofthread
h Heightofthebasicthread
d Nominalfastenerdiameter Outsidediameterofthread
d0 Diameteratsmallestcrosssectionoffastener
shank
d3 Minordiameterofthread Forthetruethread,notthebasic
profile
dsha Shankdiameter Forneckeddownfasteners
duh,brg Outerdiameterofbearingarea Eitherunderaheadorundera
nut
Dh Nominaldiameterofholeinflange
A0 Smallestcrosssectionoffastenershank
A3 Crosssectionalareaatminordiameterof
fastenerthread
Underheadbearingangle Always100forcountersunk
aerospacestandardfasteners
(rightsideof Figure56)
851H
sw Sizeofwrench Acrossflatsdimensionofeithera
headoranut
Figure 57 shows the geometric definition of the metric (M and MJ) threads. Both the male and female
852H
threads are based on the Base Profile shown in the left of the figure. The base profile defines the
maximum spatial envelope of the male threaded part (normally the fastener itself), and the minimum
spatialenvelopofthefemalethreadedpart(normallythenutorthreadedhole).Theexactthreadprofiles
aredefinedintherelevantstandards.TheonlydifferencebetweenMandMJtypethreadsisthethread
rootradius,whichislarger(togiveimprovedfatigueperformance)intheMJtypethread.
Table53showsthedefinitionsofthederivedfastenervariablesusedinthishandbook.
853H
40
ECSSEHB3223A
16April2010
h
d3 Thread root radius
h
4
Male thread
(external)
2 =60
p D3 Female thread
(internal)
h
2
d2
h
8
Base profile
Figure57ThreadGeometry
41
ECSSEHB3223A
16April2010
Table53DerivedDimensionsforMetricFasteners
Variable Notation Equation
Pitchdiameter d2 =d0.64952p [5.4.1]
Minordiameter d3 =d1.22687p [5.4.2]
Diameterusedforstress dS =0.5(d2+d3) [5.4.3]
calculation (forMtypethread)
=seestressareabelow [5.4.4]
(forMJtypethread)
Diameterusedforstiffness dsm =d3 [5.4.5]
calculation
Effectivediameteroffriction duh =0.5(Dhead+Dh) [5.4.6]
underheadornut
Stressarea AS =0.25dS2 [5.4.7]
(forstandardfasteners,Mtype
thread)
0.25(d3)2[2(d3/d2)2]
(forstandardfasteners,MJtype
thread)
=0.25d02 [5.4.8]
(forreducedshankfasteners)
Stiffnessarea Asm =0.25dsm2 [5.4.9]
Smallestcrosssectionoffastener A0 =0.25d02 [5.4.10]
shank
Ingeneral,thecorrectequationshouldbefoundintherelevantfastenertechnicalspecification,e.g.ISO
9152forTitaniumfasteners(Reference 5.1).
854H
For unified threads (UNC, UNF, etc) the same equations apply except that the pitch, p, is not normally
giveninthefastenerspecification.Inordertousetheaboveequations,itisnecessarytocalculatefirstthe
pitchbyinvertingthethreadsperinchspecification,whichisgiveninthethreaddesignation.
Tolerance information should also be taken into account in determining some or all of the dimensions
usedforstresscalculations.
ISO5855(Reference 5.2)specifiestheMJtypethread,whichisspeciallysuitedforaerospaceapplications
855H
andhasalargerdiameteratthethreadroottoimprovethefastenersfatiguelife.TheUNJthreadisthe
analogousmodificationforunifiedthreads.
42
ECSSEHB3223A
16April2010
5.5.1 Overview
121B
Table54showstherecommendedsafetyfactorsfortheanalysisofthreadedfastenersinEuropeanspace
856H
projects (e.g. on behalf of ESA). Ultimate responsibility for the definition of safety factors for a space
projectrestswiththeprojectmanagement.Sincesafetyfactorsdependstronglyonthetypeofload,the
statedsafetyfactorsaretoberegardedasminimumvalues.
The safety factors can be decreased significantly if a test with the appropriate test factor is performed
either on dedicated test structure identical to the flight structure (a prototype test) or on the flight
hardware(aprotoflighttest).
Table54Minimumsafetyfactorsforanalysesofthreadedfasteners
Verification Testtype Test Safety Safety SafetyFactorfor
approach factor factorfor factorfor Jointseparation(or
yield ultimate gapping)
strength strength (sfgap)
(sfy) (sfult) Safety Other
critical
Analysisonly N/A N/A 1.25 2.0 1.4 1.2
Analysisand Qualification 1.4 1.0 1.4 1.4 1.2
prototypetest
Analysisand Acceptanceor 1.2 1.0 1.4 1.4 1.2
protoflighttest proof
Thesafetyfactorsdescribedin Table54areapplicableforallmetallicflightstructuresandincludethe
857H
normaluncertaintiesinloadpredictionandmaterialcondition.Thetestfactorisamultiplierontheloads
thatareappliedduringthetest,anddependsontheverificationapproach.
In cases where the joint load distribution is difficult to accurately predict, an extra joint fitting factor of
1.15shouldbeappliedinadditiontothesafetyfactorsin Table54.Thisfactorshouldbeusedonyield
858H
andultimateloadsinthestructuralanalysisofallfittingswiththreadedfasteners,whosestrengthisnot
provenbylimitandultimateloadtestinwhichtheactualstressconditionsaresimulatedandmeasuredin
the joint and surrounding structure. This factor should apply to all portions ofthe fitting, the means of
fastening,andthebearingonthemembersjoined.
The fitting factor need not be used with limit and ultimate loads where the specific type of joint is
strengthverifiedbycomprehensivelimitandultimatetests.Thiswouldcommonlybethecaseforjoints
suchascontinuousrowoffastenersinsheetorplate.
43
ECSSEHB3223A
16April2010
5.6 References
30B
5.1
859H ISO9152:1998 Procurementspecification:Bolts,withMJthreads,intitaniumalloys,
strengthclass1100MPa
5.2
860H ISO58552:1999 MJThreads,Part2Limitdimensionsforboltsandnuts
44
ECSSEHB3223A
16April2010
6
The Preload
7B
6.1 31BOverview
The integrity of a threaded fastener joint, either shear or tension, is maintained by the preload in the
fastener. It has been shown that the optimum jointhasas higha preloadas possible.However, equally
importantistheneedtoaccuratelycontrolthepreloadtoaspecifiedvalue.Inmanyinstancesthereason
forfailureofboltedjointsisincorrectinitialpreload,lossofthepreloadduetoloosening,orrelaxationof
thepreloadduetoavarietyofmechanisms.
The fundamental problem, however, is to assemble the joint and accurately achieve the required initial
preloadinthefastener.Anumberofmethodsareavailableforcontrollingorachievingthepreloadduring
assembly.Thesearediscussedinthissection.
Themethodsdescribedaregenerallyamenabletotheoreticalanalysis,andwhererelevantthisisincluded.
It should be noted that, although apparently simple, the bolted joint is extremely complex. Simple
theoreticallycorrect methods suffer from inaccuracies,some quite subtle. The accuracy of the analysis
dependsonthespecificationofmaterialsandgeometricalparameters,whichinpracticeareimpossibleto
specifywithoutresortingtoempiricaldataderivedfromtestsonactualorrepresentativejoints.Friction
dataforsomecommonsjointmaterialsaregiveninAppendixC.
Thissectionrelatestothemethodoftorquetightening,whichisthemostcommonmethodoftightening
threaded fasteners. Other methods intended for obtaining more precise preload during tightening are
presentedinSection 12.
861H
Thetorquetighteningmethodutilisestherelationshipbetweenappliedtorqueandpreloadinducedinthe
fastener. The relationship is dependent on both geometry and effective coefficients of friction at the
underheadbearingregionandinthethreads.
45
ECSSEHB3223A
16April2010
Thevalueof canbevariedinordertofinetunetheperformanceofthejointhoweveritshouldtypically
lieintherangefrom0.5to0.8.Therecommendedvaluetostartthedesignprocessis0.65.
Thenominalpreloadinthefastener,FV,canthenbecalculatedby,
FV y AS [6.2.2]
Inpractice,thepreloadachievedinservicevariesaroundthisnominalvaluedueeffectssuchasfriction,
inaccuraciesinappliedtorquemeasurement,embeddingandthermalloads.Theremainderofthissection
explainstheseeffectsandprovidestheoryforquantifyingthestatisticalscatterinthepreload.
The idealised model used to describe the system of forces in action during tightening assumes the
developedshapeofthenutthreadhelixasawedgedrivenbetweenthewedgeofthefastenersthread
andsurfaceoftheclampedmaterial.Theaxialresultforcetransmittedbytheengagedthreadisassumed
toactnormaltotheidealthreadsurfaceatthepitchradius(d2/2).Theforcespresentduringtighteningare
shownintheleftof Figure61,andthecorrespondingwedgemodelwithforcesareresolvedintoaplane
862H
isshownontherightofthefigure.
46
ECSSEHB3223A
16April2010
FH
FASTENER
FH
p
FV
Figure61ForcesPresentDuringTightening(theWedgeModel)
Fromthelowerdiagramin Figure61theappliedforceactingtoslipthewedge(analogoustorotating
863H
thethread)is,
FH FV tan( ) [6.3.1]
47
ECSSEHB3223A
16April2010
Thetorquepresentatthethreadinterfaceisthereforegivenby,
d2
M th FV tan( ) [6.3.3]
2
The underhead torque due to friction at the bearing surface between nut or fastener head and the
adjacentclampedpartisassumedtoactatthemeanbearingradiusofthehead(ornut).Therefore,itis
givenby,
uh d uh 1
M uh FV [6.3.4]
2 sin( 2)
where uh isthecoefficientoffrictionbetweenthenut(orfastenerhead)andtheadjacentclampedpart,
duhistheeffectivediameteroftheunderheadfrictiontorqueasdefinedbyequation[5.4.6],and isthe
underheadbearingangle(see Figure56).
864H
Thus,thetotalinstallationtorqueapplied(toeithertheheadornut)duringtighteninginordertoproduce
thepreloadFVis,
d d uh [6.3.5]
M app M th M uh M P FV 2 tan( ) uh MP
2 2 sin( 2)
MP is subject to scatter so, in addition to quoting mean values, an indication of its expected variation
should be given by either the range or coefficient of variation. Mp is likely to vary with fastener size.
Subsection 6.3.2 contains more information about locking devices and Table 61 provides some
866H 867H
experimentaldata.
Thegeometricrelation,
tan tan
tan( ) [6.3.6]
1 tan tan
withtheapproximation tan tan 1 ,gives,
tan( ) tan tan [6.3.7]
Hence,[5.2.5]canberewrittenas,
[6.3.8]
d th d uh
M app FV 2 tan uh MP
2 2 sin( 2)
cos
2
orusingthegeometricrelation
p [6.3.9]
tan
d uh
48
ECSSEHB3223A
16April2010
[6.3.10]
p th d uh
M app FV d 2 uh MP
2 sin( 2 )
cos
2
TheappliedtorqueEquation [6.3.10]canbeseparatedintothefollowingterms:
868H
p
FV representingthetorqueabsorbedinstretchingthefastener,
2
d 2 th
FV
representingthetorqueduetofrictioninthethreadedinterface,
2 cos
2
d
FV uh uh representingthetorqueduefrictionunderthenut(orhead),and
2 sin( 2)
MP representingtheprevailingtorqueduetothelockingdevice
Atypicaldistributionofthetorquebetweenthesecomponentsisasfollows:
10% Stretchingbolt
30% Absorbedbyfrictioninthreads
50% Absorbedbyfrictionunderheadofboltornut
10% Prevailingtorqueofaselflockingthread
ThefollowingsimplifiedexpressionforMappcanbeusedforthecommoncaseofUnifiedorMetricthreads
(where =30)andstandardheads(where =180),
d
M app FV 0.16 p 0.58 th d 2 uh uh M P [6.3.11]
2
In practice manufacturing variations in p, d2 and duh are negligible. Thus, for a geometrically similar
fastener system, Mapp depends primarily on variations in th and uh . If, for a given joint, the friction
coefficientsarefixed,thetorquetopreloadrelationshipislinear.
6.3.2.1 Overview
248B
Excessivefastenertorquecanleadtotensilefailureofthefastener.Ontheotherhand,insufficienttorque
can lead to inadequate compression of the clamped parts, leading to failure by gapping or to slipping
(whenslippingisnotallowed).
Thespecifiedtorquerangeisnormallyfoundbyaniterativedesignprocesswherebythemarginsofsafety
arecalculatedforeachrelevantfailuremode.
The maximum and minimum applied torques,Mapp,max and Mapp,min, should first be calculated taking into
account the torque measurement accuracy. The most common way to specify these torque levels is to
49
ECSSEHB3223A
16April2010
calculate them relative to a nominal applied torque, Mnom. Appendix A recommends starting values of
nominaltorque,Mapp,nom,forsomeofthemostcommonlyusedfasteners.
Twomethodsarepresentedbelowforcalculatingthepreloadrangebasedonthetorquerange:
a. ExperimentalCoefficientMethod,and
b. TypicalCoefficientMethod.
TheExperimentalCoefficientMethodshouldbeappliedwheneverthepreloadinthejointiscritical(e.g.
for friction grip joints) since it considers independently the uncertainties of the fastener friction
coefficientsandtorquewrenchaccuracy.
TheTypicalCoefficientMethodismostlyusedincaseswherecontrolofthepreloadisnoncritical.This
occursmostoftenforbearingjointsorlowdutyjoints.
Thismethodtakesintoaccountallsourcesofpreloaduncertainty,namelytheprevailingtorque,preload
lossduetoembedding,thermoelasticeffects,andthetorquewrenchaccuracy.
Firstly,themaximumandminimumboundsoftheappliedtorque,Mapp,minandMapp,minaredeterminedby,
where Mnom is the nominal applied torque (normally determined by an iterative process) and is the
torquewrenchaccuracy,whichistypicallyintherange5%to15%.
Thetorquewrenchaccuracyshouldbeguaranteedbycertificatesobtainedduringperiodiccalibrations.
Next, the maximum and minimum preloads are found by the following equations, which are derived
fromEquation [6.3.10],andmodifiedtoincludethermalandembeddingeffects:
M M P ,min
869H
FV ,max FT [6.3.14]
app , max
1 th ,min 1
d
d 2 tan
2 2 uh uh ,min
cos
2
( M app ,min M P ,max )
FV ,min FT FZ [6.3.15]
1 th ,max 1
d 2 tan d
2 2 uh uh ,max
cos
2
where FT and FT arethemaximum(mostpositive)andminimum(mostnegative)incrementsin
thefastenerloadduetothermoelasticeffects(seeSubsection 6.3.3),andFZisthemaximumpreloadloss
870H
ThetermMP,mininEquation [6.3.14]isrelativelyinsignificantand,asaconservativesimplification,isoften
872H
excludedfromthecalculation.
50
ECSSEHB3223A
16April2010
The minimum preload calculated by Equation [6.3.15], includes the preload loss due to relaxation, FZ.
873H
Thistermdoesnotappearinthemaximumpreloadequation [6.3.14]sinceembeddingcanonlyloosena
874H
fastener,andthereforecannotincreasethemaximumpreload.
Inthismethodanuncertaintyfactor isintroducedtoaccountforallsourcesofscatterobservedduring
thetighteningprocess.Thefrictioncoefficientsusedarenominal(oraverage)values.
(1 ) M app ,max
FV ,max FT [6.3.16]
1 th ,nom 1
d 2 tan d
2 2 uh uh ,nom
cos
2
(1 )( M app ,min M p ,max )
FV ,min FT FZ [6.3.17]
1 th ,nom 1
d 2 tan d
2 2 uh uh ,nom
cos
2
TheEquations [6.3.16]and [6.3.17]separatetheappliedtorqueintoitsminimumandmaximumvalues,
875H 876H
Mapp,maxandMapp,min.However,sincethepreloadscatterisdominatedbyfrictionvariability,inmanycases
itisequallyconservativetousethenominalappliedtorque,Mapp,nom,inbothoftheseequations.
Table61givestypicalvaluesof forarangeoftightening.
877H
Methodofpreloadapplication Uncertaintyfactor( )
Torquemeasurementofunlubricatedbolts 0.35
Torquemeasurementofcadmiumplatedbolts(*) 0.30
Torquemeasurementoflubricatedbolts 0.25
Hydraulicfasteners 0.15
Preloadindicatingwashers 0.10
Ultrasonicmeasurementdevices 0.10
Boltelongationmeasurement 0.05
Instrumentedbolts 0.05
Theuncertaintyfactorcanbereducedwhenthepreloadismeasuredbymoreaccuratemethodssuchas
thosediscussedinSection 12,orsometimesalsobyincreasingtheaccuracyofthetorquewrench.
879H
(*) This information is provided for reference only, since the use of cadmiumplated fasteners is not
allowed
51
ECSSEHB3223A
16April2010
6.3.3 Locking Devices and Prevailing Torque
125B
A joint in a spacecraft is exposed to extensive vibration loads during the launch phase of its mission.
Therefore, all joints with threaded fasteners should have locking devices to avoid vibration loosening.
Lockingcanberealisedbyeitherselflockingthreads(applicabletobothnutsandhelicoilsorinserts)or
byothermeanssuchasadhesivesorlockwire.
Selflockingthreaddevicescommonlyusedinspacecraftcanbeputinthreecategories:
Helicoil(screwlock)orinsert:Thelockingisachievedbyoneorseveraldeformedthreadslocatedin
thethreadedhole.
Selflockingnuts:Thenutscrosssectionisdeformedtoanellipticalshapeandhastobereshapedto
acirclewhenscrewingthroughit.
Selflocking anchor nuts: The locking mechanism is the same as selflocking nuts but the nut is
anchoredtotheflange.
Avarietyofotherselflockingthreaddevicesexistforterrestrialapplications;plasticthreadinserts,etc.
Beforethefastenercanrotateitisnecessarytoovercomeacertaintorque,namedtheprevailingtorque.
Thisismeasuredbyturningthethreadedpartswhilethereisneitherunderheadnorundernutcontact.
Thispracticeisrecommendedforeverysafetycriticalfastener,whichhasaselflockingdevice.
The prevailing torque of some commonly manufactured devices meeting LN norms are listed in Table 880H
62. If an adhesive is used to guarantee locking of the threaded parts, tests should be performed to
determine its prevailing torque. It is important to note that liquid adhesives are known to act as a
lubricant prior to curing, so a lower friction coefficient should be used when determining the preload
torque.Also,theuseofliquidadhesivesaslockingmeanisnotrecommendedsinceitseffectmaynotbe
verifiedduringtheinstallationofthefastener(theinferredlockingtorquemaynotbemeasured).
52
ECSSEHB3223A
16April2010
Table62PrevailingTorquesofLockingDevicesonUnlubricatedFasteners
FastenerType TypeofNutorInsert Size Mp,min Mp,max Mp,av
Cylinderhead Helicoil(screwlock)bronze M4 [2]
fastener, (notcadmiumplated), M5 1.10 1.60 1.28
LN29949 LN9499[1]
M6 2.60 4.60 3.36
M8 4.40 5.60 4.91
Anchornut, M4
LN29693 M5
M6 0.40 2.00 1.47
M8
Nut, M4
LN9161M M5 1.10 2.00 1.56
M6 0.50 1.00 0.66
1. LN 9499 is only applicable for CRES helicoils. DIN 65536-1 is the correct standard for bronze
helicoils.
2.Missingdatacanbeprovidedonrequest
Whenusing Table62someadditionalinformationshouldbeconsidered.Generally,helicoils(screwlock)
881H
asdefinedinLN29949showthelargeststatisticalspreadofselflockingtorquewiththemostandlargest
outlier values. These outliers are not taken into account in the table. Uncommonly high values for the
prevailingtorquenormallyindicateafailureofthehelicoil.
Table62isgivenasreferenceonly.Thereaderisaddressedtolookintothedatasheetspublishedbythe
882H
manufacturers of selflocking helicoils, inserts and nuts. They usually provide the minimum and
maximumvaluesoftheselflockingtorqueforeachthreadsize.Itisnecessarytomeasuretheactualvalue
oftheselflockingtorqueduringthefastenerinstallationprocess,andtocheckthatitiswithinthelimits
providedbythemanufacturer.
To avoid galling, it is recommended that the helicoils should be made from bronze CuSn6 material
number2.1020.34.
relationshipisthevariabilityinthefrictionconditions.
Variations in fastener geometry that influence the friction radii will also contribute to scatter but to a
lesserextent.Ofthese,variationsintheunderheadandundernutradiiaremostdominant.
Inaddition,variationintheappliedassemblytorquewillalsocontributetoscatter.
Prevailingtorquescatterwillalsocontributetothevariationintheeffectiveassemblytorqueseenbythe
fastener.Tominimisethiseffectthetorqueappliedtoeachfastenershouldbeadjustedbyits(measured)
53
ECSSEHB3223A
16April2010
prevailingtorque,whichismeasuredwhileturningthethreadedpartspriortocontactattheunderhead
orundernutinterfaces.
The effect of these factors on the possible scatter in the fastener preload is shown in Figure 62. This
884H
illustratesthedominanteffectoffrictionanddemonstratesthehighscatterpossibleiffrictionconditions
arenotcontrolledbythecarefulspecificationandapplicationoflubrication.
Figure62showsanexampleofthepreloadasafunctionoftheappliedtorque.Therelationshipislinear
885H
andtheslopedependsonthefrictioncoefficients.Thegraphshowscontoursofthefastenerstressratio,
v.m. / y ,whichindicatethestrengthutilisation.
Themaximumpreload,90%ofthefasteneryieldload,isfoundatthepointX,wherethemaximumthread
torqueMth,maxandminimumfriction min occur.TheminimumpreloadloadFV,minisfoundatthepointY
wheretheminimumthreadtorqueMth,minandmaximumfrictioncoefficient max occur.
min
X
max
Fastener
Preload
b
(FV) b , y 1 .0
b
Y b , y 0.9
Torque
Mth,max Muh,min
Mth,min Muh,max
Figure62TypicalPreloadvs.AppliedTorqueGraph
Regarding Figure62itshouldbenotedthattheappliedtorqueisdeterministic,i.e.thefrictionaltorques
886H
actingonthefasteneralwaysadduptotheexternallyappliedtoque,Mapp.Thismeansthattheextreme
preloadconditionsoccurunderthefollowingconditions:
Maximumpreloadminimumthreadfrictionandminimumunderheadfriction
Minimumpreloadmaximumthreadfrictionandmaximumunderheadfriction
The joint in Figure 62 has no prevailing torque device. This is evidenced by the fact that the torque is
887H
zerowhenthepreloadiszero.Includingaprevailingtorqueelementwouldshiftthecurvestotherightof
thisgraph.
54
ECSSEHB3223A
16April2010
6.3.5 Forces Induced by Thermal Fluctuation
127B
6.3.5.1 Overview
251B
During thermal fluctuations the preload changes due to differing thermal expansion coefficients of the
fastenerandflangematerial.Also,themodulusofelasticityofthematerialscanbetemperaturesensitive.
Thethermalexpansioncoefficientrelatestheelongation f T ofanindividualitemwiththelengthL,toa
temperaturedifference T accordingto,
f T L T [6.3.18]
f T ,b b L j T ,and [6.3.19]
f T ,c c L j T [6.3.20]
whereAsmisthestiffnessareaofthefastener, isthebasic(i.e.n=1)forceratioofthejointasdefinedin
eitherSection 7.5orSection 8(dependingonwhetherthejointisconcentricoreccentric).
888H 889H
If the clamped parts have multiple material types, a more general form of Equation [6.3.22] should be
890H
usedtoaccountforthedifferingthermalexpansioncoefficients,
m
f T ,c c ,i Lc ,i T [6.3.26]
i 1
55
ECSSEHB3223A
16April2010
where c ,i is the thermal expansion coefficient of the ith clamped part, Lc,i is the length of the ith
clampedpartandmisthetotalnumberofclampedparts.
Itcanbeshownthatthisequationleadstothefollowingequationforthethermallyinducedloadinthe
fastener,
m
c ,i Lc ,i b Lb [6.3.27]
FT i 1 T Eb Asm 1
Lj
6.4.1 128BOverview
Therearefivemechanismsthatcauserelaxationpreloadlossinboltedjoints:
a. Embedding
b. FastenerGroupInteractions
c. GasketCreep
d. MetallicCreep
e. TensionalRelaxation
The user should be aware of the possibilities of such preload reductions and estimations of their
magnitude should be included in the design calculations. It should be noted that vibration loosening,
whichcancauserelaxation,isoutsideofthescopeofthissectionandstressrelaxation,usuallyduetoa
combinationofstressandhightemperature,isnotapplicable.
56
ECSSEHB3223A
16April2010
Creep is a significant factor when the temperature exceeds approximately 50% of the recrystallisation
temperatureoftherespectivematerial.
In addition to the five mechanisms listed above, preload loss can also occur due to localised plastic
deformations in the fastener, if it is subjected to repeated loadings above material yield. This effect is
furtherexplainedinSection 7.13.
892H
6.4.2 Embedding
129B
6.4.2.1 Overview
254B
Whennewfastenersarefirsttightened,see Figure63,themaleandfemalethreads,theunderheadand
893H
undernutsurfacesandtheclampedpartsinterfacescontacteachotheronlyonmicroscopicallysmallhigh
spots,wherethesurfaceasperitiestouch,asillustratedin Figure64.Thematerialatthesehighspots
894H
willbeoverloaded,wellpasttheiryieldpoint,duringinitialtighteningandwillsubsequentlycreepuntil
alargeenoughareaoftheavailablecontactsurfacehasbeenengagedtostabilizetheprocess.
Figure63FastenerFayingSurfaces
Figure64MicroscopicViewofSurfaceAsperitiesInitiallyinContact
57
ECSSEHB3223A
16April2010
Inaddition,plasticflowwilloftenoccuratthehigheststressedpointssuchasinthreadroots,atthefirst
engagedthreadinthenut,oratotherstressconcentrationscausedbyimperfectmanufacture(e.g.errors
insquareness,flatness,concentricity,etc.).
These relatively shortterm relaxation effects are known as embedding. After tightening the rate of
relaxationisa maximum,reducing exponentially, usually over the first minutes, to a constant very low
rate of creep as illustrated in Figure 65. Typically embedding accounts for only a few percent loss of
895H
initial preload, however 5 to 10% preload loss is not uncommon, and maximums of 25% for torque
tighteningand40%forhydraulictensioninghavebeenrecorded(seeReference 6.11). 896H
100
90
80
Preload 70
Initial Preload 60
[%]
50
40
30
20
10
0
0 10 20
Time from assembly [days]
Figure65EmbeddingPreloadDecay
ThepreloadlossFZtobeputinEquation [6.3.15]dependsontheplasticdeformationfZofthejointcaused
897H
byembedding.Thefollowingrelationshipsapply,
FZ f fZ fZ
Z [6.4.1]
FV fV f b,V f c ,V ( b c ) FV
where; fb,V is the elongation of the fastener due to the preload and fc,V is the elongation of the flanges
under preload, and b and c are the compliances of the fasteners and the clamped parts respectively
(seeSubsections 0and 7.7).
898H 899H
Thusitfollowsthat,
fZ
FZ [6.4.2]
( b c )
ThevalueoffZdependsonthesurfaceroughness,thenumberofintersticesinthejointandthematerial
type.Forexample,usingaluminiumforthefemalethreadcandoublethevalueoffZcomparedwiththat
ofsteel.
Foruncriticalcasesavalueof5%ofthepreloadcanbeusedforcalculationpurposes,i.e.:
FZ 0.05 FV [6.4.3]
58
ECSSEHB3223A
16April2010
Otherwisetheappropriatepreloadlossshouldbedeterminedbyexperiment.Ifnoexperimentaldatacan
be gained, Table 63 can be used to find approximate values. This table may only be used if the
900H
temperaturesarebelow50%oftherecrystallisationtemperatures.
Table63Approximatevaluesforplasticdeformationcausedbyembedding
Average Surface Type of StandardValuesforEmbedding[m]
Roughness,RZ External Load
Inthethread Per under Per inter
Applied to
(asperRef. 7.3) 901H head or flange
Joint
undernut interstice
interface
<10m Axial 3 2.5 1.5
Shear 3 3 2
10m<x<40m Axial 3 3 2
The asperities of surfaces subjected to frictional sliding during assembly, such as during conventional
torque tightening, tend to shear and flatten thereby producing more embedding during assembly and
leaving less to take place after tightening. It is therefore recommended that tightening methods that do
not induce frictional sliding (e.g. hydraulic tensioning) be avoided unless galling (excessive friction
resultinginlocalizedwelding)orauniformpulldownareoverridingconsiderations.
6.4.2.3 Retightening
256B
Themagnitudeofembeddingpreloadlosscanbegreatlyreducedincriticalcasesbytorquetightening,
loosening and retightening several times. However, attention should be given to the reduction of self
lockingtorqueduetomultipletighteningcycles(seeSubsection 12.6.3).
902H
Itisalsoanadvantage,wherepossible,touseflangesmadefromhardermaterialsorwashersthatresist
embedding.
To minimisepreload loss in the thread, bolts with the tightest tolerance classof thread that is available
shouldbeused.ForexampleClass3ispreferredtoClass2asarerootradiuscontrolledJclassthreadsto
standardthreads,androlledthreadsshouldbeusedinpreferencetocutthreads.Theuseoffastenerswith
cutmachinedthreadsissubjecttoapproval.
59
ECSSEHB3223A
16April2010
6.4.2.6 Conical Components
259B
Conicaljointmatingsurfacesasillustratedin Figure66shouldbeavoided.Forthesejointconfigurations
903H
relaxationperpendiculartotheconicalsurface(duetoembedding)resultsisafargreaterrelaxationinthe
directionofthefasteneraxis.
Figure66JointwithConicalMatingSurfaces
Long thin compliant fasteners should be used in preference to short, stubby stiff fasteners. The total
amount of embedding will be the same for both for a given preload but for the compliant fastener the
stiffness is less and hence a smaller preload loss occurs (see Figure 67). Similarly Belleville washers
904H
(conicalspringwashersasshownin Figure68)canbeusedtoincreasethecomplianceoftheclamped
905H
partsandhencereducetheeffectofembedding.
60
ECSSEHB3223A
16April2010
Figure67EffectofFastenerStiffnessonPreloadLossDuetoEmbedding
Figure68IncreasingJointComplianceUsingBellevilleWashers
Relaxation will occur by major material plastic flow if the bearing surfaces are not large enough or the
appliedstressesarenotdistributedproperly.
Itisimportanttoensureinthedesign,manufactureandqualitycontrolcheckingthat:
61
ECSSEHB3223A
16April2010
Theunderheadradiusdoesnotfoultheedgeoftheclampedpartschamfer,whichcanresultina
verysmallcontactarealeadingtoexaggeratedembeddingwithacorrespondingpreloadloss.
Conversely,thechamferisnotsogreatthattoosmallanunderheadbearingarearesultsandthe
boltheadembedsintotheclampedmaterial.
slottedholes,whichcangreatlyreducethebearingareaundertheheadornut,areavoided.
The minimum acceptable thread engagement of high strength steel or titanium fastener is 0.8 times the
nominaldiameterofthefastener.Atleast1.5dthreadengagementshouldbeusedforsuchfastenerused
inaluminiumtappedholes,evenifwirethreadinserts(e.g.helicoils)areused.
Some lubricants can gradually migrate from the fastener and joint contacting surfaces causing a small
amountofrelaxationpreloadloss.
Fromtheinitialdiscussionofthissectionitisevidentthatifsurfaceroughnessisminimisedbygrinding
orpolishingtherelaxationeffectsofembeddingcanbereduced,butnoteliminated.
Anotherobviouspointisthatthetotalamountofembeddingisrelatedtothenumberofjointinterfaces.
6.4.3.1 Overview
265B
Someshorttermrelaxationproblemsevolveduetointeractionsbetweenfastenersingroups.
Whenthefirstfastenerinagroupistightened,itiselongatedandthejointinitsvicinityiscompressed.
When an adjacent fastener is tightened, the joint in the vicinity of the first fastener can be further
compressed causing relaxation of that fastener. The amount of interaction depends on such factors as
clampedpartsstiffness,thespacingoffastenersandwhetheragasketisused.Preloadlossesintheorder
of50%to100%havebeenrecorded(seeReference 6.3).Anexampleofthisproblemisgivenin Figure69
906H 907H
thesawtootheffectofonetighteningpassasthelastfastenertightenedinaparticularquadrantendup
with a greater preload than those immediately adjacent to it. However, after four passes Figure 610
910H
illustratesthatthefinalrelaxationpatternisnotaregularsawtooth.Ateachpassthejointistightenedby
successivelyhigher,butuniform,torquestothefasteners.
Thisproblemcanbepartlyovercomebygivinghigherinitialpreloadstothosefastenersexpectedtorelax
themost,thusproducingamoreuniformfastenerloadafterrelaxation.Thehighestaccuracyisachieved
bymonitoringpreloadtests,usingatechniquesuchasultrasonics,andapplyingsmallamountsofextra
torquetothefastenerswithinsufficientpreload.
62
ECSSEHB3223A
16April2010
6.4.3.3 Bent Flanges
267B
If flanges are soft, warped or bent, tightening one fastener can cause relaxation (or additional stress) in
otherfasteners.Aprecautionaryfinaltorquepassisalwaysmadewhenassemblingjointswithmultiple
fasteners.
Figure69InitialandFinalPreloadofaTypicalFlangeJointfortheFirstTightening
Pass.
Figure610InitialandFinalPreloadLevelsafterFourTighteningPasses.
63
ECSSEHB3223A
16April2010
6.4.3.4 Motorized Tightening Speed and Interaction Effects
268B
If flanges with a large number of fasteners are tightened simultaneously using a motorized technique
thereisagreatdangerofhighpreloadlossduetoembedding.Afastenertightenedalonewillseehigher
stress than if tightened in a group, where all the bolts share the load evenly as it develops. Also this
tightening process is usually carried out at relatively high speeds. It is therefore recommended that the
torqueisappliedinaseriesofincrementsandthattimeisallowedbetweeneachpassforembeddingto
takeplace.Thisrecommendationisparticularlyrelevantwheregasketsareused.
6.4.4 Gaskets
131B
Outgassingconstraintsrestricttheselectionofgasketmaterialsthataresuitableforspaceapplications.It
is usually preferred, where possible, to eliminate gaskets and rely on metaltometal sealing of flanges.
Wherethiscannotbeachievedapreformedgasketfromfullycuredsiliconerubberorasoftmetalsuchas
leadorasoftaluminiumalloyisusedtoformtheseal.
Gasketcreepisanotherformofrelativelyshorttermrelaxation.Havingagasketinajointensureslarge
amountsofrelaxationcomparedtothatexpectedwithnormalembedding.Infact,agasketisdeliberately
designed to deform plastically to some extent in order to plug the paths through which the contained
fluidorgaswouldotherwiseescape.Likeanyplasticmaterialitwillcreepwhenfirstsubjectedtohigh
surface pressures. The amount of creep is difficult to predict since it is dependent on many factors
includingthelevelofcompressiveloadgeneratedbythefasteners.
Thefollowingrecommendationsaregivenfordesigngasketjoints:
Toobtainthemostuniformpulldown,tightenflangeboltsinacrossstarpatternor,ifthecostis
justified,tightenseveralboltssimultaneously.
Usethemostaccuratemethodoftighteningavailable.
Usethehighestdesignaimpreloadpossiblewithoutdamagingthegasketorjointmember.
Ifdesigningpressurevesselsorpipeconnectionsfollowtherelevantdesigncodesofpracticessuch
asASMEVIIIdivision1andBS5500.
The nonlinear characteristics of gaskets can significantly influence the clamped parts compliance.
Subsection 7.6.5containsanexplanationofthecompliancecalculationsforjointswithgaskets.
911H
Very high temperatures, which could cause relaxation due to a reduction of material strength, do not
normallyoccurinspaceapplications.However,differentialthermalexpansionmaydirectlygivepreload
loss and can also cause relaxation due to over stressing. Such thermoelastic effects are discussed in
Subsection 6.3.5.
912H
6.4.6 Washers
133B
Washers are used for a variety of reasons. Plain washers are used to spread the load into the clamped
parts,toallowtheuseofoversizedholes,topreventthefastenerheadornutdamagingthesurfaceofthe
clampedparts,ortogivecontrolledsurfaceconditionsattheunderheadorundernutinterfaces.Many
othertypesofspecialpurposewashersexistsuchasinsulatingwashersforthermalorelectricalreasons,
64
ECSSEHB3223A
16April2010
conicalspringwasherstominimisetheeffectofembedding,thedimpledortwopiececrushablewashers
forpreloaddetection,thestraingaugedfacewasherforuseasaminiatureloadcell,shakeproofwashersto
preventrotarylooseningandtoothedshearwashersforslipprevention,etc.Itisimportanttoconsiderthe
complianceandrelaxationpropertiesoftheparticularwashertype.
The analysis of clamped parts compliance (Section 7.7) should take into account compliance of the
913H
washers.Generally,thewashercompliancecanbecalculatedassumingthatitbehaveslikeaflange.Ifthe
washerissoft,theanalysiscanbedoneusingthenonlineartheoryforgaskets(seeSubsection6.6.4).
Whenever washers are used, embedding (Subsection 6.4.2) is increased since there is at least one more
914H
intersticeregionwithcrushablesurfaceasperities.Forthisreason,itisrecommendedtousehardwashers
wheneverpossibletominimisetheembeddinglosses.
Duringthetorquetighteningprocesstorsionalstressisdevelopedinthefastenerinadditiontotheaxial
stress,whichisnecessaryforthepreload.Thetorsionalstressissubjecttovaryingamountsofrelaxation,
and it is generally believed to be the first stress to be relieved by relaxation. In most instances it
disappearsimmediatelywhenthewrenchisremoved.Ithasbeenshownthattherelaxationoftorsional
stresscaninduceextratensilestress(seeReference 6.4).Thisselftighteningphenomenon,illustratedin
915H
Figure 611 is invariably masked by the greater axial embedding relaxation (see Reference 6.3). The
916H 917H
amount of torsional stress primarily depends on the magnitude of the thread friction torque and the
amount of relaxation. Whether relaxation is by head rotation or selftightening depends largely on the
underheadfrictionconditionscomparedwiththoseofthethreadfrictionconditions.
Figure611SelfTighteningbyTorsionalRelaxation
65
ECSSEHB3223A
16April2010
When specifying the torque tightening method for fastening, the ability of the fastener to sustain the
torquingshouldbechecked.Thestressoccurringduringthetighteningprocessiscomposedbyatension
caused by the increasing preload and a shear stress caused by the application of the torque. These two
stressesarecombinedbytheshapedeformationenergyhypothesisaccordingtotheVonMisesequation,
v.m. V ,max 2 3 max 2 [6.5.1]
recommendedtomeasure(andcompensatefor)theprevailingtorqueeachtimeafasteneristightened.
Aftertheshearstressinthefastenerexceedsitsyieldpointastateoftotalplasticdeformationisreached,
whereby the torsional stresses through the cross section are constant. To include this effect, the polar
sectionmodulusiscorrectedto,
d03 [6.5.5]
Wp
12
ForEquation[5.5.1]themaximumpretensionstressisgivenby,
FV ,max FT
V ,max [6.5.6]
A0
TheMarginsofSafetyontighteningare:
y
Yield: MoS ti , y 1 [6.5.7]
v. m.
66
ECSSEHB3223A
16April2010
ult
Ultimate: MoS ti ,ult 1 [6.5.8]
v .m.
No factors of safety are used in Equations [6.5.7] and [6.5.8] since all uncertainty sources are already
919H 920H
includedinFV,max.
IftheMoSforyieldispositive,theMoSfortheultimateisalwayspositive.Therefore,theultimateMoSin
Equation [6.5.8] is not normally needed. One possible exception to this is yield controlled tightening
921H
methods(seeSection 12.2).
922H
involvepreloadcalculations.
6.7 References
37B
6.1
924H NASANSTS08307,July6, SpaceShuttleCriteriaforPreloadedBolts
1988
6.2
925H DIN4768:1990 Determination of values of surface roughness parameters RA, RZ,
RMAX using Electrical Contact (Stylus) Instruments Concepts and
MeasuringConditions.
6.3
926H J.H.Bickford An Introduction to the Design and Behaviour of Bolted Joints,
MarcelDekker,1981.
6.4
927H FISHER.J,STRUIK.J.H.A. GuidetoDesignCriteriaforBoltedandRivetedJoints,JohnWiley&
Sons,1974.
67
ECSSEHB3223A
16April2010
7
Concentric Axially Loaded Joints
8B
7.1 Overview
38B
This section describes the recommended analysis procedure joints in which the threaded fastener/s are
subjectedtoapureaxialload.Theanalysismethodspresentedinthissectionareappropriateforjointsif
allthefollowingcriteriaaremet:
Theaxisoftheappliedloadspassesthroughthefasteneraxis(orfastenergroupcentroid)
Theflangesaresymmetricabouttheaxisoftheappliedload.
Theflangesarestiffenoughtonotdevelopapryingeffect(seeSubsection 8.1.2) 928H
Ifanyoftheabovecriteriaarenotmet,themoregeneralanalysismethodsforeccentricallyaxiallyloaded
jointspresentedinSection 8shouldbeapplied.
929H
Considerthejointillustratedin Figure71.Thiscomprisesafastener,nutandtwoflanges.Intheabsence
930H
of externally applied loads the fastener preload is equal and opposite to the clamping load (the
compressionforceactingbetweentheclampedparts).Duringservicethejointsdimensionswillvarydue
to the applied load and the preload. For a given combination of these loads, the joints instantaneous
length,Lj,isrelatedtoitsconstituentpartsaccordingto,
Lb , free Lb L j [7.2.1]
Lc , free Lc L j [7.2.2]
where; Lb and Lc are the extension of the fastener and the compression of the clamped parts
respectivelyundertheactionoftheappliedloads,andLb,free andLc,freearethefree(i.e.unloaded)lengthsof
thefastenerandclampedpartsrespectively.
For a preloaded joint the free length of the fastener is less than that of the flanges since the fastener
stretcheswhiletheflangescompress.
Combiningandrearranging [7.2.1]and [7.2.2]gives,
931H 932H
68
ECSSEHB3223A
16April2010
Lj
Figure71TypicalJointComponents
whichleadstoasimplerelationbetweentheelongationsofthefastenerandclampedparts,
Lc Lb [7.2.5]
7.3.1 Overview
135B
Thedeformationswithinajointcanbeshowninthejointdiagram. Figure72showstheelongationofthe
933H
parts.
Lc ,V
Lb,V
Figure72FastenerStiffness Figure73ClampedPartsStiffness
Inthejointdiagramclassicalanalysisthestiffnessplotsforthetwocomponents(fastenerandflanges)are
combined by inverting Figure 73 (in the horizontal axis) and moving its origin such that the preload
935H
Figure74(6.2.3a),b)&c))showthepreloadFVincreasingduringthetighteningprocess.
937H
69
ECSSEHB3223A
16April2010
Figure 6.2.3a
Figure 6.2.3b
Figure 6.2.3c
Figure74GrowthoftheJointDiagramIllustratingtheTighteningProcess
70
ECSSEHB3223A
16April2010
Applyinganexternalaxialload,FA,ontheoutersurfacesofthejointsflanges,asshownin Figure75,
938H
increasesthefastenerstensionandreducestheclampedpartscompressiveloading.Thisisreflectedby
the joint diagram in Figure 76 with the external force FA represented by a vertical line. The load
939H
incrementsexperiencedbythefastenerandclampedpartswhenthejointissubjectedtoFAaregivenby
Fb , A and Fc , A respectively.
F b , A
F b , A
F b , A
Figure75Applicationofexternalaxialload
Lb , A
Fb , A
Lb , A Lc , A
Fc , A
Fb
Lc , A
Figure76JointDiagramShowingtheEffectofanExternalAxialLoad
71
ECSSEHB3223A
16April2010
Figure76 shows that ifFA isincreased beyond a certain point,FK =0. This corresponds to the point at
940H
whichtheclampedpartsarecompletelyrelievedoftheirpreloadandgappingattheinterfacebetweenthe
flangesoccurs,asillustratedin Figure77.
941H
Figure77ExternalLoadCausingGapping
Iftheexternallyappliedaxialloadonapreloadedjointiscompressive,theresultingcompressionofthe
flangesandthereductioninbolttensioncanberepresentedonajointdiagramasshownin Figure78. 942H
All the equations that were derived for externally applied tensile loads are directly applicable for
compressiveloads,subjecttothecorrectuseofnegativesigns,(seeReference 7.3). 943H
Fc , A
Fb , A
FV
Figure78JointDiagramforCompressiveLoading
72
ECSSEHB3223A
16April2010
action is at the outer surface of the flanges. This is the situation that is shown in Figure 75. For such
945H
joints,theforceratiocanbeshowntodependontherelativestiffnesses(orcompliancees)ofthefastener
andtheclampedparts,
Fb , A Kb c
[7.4.2]
FA Kc Kb c b
where Kb & Kc are the respective stiffnesses of the fastener and clamped parts, and b & c are the
respectivecompliancesofthefastenerandclampedparts.
From Figure76itcanbeseenthat,
946H
FA Fb , A Fc , A [7.4.3]
therefore,usingEquation [7.4.2], 947H
FA FA Fc , A [7.4.4]
Rearranging,theloadreductionfeltbytheclampedpartsduetotheexternalloadFAisgivenby,
Fc , A (1 ) FA [7.4.5]
Equations [7.4.2], [7.4.3], [7.4.4]and [7.4.5]areonlyvalidforjointswherethepointofapplicationofthe
948H 949H 950H 951H
externalloadistheoutersurfaceoftheflanges.Ingeneral,theexternalloadiseffectivelyappliedatpoints
located within the flanges. Therefore, the loading plane factor, n, is introduced into equation [7.4.2] 952H
resultingin,
c [7.4.6]
n n n
c b
Methods are presented in the following Sections for calculating the various parameters in Equation
[7.4.6].
953H
1
Fb K b Lb Lb [7.5.1]
b
where; Fb is the tensile load in the fastener shank, Kb is its stiffness, Lb is its elongation and b is its
compliance.
73
ECSSEHB3223A
16April2010
Inthegeneralcaseofafastenerwithvaryingshankdiameters,suchas Figure79,thetotalelongationof
954H
thefasteneristhesumoftheelongationsofitsconstituentsegments(i.e.head,nut,andshank).Ingeneral,
thedifferentpartsofthefastenerhavedifferentlocalstiffnessproperties.Thusthetotalelongationcanbe
written,
Lb Li [7.5.2]
where Li aretheelongationsoftheconstituentsegments.
Lj
L1 L2 L3
Ln,sub
Lh,sub Leng,sub
Figure79DimensioningoftheFastenerforComplianceCalculations
ApplyingHookeslawtoeachsegment,
Li Fb
Li [7.5.3]
E b Ai
where;EbistheYoungsmodulusofthefastenermaterialandAiisthelocalcrosssectionofthesegment.
Combiningtheaboveequations,thecomplianceofthefasteneristhen,
1 Lb 1 Li
b
Kb
Fb
Eb
A [7.5.4]
i
Expanding [7.5.4] and introducing substitution lengths for deformations in the head, the fasteners
955H
engagedregionandthenutsengagedregiongives,
1 Lh , sub Leng , sub Lsha ,1 Lsha , 2 Lsha ,i L
b n , sub
[7.5.5]
Eb Anom A3 Asha ,1 Asha , 2 Asha ,i E n Anom
where;Lh,sub,Leng,subandLn,subarethesubstitutionlengthsfordeformationswithinthehead,engagedshank
and engaged nut or insert (see Table 71), Anom is the fasteners nominal crosssectional area, A3 is the
956H
fastenersminordiameterarea(see Figure55),Asha,iistheeffectivecrosssectionalareaattheithsegment
957H
ofthefastenersshankandLsha,iisthelengthoftheithsegmentofthefastenersshank.
74
ECSSEHB3223A
16April2010
InmanycaseswhenusingEquation[7.5.5]onlyonesegmentoffastenersshankneedstobeconsidered.
However, if the fastener has varying shank diameters, the length and effective stiffness area of each
segmentshouldbeincluded.
Table 71 shows typical substitution lengths for standardised fasteners, which can be used in Equation
958H
[7.5.5]. The table only includes values for typical hexagon heads, cylindrical heads and nuts. The same
959H
analysismethodcanbeappliedtootherfastenerheadsandnuts,andtheappropriatesubstitutionlengths
shouldbedeterminedbyexperimentoranalysis.
Table71TypicalSubstitutionLengthsforCommonlyUsedFasteners
7.6.1 Overview
137B
Thecalculationofthecomplianceoftheclampedparts, c ,ismorecomplicatedthanthatofthefastener
due to the 3dimensional state of stress state in the joint that is induced by the preload. The method
presentedhere(basedonReference 7.3)neglectsthecomplianceoftheinterstices,andthereforeitismost
961H
accurateforjointswithasmallnumberofclampedparts.Ifneeded(e.g.whenmanyplatesareclamped
together), the compliance of the clamped parts can be determined by experiment or finite element
analysis.
Thecomplianceofthecampedpartsiscalculatedbytheintegral,
z Lc
1 dz
[7.6.1]
c
Ec z 0
A( z )
where; A(z) is the crosssectional area of the assumed compression zone within the clamped parts
(explainedinSubsections 7.6.2to 7.6.4),zisthedistancethroughtheclampedparts,Lcisthelengthofthe
962H 963H
clampedparts,andEcisthemodulusofelasticityoftheclampedparts.
Whenmultiplematerialsareusedintheclampedparts,themoregeneralformofEquation [7.6.1]should 964H
beused,
75
ECSSEHB3223A
16April2010
z Lc
dz
[7.6.2]
c
z 0
E c ( z ) A( z )
7.6.2.1 Overview
269B
Theconfigurationofthecompressionzonedependsonthegeometryoftheclampedparts.Thefollowing
paragraphs explain the possible configurations. Equations for calculating the compliance of the various
configurationsaregiveninSubsection 7.6.4. 965H
Figure 710 shows the three possible compression zones (hatched) for cylindrical clamped parts. For
966H
clarity, the figure assumes that the flanges are compressed between infinitely stiff washers with under
head bearing diameter Duh,brg. When the flanges are wider than the underhead bearing diameter, the
compression zone spreads out. For the theory presented here, it is assumed that the compression zone
spreads within a conical surface. In many cases the compliance of the washers is also significant. If so,
theyshouldalsobetreatedasclampedparts.
Duh,brg
Lc
Dlim
Davail Davail Davail
Figure710Compressionzonesincylindricalclampedparts
Theconfigurationontheleftof Figure710hasflangesthataresufficientlywidetoallowfullspreading
967H
ofthecompressionzonetothelimitingdiameter,Dlim.Inreality,theshapeofthe3dimensionalzoneof
compressioninanisotropicmaterialisaparaboloid.However,themethodapproximatesthecompression
zone as a pair of compression cones. The two cones are symmetric about the midpoint of the clamped
partslength,Lc,(whichdoesnotnecessarilycorrespondtotheintersticebetweentheflanges).
76
ECSSEHB3223A
16April2010
7.6.2.4 Compression Sleeve
272B
Themiddleconfigurationof Figure710hasnarrowflangesthatfitentirelywithintheunderheadbearing
968H
diameter, i.e. Davail < Duh,brg. In this case, the compression zone is a sleeve (a uniform compression
cylinder).
The configuration to the right of Figure 710 has flanges of intermediate edge clearances so a partial
969H
length sleeve of diameter Davail approximates the compression zone, and two deformation cones are
formedaboveandbelowthesleeve.
In the general case with clamped partsthat are not axiallysymmetric about the fasteners axis multiple
edgedistancesarepresent(see Figure711).Forthistypeofjoint,theconfigurationofthedeformation
970H
zoneshouldbedeterminedbytheminimumedgedistanceaccordingto,
al,min=Davail,min/2. [7.6.3]
Equation [7.6.3] results in a value of clamped parts compliance that is higher than reality, which may
971H
introducesignificant(nonconservative)inaccuraciesintheMoS.Paragraph 7.6.2.9discussesthisissuein
972H
moredetail.
al,min
Davail,min
Figure711Thecompressionzonewhenmultipleedgedistancesarepresent
Unlikenuttightenedjoints,threadedholejointsdeveloponlyonecompressioncone.Thisconedevelops
fullyiftheflangehasanavailablediameterlargeenoughtoallowtheconetoextendtotheintersticeof
thethreadedholeflange(seeleftof Figure712). 973H
77
ECSSEHB3223A
16April2010
Whenthecompressedflangehasinsufficientdiameter,thecompressionconestopsattheflangeedgeand
a compression sleeve forms between the base of the compression cone and the interstice of the insert
flange(seerightof Figure712).
974H
The threaded hole flange does not contribute to the clamped parts compliance, although it should be
includedinthecalculationofthefastenercomplianceaccordingto Table71.
975H
Dlim
Lc
Figure712ApproximationoftheCompressionZoneforInsertJoints
Whenajointhasmultiplefasteners,theavailablediameterofthedeformationzonearoundeachfastener
isaffectedbytheinterferenceeffectwithitsneighbour/s(see Figure713).Theresultofthisinterferenceis
976H
adecreasedcompliance,relativetoasimilarjointwithoutoverlappingcompressionzones.Thisleadstoa
conservative estimate of the fasteners tensile loading, but a nonconservative estimate of the flange
compression(seeParagraph 7.6.2.9below).
977H
Whencalculatingmarginsofsafetythatdependonflangecompression(e.g.gapping,andslipping),itis
conservative to neglect the overlapping compression zones when calculating the compliance of the
clampedparts.
78
ECSSEHB3223A
16April2010
Figure713Thecompressionzonewithmultipleinteractingfasteners
Sincethemethodspresentedinthissectionforcalculatingthecomplianceoftheclampedpartsaresubject
to some approximations, it is worthwhile examining the effects of overestimating and underestimating
thiscompliance.
Referring to Equation [7.4.6], overestimating the clamped parts compliance gives rise to an artificially
978H
high force ratio, . This leads to an overestimate of the force increment experienced by the fastener,
whichisthereforeconservativeforthecalculationoffastenertensilefailuremodessuchasfastenershank
failureorthreadpullout.However,italsoleadstoanunderestimateforforcetheincrementexperienced
by the clamped parts, which gives nonconservative for the flange performances related to separation,
sealing,frictiongrip,crushing,etc.
Intheoppositecaseoftheclampedpartscompliancebeingunderestimated,theforceratioisartificially
low.Thisleadstononconservativevaluesforthefastenerandthreadstrengths.
Accordingly,it is recommended that more detailedanalysis or testing is undertaken in cases where the
clampedpartscompliancemightbeacriticalfactorinthejustificationofthejointsdesign.
7.6.3.1 Overview
278B
The appropriate configuration of the compression zone should be determined for each fastener in the
joint. The following paragraphs present the equations necessary to determine which compression zone
configurationisapplicable.
First,theedgedistanceoftheflanges,al,isusedtocalculatetheavailablediameterforthecompression
zone,
Davail=2al [7.6.4]
79
ECSSEHB3223A
16April2010
Then,thisiscomparedwiththelimitdiameterofthecompressionconeasgivenby,
D lim Duh ,brg w Lc tan [7.6.5]
where; Duh,brg is the fasteners underhead bearing diameter, Lc is the compression length, and is the
compressionconehalfangle(seeParagraph 7.6.3.3below),andw=1foranuttightenedjointandw=2for
979H
athreadedholejoint,.
The compression cones half angle, , depends strongly on the available area of the flanges. It is also
influenced by the fasteners underhead diameter, the hole diameter and the distribution of the
compressionload.Reference 7.3provesempiricallyderivedequationsforcalculating .
980H
Fornuttightenedjointstan( )isgivenby,
tan 0.362 0.032 ln x 2 0.153 ln( y ) [7.6.6]
andforinsertjointsitisgivenby,
tan 1.295 0.246 ln( x ) 0.94 ln( y ) [7.6.7]
wherethefollowingnondimensionalparametersareused,
Lc
x [7.6.8]
Duh ,brg
Davail
y [7.6.9]
Duh ,brg
Theexistenceofthesleeveisdeterminedasfollows:
If Davail > Dlim, the compression zone is fully developed into a cone (or pair of cones for nuttightened
joints)
IfDuh,brg>Davail,nocompressionconedevelopsandonlyasleevehastobeusedforcalculationpurposes
IfDuh,brg<Davail<Dlim,thatcompressionzoneisformedbyapartialcompressionsleeveandcompression
cone/s
Theabovecases,1to3,correspondtotheleft,middleandrightconfigurationsof Figure710.
981H
The compliance of the clamped parts, c , for the case of fully developed compression cone/s (i.e. with
Davail>Dlim)isgivenby,
( Duh ,brg d )( Dlim d ) [7.6.10]
2 ln
( Duh ,brg d )( Dlim d )
c
w E c d tan( )
whereEcisthemodulusofelasticityoftheclampedparts.
80
ECSSEHB3223A
16April2010
Forpartiallydevelopedcompressioncone/s(i.e.Duh,brg<Davail<Dlim),theclampedpartscomplianceis
givenby,
2 ( Duh ,brg d )( Davail d ) 4 Davail Duh ,brg [7.6.11]
ln L
w d tan( ) ( Duh ,brg d )( Davail d ) Davail 2 d 2 w tan( )
c
c
Ec
Equation [7.6.11]includestheeffectofboththecompressioncone/sandsleeves,andisappropriatewhen
982H
theclampedpartshavethesamemodulusofelasticity.
If the compliances of the cone and sleeve should be calculated separately (for example when multiple
materialsarepresentintheclampedparts),theclampedpartscomplianceshouldbecalculatedby,
2
c cone sleeve [7.6.12]
w
wheretheconescomplianceis,
( Duh ,brg d )( Duh ,brg 2 Lcone tan( ) d ) [7.6.13]
ln
( Duh ,brg d )( Duh ,brg 2 Lcone tan( ) d )
cone
E c d tan( )
andthesleevescomplianceis,
4 Lsleeve
sleeve [7.6.14]
E c ( Davail d 2 )
2
andthelengthsoftheconeandsleevearegivenby,
Davail d
Lcone [7.6.15]
2 tan( )
and,
2 Lcone [7.6.16]
Lsleeve Lc .
w
Forthecasewhereonlyacompressionsleeveexists(Duh,brg<Davail),Equation [7.6.16]isusedforthesleeve
983H
complianceandLsleeveisequaltothecombinedlengthoftheclampedparts,Lc.,andw=1foranuttightened
jointorw=2forathreadedholejoint.
Iftheclampedpartsincludeagasket,itispossibletoestimatethetotalcomplianceofallclampedparts
using,
c e g [7.6.17]
714) and the compliance of the gasket is much greater than that of the other clamped parts. Therefore,
81
ECSSEHB3223A
16April2010
deformationofthegasketoftendominatestheelasticbehaviourofthejoint.Notonlydoes g varywith
theappliedload,butinadditionthegasketbehaviourisoftenpartiallyplastic(see Figure715),leading
985H
to a hysteresis effect. If the load is dynamic, the hysteresis can disappear over a period of time, and
thereforegasketjointsshouldbeanalyseddifferentlydependingonwhethertheyareintendedforstatic
ordynamicservice.
Figure714TypicalGasket Figure715GasketShowingHysteresis
Deflection
Thecoefficientofthermalexpansionofagasketmaybedifferentfromthatoftheotherclampedpartsor
thefastenerhencetemperaturechangescanaffectjointperformance.
Includinganelasticplasticnonlinearspringappliedinthejointdiagramresultsinajointnonlinearjoint
diagramsuchasshownin Figure716.Inthisfiguretheamountofcreepisindicatedbythedifference
986H
betweenthedottedandfulllines.
Figure716JointDiagramforaGasketJointShowingGasketCreep
Thedifficultiesinpredictinggasketcreepandspringrateshasledtoanempiricalapproach,whichuses
experimentallyderivedgasketfactors,yandmforeachtypeofgasketmaterial:
82
ECSSEHB3223A
16April2010
Theyfactorgivestheinitialgasketstresswhichcanbeconvertedtoaminimuminterfaceclamping
load, FK,req, assuming uniform gasket loading. A typical value for a soft aluminium gasket is 60
N/mmandsiliconrubber3N/mm.
Themfactorisashapefactorexperimentallyderivedforeachgaskettype.Itistheratioofgasket
pressuretocontainedpressureforaneffectiveseal.Typicalvaluesforsoftaluminiumandsilicon
rubberare3.25and1.25respectively.
Subsection 6.4.5 contains some recommendations for designing joints in order to minimise unwanted
987H
complianceeffectsduetogaskets.
7.7.1 Introduction
142B
7.7.1.1 Overview
282B
These equations in the section are only applicable for concentric axially loaded joints. Section 8.2.4
988H
providesanextensionofthismethodforthecaseofeccentricaxiallyloadedjoints.
For most joints, the effective loading planes are considered to be within the joint and separated be a
distanceof n L j ,asshownin Figure717.
989H
n Lc Fb
Lc
Fb Fc = FA + FK
Fc = Fb
Figure717TensionJointLoadingPlanesandtheForcesActingwithintheJoint
Between the loading planes the joints internal compressive stresses are relieved by the effect of the
externalload,FA.Outsidetheloadingplanestheclampedmaterialissubjectedtoextracompressioneffect
ofFAinadditiontotheclampingload,FK.
83
ECSSEHB3223A
16April2010
The effect of the loading plane factor on the joint diagram is shown in Figure 718. As n reduces, the
990H
effectivefastenerstiffnessreducesandtheeffectivejointstiffnessincreases(seeReference 7.4).Hence,the
991H
closertheloadingplanes,thesmallerthefastenerloadincrementandthelargerthepreloadreliefinthe
clampedparts.
Fb , A
Fc , A
Fb , A n FA
Fc , A 1 n FA
Figure718JointDiagramforLoadingPlaneswithintheJoint(n<1)
The loading plane factor depends on the deformation of the joint caused by the preload. For uncritical
verification purposes with simple joint geometries the loading plane factor may be set to 0.5, which
assumesthattheloadingplanesareatthecentreofeachflange.
Amoreprecisevaluefortheloadingplanefactorcanbedeterminedfromtestingbyuseoftherelation,
j f l ,1 f l , 2
n [7.7.1]
c f uh ,1 f uh , 2
where; j istheeffectivecomplianceofthejoint(equaltotherelativedeflectionoftheloadingplanesfor
1N of externally applied axial load), f l ,1 and f l , 2 are the deflections of the points of action of the
externalload,andfuh,1andfuh,2arethedeflectionsoftheunderheadandundernutbearingsurfacesunder
thesameexternalload.
Equation [7.7.1]cannotbeappliedwithoutknowingoftheaxialdeflectionsoftheunderheadandunder
992H
nutbearingsurfacesandtheloadingplanesduetotheexternallyappliedaxialload.Thesedeflectionscan
bedeterminedbytestorFEMmodelling.
Thedeformationofthejointisdependsstronglyonitsgeometry.Foraconcentricaxiallyloadedjoint,the
determinationoftheloadingplanefactorisrelativelysimpleprovidedthereiscontinuousanduniform
84
ECSSEHB3223A
16April2010
contact pressure between the flanges. In such cases, the loading plane factor depends only on the
followingparameters:
Duh,brg The underhead bearing diameter (or diameter of the washer if it is assumed to be
stiff)
d Theholediameter
a k Thedistancebetweentheedgeofthebearingsurfaceunderthefastenerheadandthe
pointloadintroductionoftheexternalforce
a r The shortest distance between the edge of the bearing surface and the edge of the
clampedparts
hk Thedistancebetweenthepointofloadintroductionandtheinterstice
Thesegeometricparametersaredefinedin Figure719.
993H
Duh,brg d ak
LA
h
hk
Figure719GeometryforDeterminationofLoadingPlaneFactor
Usingthegeometricanalysismethod,thejointisfirstextractedfromitsenvironmentbymakingacutata
position where the inner moment is zero (see Figure 720). In joints with multiple fasteners there are
994H
interferenceeffectsbetweenneighbouringfasteners,causingtheclampedpartstobestifferthanwouldbe
thecaseinasinglefastenerjoint.Thiseffectisduetothemutualobstructionoflateraldeformationsbythe
neighbouringfasteners.Thisproblemcanbesolvedbyusingthedistancebetweentheholeedgesoftwo
neighbouring holes as the region of the deformation cone for the calculation of the clamped parts
compliance, c .
85
ECSSEHB3223A
16April2010
Complete joint FA
2
FA
2
Bending moment
distribution in flange
LA
FA
Figure720ExtractionoftheJointfromitsEnvironment
Afterextractingthejointfromitsenvironment,itbesplitintoabasicbodyandaconnectorbody.The
basic body determines the elastic properties of the joint, including the deformation cone. The connector
bodyisthepartofthejointwheretheexternalforcesareapplied.Thepointofexternalforceintroduction
intothebasicbodyKGisassumedtobeatthehorizontalmidplaneoftheconnectorbody.
86
ECSSEHB3223A
16April2010
ak LA
LC
Figure721TheBasicandConnectorBodies
7.7.2.1 Overview
286B
ThisSubsectionpresentsasimplemethodfordeterminingtheloadingplanefactor.Themethodisbased
ona2dimenionalrepresentationofthejointanditmakesthefollowingassumptions:
Noexternalmomentsactonthefastener(butthemomentactingonthefastenerduetotheexternal
forceisconsidered)
Thejointsflangeshavenogapping
Allclampedpartsinthejointhavethesamemodulusofelasticity
Thejointbodyisofprismaticshape.
The product of the parameters concerning the 3dimensional nature of the joint can be
conservatively estimated by kas * kdh *kdw = 0.8 (see Section [7.7.3.3] for the analytical definition of
theseparameters).
Most commonly used joints can be approximated by a class of simplified joints shown in Figure 722
995H
accordingtothepointofforceintroductionintothejoint.Theintersticeshouldlieinsidetheshadedzone
of the corresponding diagram. This zone defines the region where evenly distributed compression is
assumed,andisassumedtohaveaconehalfangleof30.
Figure 722 shows only nuttightened joints, however the joint types 1, 2 and 4 are also applicable to
996H
threaded hole joints. In these cases the lower flange acts as the insert and the dimension Lc is only the
thicknessoftheupperflange.
87
ECSSEHB3223A
16April2010
1 2 3 4 5 6
Figure722JointTypesAccordingtothePointsofForceIntroduction
applicable.
Table72LoadingPlaneFactorsforSimplifiedMethod
JOINT TYPE (FROM FIGURE 7-22)
LA/LC ak/LC 1 2 3 4 5 6
0.0 0.0 0.70 0.57 0.44 0.42 0.3 0.15
0.1 0.55 0.46 0.37 0.34 0.25 0.14
0.3 0.30 0.30 0.26 0.25 0.22 0.14
0.5 0.13 0.13 0.12 0.12 0.10 0.07
0.1 0.0 0.52 0.44 0.35 0.33 0.24 0.13
0.1 0.41 0.36 0.30 0.27 0.21 0.12
0.3 0.22 0.21 0.20 0.16 0.15 0.10
0.5 0.10 0.10 0.09 0.08 0.07 0.06
0.2 0.0 0.34 0.30 0.26 0.23 0.19 0.11
0.1 0.28 0.25 0.23 0.19 0.17 0.11
0.3 0.16 0.16 0.15 0.12 0.12 0.09
0.5 0.07 0.07 0.07 0.06 0.06 0.06
0.3 0.0 0.16 0.16 0.16 0.14 0.14 0.10
0.1 0.14 0.14 0.14 0.13 0.13 0.10
0.3 0.12 0.12 0.12 0.10 0.10 0.08
0.5 0.04 0.04 0.04 0.03 0.03 0.03
88
ECSSEHB3223A
16April2010
7.7.3 Analytical Calculation of the Loading Plane Factor
144B
7.7.3.1 Overview
289B
The loading plane factor can be determined analytically by the method presented in this subsection,
althoughitisnormallyconsideredtoocomplexforhandcalculations.Thus,itisexpectedthatthemethod
presentedhereismainlyusedincomputerprograms(orspreadsheets,etc.)forfasteneranalysis.
Itisassumedthatheightoftheconnectorbodyisrelativelysmallcomparedtothethicknessoftheflanges
and thus, the load introduction can be represented by a point force. Also, it is assume that the joint
consistsoftwoflanges(suchasshownin Figure723),eachofwhichcanbetreatedseparately.
998H
Connector body FV
FA
KG
h
hk LA
Figure723BasicandConnectorBodiesofaFlange
The (overall) loading plane factor of the joint is based on loading plane factors for the two flanges
accordingto,
upper lower
n nupper nlower [7.7.2]
j j
where; nupper and nlower are the loading plane factors for the upper and lower flanges respectively, and
uypper and lower arethecompliancesoftheupperandlowerflangesrespectively.
Theloadingplanefactorforasingleflangenflangeisgivenby,
L [7.7.3]
n flange nG n M A (onlyvalidforLA/h 1/3)
h
where nG and nM are the general loading plane and moment conduction factors respectively, which are
giveninthefollowingsubsections.
89
ECSSEHB3223A
16April2010
7.7.3.3 The Moment Conduction Factor
291B
The moment conduction factor, nM, accounts for the influence of the position of the connector body
relativetothecompressionconeandtheparameterLA,andiscalculatedby,
*
n [7.7.4]
nG nG 2 D*
*
nM n2 D
0.1
where;
nG istheloadingplanefactorofthebasicbody(seebelow),
nG* is the loading plane factor of the basic body whose parameter ak/h is
reducedby0.1,and;
*
n2 D istheratioofthe2Dloadingplanefactors,wheren2D*iscalculatedwith
* aratioak/hthatisreducedby0.1,andhk/hismovedintotheregionof
n2 D evenlydistributedcompression(forthecalculationofbothfactors)
Figure724showstheprincipaltrendsofnMovertheheightofaflange.
999H
thespreadingzoneofthecompressioncone.Insomecasesthisleadstonegativevaluesforn,whichcan
beinterpretedasaonesidedgappingofthejointduetothebendingeffectoftheappliedload.
hk
h
KG,a
Compression KG,b
cone
Compression KG,c
sleeve
Figure724MomentConductionFactorVariationwiththeConnectorBodyPosition
90
ECSSEHB3223A
16April2010
7.7.3.4 The General Loading Plane Factor
292B
Thegeneralloadingplanefactorofthejointiscalculatedby,
nG n2 D k ar k dh k dw [7.7.5]
whereandn2Distheloadingplanefactorofanequivalenttwodimensionaljointcalculatedby,
ak a k hk
2
[7.7.6]
n2 D 0.71 0.29 1.41
h h h
ak a k hk
2 2
fortheinfluenceofthe3Dspatialextensionofthejointandaregivenby,
2
a a [7.7.7]
k ar 1 1.74 r 1.24 r
h h
d
k dh 1 0.25 [7.7.8]
D
uh ,brg
k dw 1 [7.7.9]
Equations [7.7.5]to [7.7.8]arevalidintheranges,
1003H 1004H
h a a
2
0 k 1 [7.7.10]
0 r k 0 .7
[7.7.11]
h h h
d [7.7.12] d [7.7.13]
0.2 uh ,brg 0.6 0.6 0.8
h D
uh ,brg
7.8.1 Introduction
145B
Underanexternallyappliedtensileloaditisgenerallyrequiredthatnogappingoccursatanyinterstice
betweenclampedpartsinajoint(i.e.theclampingforcedoesnotreachzero).Manyjointshaveadditional
requirementsforapositiveclampingforceforsealingandotheroperationalfunctions.Thus,thefailure
modeofjointseparationisdefinedtooccurwhentheclampingforceisreducedtoitscriticallevel,FK,req
(whichisgreaterthanorequaltozero).
Thejointdiagramatthecriticalconditionforseparationfailureisdepictedin Figure725.Itcanbeseen
1005H
that,undertheactionofFA,thecompressionintheclampedpartsisreducedfromtheinitialpreloadof
91
ECSSEHB3223A
16April2010
FV,mintotheminimumspecifiedlevelofFK,req.AnyincreaseinFAwouldresultininsufficientcompression
intheclampedparts,therebyconstitutingfailurebyseparation.
Fb , A
Fc , A
FV,min
FK,req
Figure725JointDiagramatOnsetofSeparationFailure
The MoS for joint separation is not intended to cover the case of slipping of a friction grip joint. The
appropriateMoSforslippingisdefinedinSubsection 9.2.1. 1006H
TheMarginofSafetyforjointseparationisgivenby,
FV ,min FK ,req
MoS sep [7.8.1]
(1 ) FA sf sep
wheresfsepisthesafetyfactordefinedforseparationasdefinedin Table54,
1007H
andistheforceratiodefinedinequation [7.4.1]
1008H
The calculations of the MoS for axial loads depend on how the joint transmits externally applied shear
loads.Twomodesofshearloadtransferareincludedintheseguidelines,FrictionGripandBearing(see
Section 9).
1009H
Thefailurecriterionforoverallloadinthefastenershouldbecheckedforbothtypesofjoint.Theoverall
load,Ftot,isthesumofthemaximumpreloadandthepartoftheexternallyappliedaxialloadthatisseen
bythefastener,i.e.,
92
ECSSEHB3223A
16April2010
Ftot=FVmax+ Fb , A [7.9.1]
Failure criteria for external load are only relevant when there is gapping (i.e. zero clamping force
betweentheflanges),whichleadstothefastenertransmittingthefullexternalaxialload(i.e.Fb=FA).
Frictiongripjointsgenerallydonotnormallyneedassessmentofthiscriterionsincetheyareintendedto
transmitshearforcesincaseofminimumcompressionforcesandthereforetheyaremorecriticalinthe
slippingfailuremode.
Table73showswhichaxialstrengthcriteriashouldbecheckedforeachtypeofjoint.
1010H
Table73CriteriaforAxialLoadAnalysis
FastenerLoading Case JointType Checktobe
performed
Overallload(Ftot) NoGapping FrictionGrip Yes
Bearing Yes
Externalload(FA) Gapping FrictionGrip No
Bearing Yes
ThefollowingsubsectionpresentstheequationsfortherelevantMoS.
Themarginsofsafetyforexternalloadcriteriaareallbasedonthefastenermaterialsyieldstrength, y ,
andultimatestrength, ult .
Themarginsofsafetyfortheoverallload,Ftot,are,
AS y [7.9.2]
MoS tot , y 1 (for yield
FV ,max n FA sf y
failure)
AS ult [7.9.3]
MoS tot ,ult 1 (for
FV ,max n FA sf ult
ultimatefailure)
andforaretheexternalload,FA,
AS y [7.9.4]
MoS A, y 1 (foryieldfailure)
FA sf y
AS ult [7.9.5]
MoS A,ult 1 (forultimatefailure)
FA sf ult
93
ECSSEHB3223A
16April2010
Equations [7.9.2]and [7.9.3]expresstheforceincrementseenbythefastenerintermsoftheexternalaxial
1011H 1012H
load, and include the appropriate loading plane factor and safety factors. The safety factors are only
appliedtotheexternalaxialload,FA,sincetheuncertaintyinpreloadisalreadyfactoredintoFV,max.
7.10.1 Introduction
149B
Toguaranteethepullout(striping)strengthcapacityofthethreadacertainlengthoftheengagedthread,
Leng,isnecessary.Innuttightenedjointsthefulllengthofthefemalethreadisnormallyengaged,i.e.Leng=
Ln.Formanufacturingreasons,threadedholejointsoftenonlyhaveengagedthreadforonlyaportionof
theholeslength.
Fastenersystemswithnormednutsarecompletelybearingcapablewhenthestrengthclassofthenutis
atleastashighthestrengthclassofthefastener.
Itisrecommendedtousenutswithlowerstiffnessthanthefastenerinordertogivemoreuniformload
transferthroughthethreadedinterface.Ifcalculationsindicatethatthefemalethreadistooweak,theuse
of a thread insert of higher strength (e.g. a helicoil) should be considered. The calculation methods
presentedinthissectionareapplicabletoanythreadsystem,includinghelicoils.
Thecriticalfastenerloadforfailureofthefemalethreadisgivenby,
Fult ,th ,n ult ,n Ath ,n c1 c 2 [7.10.1]
94
ECSSEHB3223A
16April2010
whereswisthewrenchsizeasindicatedin Figure56. 1013H
sw
Equation [7.10.4]isonlyvalidformetricthreadswith 1.4 1.9 .
D
1014H
Thecoefficientc2isgivenby,
c 2 0.897 forRS 1.0,or [7.10.5]
where;RSistheshearstrengthratioofthefemaleandmalethreadsaccordingto;
ult ,n Ath,n
RS [7.10.7]
ult ,b Ath,b
where; ult ,b istheultimatetensileshearstrengthofthematerialformingthemalethreadandAth,bisthe
failuresurfaceareaofthefemalethreadcalculatedby,
Leng ,eff p
Ath ,b D1 d 2 D1 tan( ) [7.10.8]
p 2
whereD1istheminordiameterofthefemalethreadandd2isthepitchdiameterofthemalethread(see
Section 5.4).
1015H
Iftheboththreadsaresteel,Equation [7.10.7]canbereplacedwith;
1016H
ult ,n Ath ,n
RS [7.10.9]
ult ,b Ath ,b
where; ult ,n and ult ,b are the ultimate tensile strengths of the female and male thread materials
respectively. If these are not known, they can be approximated from the tensile strength or Brinell
Hardness(HB)accordingtotheratiosdefinedin Table74. 1017H
95
ECSSEHB3223A
16April2010
Table74ShearStrengthRatiosforDifferentMaterial
Materialtype Shearstrengthratio
uts / ult ult /HB
Temperedsteel 0.60.65 2
Austeniticsteel Normal 0.8 3
F60/90 0.650.75 2.02.5
Greycastiron GG (GJL) 1.4 1.5
GGG(GJS) 0.9 2.0
Aluminiumalloy 0.7 1.5
Titaniumalloy 0.6 2
Fth p
l eng ,req 0.8 p [7.10.10]
p
c1 c 2 ult ,n d (d d 2 ) tan( )
2
whereFthistheaxialloadtransmittedbythethread.
Normallythefastenersmaterialstrengthishigherorequaltothatofthenut,sowhennormednutswith
equalorlowermaterialstrengthtothefastenerareused,thereisnoneedtocalculatefailureofthemale
thread.
If the fasteners material is weaker than the nut or insert, the analysis is the same except for Equation
[7.10.1]inwhichthevariables, ult,n andAth,n,arereplacedby ult,b andAth,brespectively.
1020H
Thus,thestrengthofthemalethreadiscalculatedby,
Fult ,th ,b ult ,b Ath ,b c1 c 2 [7.10.11]
where ult ,b is the ultimate shear strength of the fastener material, Ath,b is the failure area of the male
thread as calculated by Equation [7.7.10], and the coefficients c1 and c2 are calculated as for the female
1021H
thread.
96
ECSSEHB3223A
16April2010
7.10.4 Margin of Safety on Thread Pull-Out
152B
Themarginofsafetyonshearpulloutofthethreadundertheexternalloadiscalculatedby,
Fth ,crit
MoS th , A 1 [7.10.12]
FA sf ult
andfortheoverallload,
Fth ,crit
MoS th ,tot 1 [7.10.13]
FV ,max FA sf ult
whereFth,critisthecriticalthreadfailureload(lowerofthemaleandfemalethreadstrengths).
NoMoScalculationsarenecessaryforyieldfailureofthethread.
Themarginsofsafetyonflangecrushingarecalculatedwiththefollowingequations,
br , y [7.11.1]
MoS crush, y 1 (foryield)
uh,max sf y
br ,ult [7.11.2]
MoS crush ,ult 1 (foryield)
uh ,max sf ult
where uh ,max isthemaximumcompressivestressthatoccursundertheheadornutforeithertheFtotor
FA(referto Table73forapplicablefastenerloadcriteria).
1022H
PointEof Figure726representsajointtightenedtoapreloadFV,1,whichinducesstressesinthefastener
1023H
thatarebelowitsyieldstress.Atthefirstapplicationoftheexternalload,FA,thefastenerfeelsanextra
load Fb , A,1 ,whichtakesitshigheststressedregionbeyonditsyieldstrength(alongthepathEFG.).This
causespermanentdeformationofthefastener Lb andlocalworkhardeningofthedeformedmaterial.If
the applied external load is then removed, the fasteners preload will reduce from point E to a lower
value, point H. If FA is reapplied, the fastener load needs to exceed point G before that work hardened
partofthefastenercanyieldagain.However,therewillbeanadjacentpartofthefastenerthatyieldsata
lowerload(asithasnotworkhardened)allowingfurtherpermanentdeformationtotakeplace.
97
ECSSEHB3223A
16April2010
Force
Fb , A,1 F
b , A, 2
Fy
L
Lb
Figure726JointDiagramShowingRepeatedLoadingintothePlasticRegion
The result of continued loading as shown in Figure 726 is a nonlinear preload reduction with each
1024H
andtherefore,innormalcircumstancesloadingabovethefasteneryieldshouldbeavoided.
Figure727PreloadLosswithRepeatedLoadingintoYield
98
ECSSEHB3223A
16April2010
A simple example showing the effects of dynamic loading is given in Figure 728. Here, an alternating
1027H
externalaxialloadvaryingfromzerotoFAisappliedtoajoint.Thefastenerexperiencesadynamicload
withamplitudeof Fb , A /2.
Fb , A
Fc , A
Figure728JointDiagramShowingaDynamicExternalLoad
Thedynamicstressamplitudeinthefastenerduetotheexternalload b , A isgivenby,
Fb , A
b , A [7.13.1]
2 AS
whereASisthefastenerstressarea(seeSection 5.4).
1028H
Figure729isatypicalSNcurvethatrelatestheappliedstressinacomponenttoitsnumberofcyclesto
1029H
failure. As b , A reduces, the number of cycles to failure increases, until the stress is below a certain
threshold b , ,afterwhichthefasteneristermedfatigueresistant.
the clamped parts. Hence, to improve fatigue resistance, it is recommended to use reduced shank
fasteners.
99
ECSSEHB3223A
16April2010
b, A
Figure729TypicalSNCurve(ForaGivenMeanStress)
7.14.1.1 Overview
293B
Thegeometryofthisexampleproblemisshownin Figure730.
1031H
Figure730Exampleproblemofconcentricaxiallyloadedjoint
100
ECSSEHB3223A
16April2010
7.14.1.2 Initial Data
294B
ThefastenertobeusedforthisjointisaLN29949M6x14withananchornutLN29693.Thisfastenerhas
thefollowingmaterialproperties:
Table75:PreloadinaConcentricAxiallyLoadedJoint,fastenermaterialproperties
Material A286steel
Yieldallowable 950MPa
Ultimateallowable 1100MPa
Shearyieldallowable 548MPa
Shearultimateallowable 655MPa
Youngsmodulus(Eb) 201GPa
Thermalexpansioncoefficient( b ) 1.68e5K1
Thefollowingthreadparameterscanbecalculatedusingthestandardrelationsformetricthreads:
Table76:PreloadinaConcentricAxiallyLoadedJoint,Threadparameters
Threadpitch(p) 1mm(standard)
Fastenernominaldiameter(d) 6mm(M6fastener)
Pitchdiameter(d2) 5.35048mm
Minordiameter(d3) 4.77313mm
Stressdiameter(dS) 5.062mm
Crosssectionareaatminordiameter(A3) 17.894mm2
Stressarea(AS) 20.12mm2
The upper flangeis2mm thick withan edgedistance of16mm. The lower flangeis3mm thick withan
edgedistanceof12mm.Thediameteroftheholesthroughtheflangesis6.5mm.
Theflangeshavethefollowingmaterialproperties:
101
ECSSEHB3223A
16April2010
Table77:PreloadinaConcentricAxiallyLoadedJoint,flangematerialproperties
Material Al7075T7351
Youngsmodulus(Ec) 71GPa
Thermalexpansioncoefficient( c ) 2.2e5K1
Shearultimateallowable 262MPa
br , y , 2.0 613MPa
br , y ,1.5 524MPa
Thejointisassembledatareferencetemperatureof21C.Themaximumoperatingtemperatureis50C
andtheminimumoperatingtemperatureis4C.
Thejointissubjectedtoanexternalaxialloadof1000N.
The strength of this structure will be verified only by analysis. Thus, the safety factors for yield and
ultimateloadsshouldbereadfrom Table54.Thesevaluesarethenmultipliedbythejointfittingfactor
1032H
of1.15.Thus,thesafetyfactorstobeappliedtothedesignjustificationare:
sfy =1.25x1.15=1.4375
sfult =2x1.15=2.3
For the preload ratio, the default value of 0.65 is chosen. With this starting point the preload stress is
calculated,
V 0.65 y =617.5MPa
Withthis,thenominalpreloadcanbecalculated,
FV V AS 617.5e 6 20.12e 6 =12,424N
Firstthecomplianceofthefastenerhastobedetermined(seeSection 7.5).Thejointisnuttightenedand
1033H
has a cylindrical fastener head. Therefore the appropriate substitution lengths from Table 71 are as
1034H
follows:
Lh,sub=0.4d
Leng,sub=0.4d
Ln,sub=0.4d
Usingthesesubstitutionlengthsthefastenerscompliancecanbecalculatedusingequation[7.5.5],
102
ECSSEHB3223A
16April2010
Therefore,thefastenercomplianceisevaluatedas,
1 0.4 6 0.4 6 0.4 6 5
b
201000 28.27 17.89 28.27 17.89
=2.902e6mm/N
Now,thecomplianceoftheclampedpartsiscalculated.Theavailablediameterforcompressionconeis
determinedfromtheminimumedgedistance,
Davail=24mm(minimaledgedistanceof12mmoccursinupperflange)
Thelengthofthecompressionzoneisthefulllengththroughtheclampedparts,
Lc=5mm
Theunderheadbearingdiameterisequalthediameterofthefastenerhead
Duh,brg=10mm
Thenondimensionalcompressionconeparametersxandyareevaluatedas,
Lc 5
x 0.5
Duh ,brg 10
Davail 24
y 2.4
Duh ,brg 10
Thus,tanofthecompressionconehalfangleiscalculatedfromEquation [7.6.6],
1035H
Davail>Dlim.Therefore,Equation [7.6.10]isusedtocalculatethecomplianceoftheclampedparts,
1037H
103
ECSSEHB3223A
16April2010
10 612.258 6
2 ln
10 6 12.258 6
1 71e 3 6 0.4516
2 ln(1.3710)
604384
1.044e 6
c 1.044e 6
=0.265
c b 1.044e 6 2.902e 6
The geometry of the joint is simple, so the loading plane factor, n, can be assumed 0.5 with adequate
accuracy.Thus,theforceratioofthejointwithloadingplaneadjustmentisgivenby,
n n 0.5 0.26 0.132
The preload loss due to embedding is assumed to be 5% of the maximum preload. Therefore,
FZ 0.05 12426 =621N.
FT c b Tmax Eb Asm 1
2.20e 5
1.68e 5
50 21 201e 9
21.287e 6 (1 0.26)
474.5 N
FT c b Tmin Eb Asm 1
2.20e 5
1.68e 5
4 21 201e 9
21.287e 6 (1 0.26)
278.2 N
ThefollowingvaluesoftheunderheadandthreadfrictioncoefficientsarereadfromAppendixBforthe
specificfastenerandnutcombination:
104
ECSSEHB3223A
16April2010
Mp,max=2.0Nm Mp,min=0.4Nm
TheeffectivediameteroftheapplicationoftheunderheadtorqueisdeterminedusingEquation[4.4.5],
d uh 0.5 Dh D 0.5 10 6.5 =8.25mm
ThetanofthethreadhelixangleisdeterminedusingtherelationinEquation [6.3.9], 1042H
p 1
tan 0.0595
d 2 5.35048
Thethreadtoothangleis60therefore,
cos =cos30=0.8660
2
Nowatrialvalueofthemaximumappliedtorqueissetto13.65Nm.Assumingatorquewrenchscatterof
5%,thisleadstoamaximumappliedtorqueof14.3Nmandaminimumappliedtorqueof13.0Nm.The
appliedtoqueisspecifiedaspartofthemanufacturingprocesssoitshouldberoundedtoanaccuracyof
0.1Nm.
Now the maximum and minimum inservice preloads are calculated using Equations [6.3.14] and 1043H
[6.3.15],
M M p , min
1044H
FT
app , max
1 th ,min 1
FV , max d 2 tan d
2 2 uh uh ,min
cos
2
13.9 1000
474 .5 12553 N
0.4248 0.726
105
ECSSEHB3223A
16April2010
M M p ,max
FT FZ
app , min
1 1
FV ,min d 2 tan
th , max
d
2 2 uh uh ,max
cos
2
11.0 1000
278.2 621 4819 N
0.7028 1.221
Finally,themarginofsafetyonthetighteningprocessshouldbedetermined.Theloadinthefasteneris
composedofatensilestressduetothepreloadandatorsionalstressduetotheappliedtorque.Firstthe
torsionalstressiscalculated.
The polar section modulus is calculated with equation [6.5.5], which considers the case of total plastic
deformation(ultimate),
d0 5.062
Wp =33.95mm3
12 12
TheminimummomentabsorbedbyfrictionunderthefastenerheadisdeterminedwithEquation[5.6.3],
M uh ,min
d uh
FV ,max FT uh,min 8.25 12424 474.5 0.179 =8.823Nm
2 2
ThemaximumshearstressduetotorsionisthencalculatedwithEquation [6.5.2], 1045H
ult 110
MoS ti ,ult 1 1 =0.676
v .m. 656.3
Thenextstepintheanalysisprocessshouldbetochecktheshearpulloutofthread.Becausethenutis
madeofthesamematerialthanthefastener(thenutislikethefastenermadeofA286steel.),theshear
pulloutofthreadshouldnotbecritical.Nevertheless,thefollowingexampleshowsacalculationofthe
nutspulloutstrength.
106
ECSSEHB3223A
16April2010
7.14.2 Thread Shear Pull-Out Example
154B
The thread of joints that use normed nuts of the same material grade as the fastener (or higher) will
alwaysbeatleastasstrongasthefasteneritself.However,anexampleofthreadpulloutforsuchajoint
isdocumentedheretoshowtheanalysisprocess.
First,theeffectivelengthoftheengagedthreadiscalculatedwithEquation [7.10.3].Theeffectiveengaged
1049H
threadlengthleffoftheLN29693anchornutis5mm(althoughthefastenershouldprotrudebeyondthe
lengthofthenut).
Leng ,eff Leng 0.8 p 5 0.8 1 =4.2mm
Now,thefailuresurfaceareasofthethreadsarecalculatedwithEquations [7.10.2]and [7.10.8].Thepitch
1050H 1051H
diameterofthefemalethread,D2,isequaltothatofthefastener,d2.
Leng ,eff p
Ath ,n d d D2 tan
p 2
4.2 1
6 6 5.35 tan 30
1 2
69.29 mm 2
Leng ,eff p
Ath ,b D1 d 2 D1 tan
p 2
4.2 1
4.92 5.35 4.92 tan 30
1 2
48.57 mm 2
ThefemaletomalethreadstrengthratioiscalculatedwithEquation [7.10.7], 1052H
an anchor nut, and therefore the wrench size, sw, is not defined. Therefore, in order to evaluate the
constantc1,theouterdiameteroftheanchornutshouldbeused,whichis7.5mm.
2 2
s s 7.5 7.5
c1 3.8 w w 2.61 3.8 2.61 0.578
d d 6 6
RS 1 c 2 0.897
Now the critical axial forces for shear pullout of the male (fastener) and female (nut) threads are
determinedwithEquations [7.10.1], 1055H
107
ECSSEHB3223A
16April2010
Thatgivesallnecessaryinputforthecalculationofthemarginsofsafetyforthreadfailureunderexternal
andoverallaxialloadswithEquations [7.10.12]and [7.10.13],
1056H 1057H
7.15 References52B
7.1
1058H G.Meyer&D.Strelow SimpleDiagramsAidinAnalysingForcesinBoltedJoints,
AssemblyEngineering,Jan.1972,pp.2833
7.2
1059H VDIRICHTLINIEN VDI 2230, Oct. 2001, Systematic Calculations of High Duty Bolted
Joints
7.3
1060H S.D.Rossides Behaviour of a Simple Tension Joint with Fasteners Tightened into
Yield,BritishAerospaceReportBT.12065July1981
108
ECSSEHB3223A
16April2010
8
Eccentric Axially Loaded Joints
9B
8.1.1 Overview
155B
Inthecaseofaneccentricallyfastenedjoint,suchasshownin Figure81,therearethreeimportantaxes:
1061H
thejointaxis(OO)
thefasteneraxis(SS)
theaxisoftheexternalaxialload(AA)
Therelativepositionsoftheseaxesaredefinedtwovariables;theclampingeccentricity,s,andtheloading
eccentricity,a(see Figure81).
1062H
Figure81ATypicalEccentricJoint
109
ECSSEHB3223A
16April2010
Duringassemblythefastenerwillbepretensionedcausinganinterfaceclampingpressurewithapeakat
(ornear)thefasteneraxis(SS)asshownin Figure82.Thejointaxis(OO)isdefinedtolieatthecentroid
1063H
of the interface pressure distribution.For eccentrically clamped joints the jointaxis does not necessarily
coincidewiththefasteneraxis.
Figure82InterfacePressureDistributionafterPreloading
As the externaleccentric tensile load FA,1 isappliedthe reactionunder thefastener head the centroid of
interfacepressure(jointaxis)FKmovesawayasindicatedin Figure83and Figure84.Atthesametime
1064H 1065H
Fb , A,1 FV
Fb , A,1 FV
Figure83TypicalJointEccentricallyLoaded
110
ECSSEHB3223A
16April2010
Figure84Redistributionofinterfacepressure(withforceFA1notsufficienttocause
gapping)
Figure 85 and Figure 86 show a typical interface pressure distribution caused by an eccentric load
1066H 1067H
sufficient to cause onesided opening of the joint. Such separation occursif the axialload FA,2 exceedsa
certainvalue,whichisdependentonthepreloadandtheeccentricities,sanda.Onesidedopeningofthe
jointinterfacecausessubstantialincreaseinthefastenerstress.ThisisdiscussedinSection 8.4(seealso
1068H
Fb , A, 2 FV
Fb , A, 2 FV
Figure85AnEccentricallyLoadedJointwithGapping
111
ECSSEHB3223A
16April2010
Figure86RedistributionofInterfacePressure(ForceFA2CausesGapping)
Thejointindicatedintheabovefigureshasflangecontactsurfacesthatarerelativelywidecomparedto
the fastener diameter. Such joints are referred to as Joints With Large Areas of Contact, and the
complianceequationsareprovidedinSubsection 8.2.1. 1071H
Jointsthatemployagasket,sealorsealingringwhichisnarrowwithrespecttotheflangewidth,andwith
a flange that is stiff enough to prevent contact elsewhere on the interface surface, are referred to as
CantileveredFlangeJoints(seefigure7.3.2).Forsuchjointstheinterfacepressuredistributiondoesnot
significantlychangewithFAandthecomplianceequationsaregiveninSection 8.3. 1072H
8.1.2 Prying
156B
Injointswitheccentricloads,thefastenerloadincrementduetotheactionoftheexternalloadisgreater
than that of the equivalent concentric joint. This effect is often called prying or prising. Reference 8.3
1073H
providessomeanalyticalmethodsandexamplesofthiseffect.
8.2.1.1 Overview
300B
The equations for calculating the compliance of the clamped parts in the concentric loading case were
giveninSection 0.Theeccentriccaseismorecomplicatedsince,inadditiontolongitudinaldeformation,
1074H
the eccentricity of the externalload produces bending deformations in the flanges leading to additional
extension of the joint. Hence, the longitudinal compliance of the flanges in an eccentric joint is greater
112
ECSSEHB3223A
16April2010
thanthatofaconcentricallyloadedjoint.Also,forjointswithlargeareasofcontact,thedistributionofthe
clampingpressuredependsonthemagnitudeoftheexternalload(asdiscussedinSection 8.1),producing
1075H
furtherchangesincompliance.
Arelativelysimpleanalysis(modifiedfromReference 8.4)ofeccentricjointdeformationispossibleifthe
1076H
followingassumptionsaremade:
Theclampedpartsformaprismaticbendingsolid
Thereisnoonesidedgapping(i.e.theinterfacepressureisalwayspositive)
Allthecrosssectionsofthisprismaticsolidremainflatunderloading
Alinearstressdistributionisestablishedinthesecrosssections
The bending resistance of the fastener is much lower than that of the bending solid and may be
neglected
Thesesimplifyingassumptionslimitthejointsthatcanbeconsideredtothoseinwhichtheinterfacearea
loadedbyonefastener,Aj,isnotmuchgreater(3to5times)theunderheadarea.Seriouserrorsmayoccur
iflargerareasareassumed(seeReference 8.2).Additionally,thejointstransversewidth,C,andclamping
1077H
eccentricity,s,shouldsatisfythefollowingrelations:
C Duh,brg hmin [8.2.1]
D hmin
s [8.2.2]
uh ,brg
2
whereCisthedistancefromthefastenerjointsaxistotheflangeedgeasindicatedin Figure87.Joints
1078H
thatdonosatisfythesecriteriashouldbeanalysedusingFEMortesting.
FA
Duh ,brg
FA
Figure87LimitationsoftheEccentricJointAnalysisMethod
113
ECSSEHB3223A
16April2010
Inamultifastenerjointsuchas Figure88thespacingbetweenfasteners(fastenerpitch)isdefinedast,
1079H
andtheflangelengthassumedtobeloadedbyeachfastenerisdefinedasb,whichisassumedtobegiven
by,
Duh ,brg hmin if t Duh ,brg hmin
b [8.2.3]
t if t Duh ,brg hmin
ASSUMED
INTERFACE
AREA AB FOR
ONE FASTENER
Figure88TheInterfaceAreaforMultipleFastenerJoints
Theclampedinterfaceareashownin Figure88hasanarea,Ac=b.C,andaradiusofgyration,
1080H
1
I 2 [8.2.4]
Gc c
Ac
whereIcisthesecondmomentofareaoftheclampedinterface(aboutitsneutralaxis),whichisgivenby
bt3/12fortheinterfaceshownin Figure88.
1081H
Subjecttotheaboveassumptions,clampedpartscompliancevaluescanbecalculatedforthetwoclasses
ofeccentricjointsshowninthefollowingparagraphs.
114
ECSSEHB3223A
16April2010
8.2.1.2 Eccentric Joints Loaded at the Fastener Axis
301B
89.
Figure89AnEccentricallyLoadedJointwheres=a
The compliance of the clamped parts, c * , is higher than c and is calculated by multiplying the
complianceofanequivalentconcentricjointbyaneccentricityfactor,accordingto,
[8.2.5]
s2
c* c 1 2
Gc Ac
Asub
whichcanbewritteninamoresimpleformas,
c * c 1 2 [8.2.6]
where isthelengthratiogivenby,
s / Gc
[8.2.7]
Ac / Asub
whereAsub,istheeffectiveclampingareadefinedbelowinParagraph 8.2.1.4.1083H
This(moregeneral)classofjointsisshowninFigure7.2.4,andhasclampedpartscompliancegivenby,
a
c ** c 1 2 [8.2.8]
s
115
ECSSEHB3223A
16April2010
where; isdefinedbyequation [8.2.7]. 1084H
Thevalueofainequation [8.2.8]isalwaystakenaspositive.ThevalueofsispositiveifSSandAAare
1085H
onthesamesideofOOornegativeiftheyareonoppositesides.Thus, c
**
maybeeithergreaterthanor
lessthan c .
A S O
A S O
Figure810TheGeneralEccentricallyLoadedJoint
The effective clamping area, Asub, used in equation [8.2.7] is the crosssectional area of the assumed
1086H
compressionzone(seeSubsection 7.6.2)atitsgreatestextent,andiscalculatedasfollows,
1087H
a. ForaclampedsleevewithnarrowflangeswithDavail Dlim,
Asub
4
D lim
2
d [8.2.9]
b. ForwideflangeswithDavail>3Dlim;
2
Lc [8.2.10]
Asub Duh ,brg d2
4 10
c. FortheintermediatecasewithDlim<Davail 3Dlim
116
ECSSEHB3223A
16April2010
The joint diagram developed for concentric joints in Section 7 can also be used to describe the
1088H
characteristicsofeccentricaxiallyloadedjoints.
TriangleS1,V1,P1of Figure811representsthepreloadedconditionwherethefastenertensionFbisequal
1089H
andoppositetothecompressiveloadontheclampedpartsFK.AstheexternaleccentricloadFAincreases
from zero the fastener extends under the partial load Fb , A,1 . This causes the clamped parts to lose
compression linearly until the point when the joint begins to separate at the side nearest the load FA.
DuringthisphasethestraightlineV1toH1representsthebehaviouroftheclampedparts.IfFAisfurther
increased, the joint characteristics change to that of a hinge and lever producing a nonlinear load
deflection response of the clamped parts. The fastener therefore feels a high proportion of the external
load Fb , A,1 thanwouldbepredictedbyextrapolationofthelineV1H1.
Fb , A,1
Figure811TheJointDiagramforEccentricallyLoadedJointsshowingtheNon
LinearityoftheClampedParts
Figure 812 shows the load increment in the fastener due to the same dynamic external load FA, and
1090H
illustrates the fact that higher preload (indicated by FV2) delay the onset of separation and thus (V1
H1)<(V2H2). The figure also indicates that, after onesided opening commences, the higher the preload
reducestheloadincrementinthefastener,i.e. Fb , A, 2 < Fb , A,1 .
117
ECSSEHB3223A
16April2010
Fb , A, 2
Figure812JointDiagramwithaHigherPreload
Foraneccentricaxiallyloadedjointwithalargecontactarea,assumingtheclampedpartsbehavelinearly,
itcanbeshownthattheforceratio, e,n ,priortoonesidedopeningisgivenby,
** [8.2.12]
e,n necc c *
b c
As was the case for concentric joints, the force ratio is used to determine the incremental loads
experienced by the fastener and clamped parts due to the externally applied force according to the
relations; Fb , A e, n FA and Fc , A 1 e ,n FA .
Thedifferenceintheforceratiocomparedwiththatoftheequivalentconcentricjointisthequantitative
consequenceofthepryingeffect(seeSubsection 8.1.2). 1093H
For eccentric axially loaded joints, the loading factor derived in Section 7.7 need to be corrected. The
1094H
assumptionismadethatallcrosssectionsstayplanarduringdeformationsofthejoint.Firsttheloading
planefactorhastobedeterminedforatheoreticaljoint,wherethefastenerslongitudinalaxisismovedto
apointthatleadstoaconcentricaxiallyloadedjoint.Thentheeccentricloadingplanefactoriscalculated
with,
Lc
ncon a s [8.2.13]
Ec I c
necc
Lc
1 s 2
Ec I c
where;nconistheconcentricloadingplanefactorfromSection 7.7. 1095H
118
ECSSEHB3223A
16April2010
8.2.5 The Interface Opening Limit
161B
Figure7.2.8showsthelimitingvalueoftheexternalload,FA,sep,atwhichtheinterfacebeginstoopen.This
iscalculatedby,
FV
FA, sep [8.2.14]
a u su
1 2 1 2 e,n
Gc Gc
whereuistheeccentricityfthepointoffirstgapping(relativetothejointaxisOO).
Converselyforagivenexternalload,FA,therewillbeavalueofpreload,FV,sep,atwhichonesidedopening
willjustoccur,
FV , sep FK , sep 1 e,n FA [8.2.15]
whereFK,sepistheclampingforceattheonsetofinterfacegappingasgivenby,
FK , sep
a s u F [8.2.16]
A
Gc s u
2
FA, sep
Figure813AnEccentricallyJointLoadedtotheInterfaceOpeningLimitwiththe
LoadingPlaneInsidetheJoint
SeeReference 8.4forthederivationoftheaboverelationships.
1096H
Theeffectsofpreloadscatterduetofrictionvariablesduringassemblyarethesameforeccentricjointsas
forconcentricjoints,describedinSection5.1.2.1.
119
ECSSEHB3223A
16April2010
8.2.6.2 Preload Loss
305B
Embedding,asdescribedinSection5.1.3.2,mainlyoccursafterpreloadingbutbeforeexternalloadingis
applied.Theamountofembeddingpreloadlossforaneccentricjointcanbecalculatedasforaconcentric
jointfrom,
f Z (1 e ,n )
FZ [8.2.17]
c
where fZ is the plastic deformation caused by embedding as given by Table 6.3.1, and e,n is the force
ratio.
8.3.1 Overview
163B
These joints have small contact areas at the interstice, leading to bending moments being transmitted
through the flange. This results in the axis of the clamped parts being fixed rather than moving with
changeinexternalloadFA(thecaseforjointswithlargeareasofcontact).
For this class of joints, the procedure for determining the flange compliance can be complex and time
consuming.ThetheorypresentedinthefollowingSubsectionsappliestothespecialcaseofcircularjoints
with cantilevered flanges. For more general cases with noncircular flanges it is recommended to use
testingorFEM.
8.3.2.1 Overview
306B
References 8.3 and 8.5 give empirical relationships for joints with cantilevered circular flanges. This
1097H 1098H
categoryofjointsoftenoccurswhengasketsorsealingringsareused.
Forcantileveredflangejoints,threecompliancesareusedincombinationtodescribetheflangesbending
displacements. Figure 814 indicates three modes of deformation that determine the deflections of the
1099H
flangebetweentheaxisofthefastenerandtheaxisoftheclampedparts.Thesecompliancesarecalculated
asfollows:
L1 Thisisthecomplianceoftheflangesunderthepreloadtakinginto [8.3.1]
1 accountthebendingeffectsduetotheeccentricclampingload.
FV
L2
2 This is the bending compliance of the flanges due to the [8.3.2]
Fb , A corresponding portion of the applied load seen by the fastener. In
other words it is the flange bending due to the couple Fb , A with
forces acting at the axes of the fastener axis and externally applied
load.
120
ECSSEHB3223A
16April2010
L3
3 Thisisthebendingcomplianceoftheflangesduetotheloadseenby [8.3.3]
Fc , A the clamped parts and the corresponding portion of the applied
load.Inparticular,itistheflangebendingduetothecouple Fc , A
withforcesactingattheaxesoftheclampedpartsandtheexternally
appliedload.
Assumingnogappingoccurs,theelongationoftheflangeduetotheexternalloadisthelinearsumofthe
threeextensionmodesof Figure814,ie.,
1100H
Lc , A L3 L2 Lb , A [8.3.4]
Thusthecomplianceoftheflangeis,
L3 L2 Lb , A
c [8.3.5]
Fc , A
Axis of
clamped parts
contact area
Axis of Axis of
applied load fastener
FV
L1 FV 1
Fb , A FV
L2 Fb , A 21
Fb , A
Fc , A
L3 Fc , A 31
Fc , A
Central Axis of
Circular Flange
Figure814DefinitionofFlangeComplianceduetoBendingLoadsatDifferent
Points
121
ECSSEHB3223A
16April2010
Two types of flanges are frequently employed for circular joints, fixed flanges and loose flanges. Fixed
flanges are rigidly connected to the remaining structures. Loose flanges are separate annular rings that
compress the mating parts together. The following two paragraphs provide analytical equations for the
complianceofbothtypesofflanges.
Figure815showsatypicalfixedflangewithasmallcontactareasuchasasealingring.Forthistypeof
1101H
1 aD 2
[8.3.6]
2 aD aR [8.3.7]
3 a D a R a D [8.3.8]
where istheinversioncompliance(definedinParagraph 8.3.2.5below).
1104H
Fb
di
Fc d
Figure815DimensionsofaFixedCircularFlange(withaWeldneck)
In general, the contacting interstice between the fixed flanges can be either alargearea ora small area.
Onlythecaseofthesmallcontactareaisconsideredinthistheory.
122
ECSSEHB3223A
16April2010
8.3.2.4 Compliance of Loose Circular Flanges
309B
Figure816showsthebendingdimensionsandbendingdeformationsofaloosecircularflange.Forthis
1105H
flangethejointaxis(OO)andtheexternalloadaxis(AA)aregenerallycoincidentandasaresult 3 is
zero.Thus,thebendingdeformationduetoanexternalloadFAisonlyduetotheincreaseinfastenerload
Fb ,a andisgivenby,
L Fb , A a F
2
[8.3.9]
Thisisrearrangedtogivetheflangecompliance,
L
c aF
2
[8.3.10]
F
where istheinversioncompliance(definedinParagraph 8.3.2.5below).
1106H
di Dh
Figure816BendingDeformationofCircularLooseFlange
Theinversioncompliance, ,isgivenby,
da di
[8.3.11]
4 E hW
whereWisafactoraccountingforthepresenceoftheholesintheflangegivenby,
W=
1
12
da di d * h2 (forlooseflanges) [8.3.12]
or
1
12 12
d i S F S F 2 1 d a d i d * h 2 (forfixedflanges) [8.3.13]
radialthicknessoftheweldneck(see Figure815),histhethicknessoftheflange,andd*isdefinedby,
1109H
123
ECSSEHB3223A
16April2010
d [8.3.14]
for Wnom 500
2
d*
Wnom
1 d for Wnom 500
1000
wheredisthediameterofthefastenerholes,andWnomistheflangenominalwidthspecifiedinmm.
Consider a weld neck flange such as the one illustrated in Figure 815, fastened to an infinitely stiff
1110H
surfacesuchthatonlytheflangeundergoesbendingdeflections.Duringpreloadingthefastenerextends
by Lb andtheflangedeflectsby,
L1 1 FV [8.3.15]
WhentheexternalloadFAisaddedthefastenerisstretchedby,
Lb , A b Fb , A [8.3.16]
Theforceratioforthecantileveredflangejoint, C ,canbedefinedasfollows,
Fb , A 1
C [8.3.17]
Fb , A Fc , A Fc , A
1
F
b , A
UsingEquations [8.3.1], [8.3.2], [8.3.3]and [8.3.5],
Fb , A b 2 Fc , A c 3
1111H 1112H 1113H 1114H
[8.3.18]
Therefore, combining Equations [8.3.17] and [8.3.18] the force ratio expressed as a function of the
1115H 1116H
compliancesis,
1
2 [8.3.19]
C 1 b
c 3
simplifiedto,
1
2 [8.3.20]
C 1 b
c
124
ECSSEHB3223A
16April2010
8.3.3.2 Twin Flanges
312B
Thecompliancesderivedintheaboveparagraphareforindividualcantileveredflanges.Inmanycases,
thejointhastwosymmetriccantileveredflanges.Forsuchajoint,withfixedflangesandmboltstheforce
ratioisgivenby,
1
b 2 [8.3.21]
2
C 1 m
c 2 3
where b and c includetheeffectsofbothflangesandtheincreasedLc(unlikethosevaluesusedinthe
aboveparagraph).
1
b 2 [8.3.22]
2
C 1 m
c
8.3.4.1 Overview
313B
Consider a weld neck type flange, as shown in Figure 815 with m bolts. If the fasteners are equally
1119H
bythetriangleSPV.
Figure817SimplifiedJointDiagramforCantileveredFlangedJoints
125
ECSSEHB3223A
16April2010
8.3.4.2 Onset of Separation
314B
In the case of symmetrically flanged joint with a gasket or seal which has non linear deformation
characteristics,theclampedpartscompliance c isdependentupontheload.
Assumingmfasteners,Equation [8.3.23]canberearrangedtogive;
1122H
3 Lc , A Fc , A 3 [8.3.24]
Fb , A c Fc , A
b 2 b
m m 2
Since Lc , A and Fc , A are nonlinear, the joint diagram is also linear. To construct the joint diagram, a
range of values of Lc , A are assumed and using the gasket characteristics, the corresponding values of
Fc , A can be determined.Hence, the valuesof Fb , A can be calculatedandused to construct the joint
diagram.
Figure818TheJointDiagramforNonLinearGasketJoints
126
ECSSEHB3223A
16April2010
8.4.1.1 315BOverview
The bending moment in an eccentric joint causes bending stress in the fastener. This stress needs to be
consideredinadditiontothedirectstressescausedbyFAand Fb , A andtakenintoaccount,particularly
whendeterminingthefatigueendurancelimit.
Consideringtheloadsactingontheupperclampedpartof Figure819,thesumofmomentsaboutthe
1123H
jointaxisOOis,
M FA a Fb , A s FV s [8.4.1]
ThechangeinMduetotheexternallyappliedloadFAisgivenbythefirsttwotermsintherighthandside
ofEquation [8.2.17].Thus,theincrementtothemomentcanbewrittenas,
1124H
s
M A FA a Fb , A s 1 e ,n a FA [8.4.2]
a
AccordingtoReference 8.3thismomentincrementispartlyabsorbedbytheclampedparts,as M c , A ,
1125H
andpartlybythefasteneras M b , A (asisthecaseforaxialloads),accordingtotheratioofthebending
compliances.Thisleadstotherelation,
1 [8.4.3]
M b , A M A
1 b c
whichcanbesimplifiedfor b c togive,
c s
M b , A M A c 1 e ,n a FA [8.4.4]
b b a
where c isthebendingcomplianceoftheclampedpartsgivenby,
Lc
c 2
[8.4.5]
Gc Ac E c
and b isthebendingcomplianceofthefastenergivenby,
1 Li
b
Eb
I
i
[8.4.6]
i
where Li is the length of the ith segment of thefastener,andIi isthe second moment ofarea of the ith
segmentofthefastener,andEbistheyoungsmodulusofthefastener.
127
ECSSEHB3223A
16April2010
Lc nLc
FV Fb , A
Figure819IdealisedLinearPressureDistributionacrossanEccentricallyLoaded
JointsInterface
Thedeterminingfactorfortheendurancelimitofthefastenerinaneccentricallyloadedjointisgenerally
thenominalstress nom .Thisarisesfromthelongitudinalforce Fb , A andthemoment M b , A forthe
tensionfibresubjecttothegreateststressintheminorcrosssectionatthefirstloadbearingthreadturn.
Thisiscalculatedfrom,
nom axial bending [8.4.7]
Fb , A M b , A d 3
A3 I3 2
FA c s F d
e,n 1 e , n a A 3
A3 b a I3 2
Substitutingfor j , b andA3, [8.4.7]canberewrittenas,
1126H
1 s Lc Eb a d 3 3 e , n FA [8.4.8]
nom 1
e ,n a Lsub E c 8 Ac Gc
2
A3
where;theexpressiontotherightofthebracketsisthestressarisingonlyfromtheforce Fb , A ,andthe
expressioninbracketsisitsrelativeincreaseasaresultoftheadditionalbendingstress.
128
ECSSEHB3223A
16April2010
8.5 References
57B
8.1
1127H G.H. Junker & P.W. The Bolted Joint: Economy of Design Through Improved Analysis
Wallace andAssemblyMethods,Proc.Instn.Mech.Engrs.Vol.198BNo.14.,
1984
8.2
1128H AGATONOVICP Beitrag zur Berechnung von SchraubenVerbindungen, Dralitwelt
S8,Part2P.130137,1972
8.3
1129H WheelerA.T.,ClarkeM.J., Design Model for Bolted Moment End Plate Connections using
Hancock G. J., Murray T. Rectangular Hollow Sections, University of Sydney, Centre for
M. AdvancedStructuralEngineering,June1997.
8.4
1130H VDI2230,Oct.2001, SystematicCalculationsofHighDutyBoltedJoints
VDIRICHTLINIEN
8.5
1131H VDI2505 CalculationofFlangedJoints
129
ECSSEHB3223A
16April2010
9
Shear Loaded Joints
10B
9.1 Introduction
58B
9.1.1 Overview
169B
Shearloadedjointsareoftenusedtoconnectsheetorplatematerialandusuallyhavemultiplefasteners.
Thesetypesofjointstransmitloadstransversetothelongitudinalofthefastener.Atypicalshearjointis
illustratedin Figure91.
1132H
Figure91TypicalDoubleLapShearJoint
Shearloadedjointscanbedesignedaccordingtotwofundamentallydifferentphilosophies.
a. Frictiongrip(orslipresistant)design
b. Bearingtype(orslippedjoint)design
130
ECSSEHB3223A
16April2010
Asthenamesuggests,frictiongripjointsrelyonfrictionbetweentheflangestotransmittheshearloads.
Bearing joints rely on the fasteners to transmit the shear loads, which are generated at the interface
betweenthefastenershankandtheflangehole.
Inpracticejointswillgenerallybehaveasfrictiongripjointsuntiltheloadissufficientlyhightoslipthe
flanges. Figure92showsatypicalforcevs.extensioncurveforashearjointshowinghowafrictiongrip
1133H
joint can slip into a bearing joint configuration as the transverse force FQ increases. The extent of the
deformation occurring during the slip depends on the fastener pattern, the hole clearance and the
alignmentoftheholes.
ForthereasonsexplainedbelowinSubsection 9.2.1,whenfeasible,frictiongripdesignisthepreferred
1134H
optionforjointsinspacecraftstructures.
Figure92LoadDeformationCurveforaFastenerinDirectShear
Frictiongripjointsandbearingjointsneedanalysisofdifferentfailurecases.Themarginsofsafetyrelated
forshearloadsthatneedtobecalculatedforeachtypeofjointaresummarisedin Table91.
1135H
131
ECSSEHB3223A
16April2010
Table91FailureModesofShearJointTypes
Marginofsafety Friction Bearing Referto
gripjoint joint section
Slipping Yes No 9.2.2
1136H
Flangecrushing Yes No 7.12
1140H
[1]Thisfailuremodeisnotcoveredinthisversionofthehandbook.
Friction grip design relies on a sufficiently high clamping force maintained by the fasteners to prevent
transverse slippage of the clamped plates due to transverse load. It has the principal advantage that,
providing slip does not occur, the fastener feels only the tensile preload in the case of symmetric joints
(doubleshearandtwofayingsurfaces)shownin Figure93and Figure94.Intheunsymmetricalcaseof
1141H 1142H
relativelysmalltensionloadinthefasteners,whichisnormallyinsignificant.
Therearealsosomeotheradvantages:
Largeclearanceholescanbeused,whichfacilitatesmanufacture,assemblyandinterchangeability.
For stiffness and load transfer considerations the mating flanges in a friction grip joint can be
treatedasasingleelasticbody.
Intheabsenceofslip,vibrationlooseningdoesnotoccur.
Frictiongripdesignissimpletoimplementsincethefastenerpreload,provideditisabovetheslip
threshold,isnotcritical.Therefore,frictiongripfastenerscanbesizedontensileclampingstrength
alone.
Sincethereisgenerallyadditionalstrengthinthebearingload(atleastbeforeitsultimatefailure),
frictiongripdesigncouldbeconsideredfailsafe.
For friction grip joints the clamped parts strength should be calculated by the net section tension
method(seeSubsection 9.3.3).
1145H
Ifdesired,africtiongripjointcanalsobeanalysedforslippageintobearingusingthemethodsinSection
9.3.
1146H
132
ECSSEHB3223A
16April2010
Figure93ASymmetricalShearJoint
Lb Lc
Figure94TheJointDiagramforEachFastenerinFigure6.3
Figure95TypicalUnsymmetricalFrictionGripShearJoint
133
ECSSEHB3223A
16April2010
Figure96FastenerBendingandTensioninanUnsymmetricalShearJoint
9.2.2.1 Overview
317B
Slip resistance depends on the friction force at the interfaces faying surfaces. This force is in turn
dependentonthefastenerclampingloadandthefrictioncoefficientatthosefayingsurfaces.
Whennoexternalaxialloadisappliedtothejoint,theclampingloadisequaltothesumofthepreloads
andtheslipresistanceisgivenconservativelyby,
m
S r x s FV ,min,i [9.2.1]
i 1
where; s istheslipcoefficientoftheflangeinterface,xisthenumberoffayingsurfaces,andFV,min,iisthe
minimum preload in the ith fastener taking into account variations in preload and thermal effects (see
Equation[5.1.17b]).
Ifallofthefastenersinthejointhavethesameminimumpreload,FV,min,Equation [9.2.1]issimplifiedto,
1147H
S r x s m FV ,min [9.2.2]
wheremisthenumberoffastenersinthejoint.
InmultifastenerjointsEquation [9.2.2]canbeoverlyconservativesinceitassumesthatallfastenersare
1148H
tightenedtothelowerboundoftheirpreloadrange.Therefore,ifthejointhasenoughfasteners(typically
5ormore),theslipresistanceofthejointcanbecalculatedwith,
S r x s m FV ,av [9.2.3]
where FV,av, is the preload evaluated at the most extreme thermal conditions but with the friction
coefficients that influence the preload (i.e. th and uh ) taken as the average of their respective upper
andlowerbounds(seeSubsection5.1.2).
134
ECSSEHB3223A
16April2010
9.2.2.3 Slip Resistance For Axial Loaded Joints
319B
Providedthejointdoesnotslip,thefastenerwillexperienceonlytensileload.Forthesimplejointsshown
in Figure97theexternalaxialloadFAispartlyabsorbedinstretchingtheboltbutismainlyabsorbedin
1149H
releasing the clamped parts. The axial load reduces the interface clamping, FK, and thereby reduces the
shearloadcapacityofthejoint.
Figure97SimpleFrictionGripJointwithAddedTensionLoading
A friction grip joint is considered to have failed when the joint has slipped. It is therefore necessary to
ensurethat,
S r x s m Fc FQ [9.2.4]
whereFcistheflangecompressionforce.
Theflangecompressionisrelievedaccordingto,
Fc FV 1 e ,n FA [9.2.5]
where the force ratio term e,n takes into account any adjustments for the loading plane factor (see
Section 7.5)oreccentricity(seeSection 8.2.3).
1150H 1151H
Therelationshipbetweentheaxialloadandslipresistanceislinear(see Figure98).
1152H
135
ECSSEHB3223A
16April2010
Figure98ReductionofShearCapacitywithIncreasingExternalAxialLoad
Friction is subject to high variability. Therefore, to ensure safety in the design calculations, it should be
assumedthat s isattheminimumofitsrange,andthesurfaceroughnessofinterfacesshouldbetaken
intoaccount.Itshouldberememberedthatalthoughanincreaseinsurfaceroughnessincreasestheslip
coefficient,italsoincreasesthepreloadlossduetoembeddingwhichmaybethemoreimportanteffect
(seeSubsection 6.4.2).
1153H
AnnexCliststypicalfrictioncoefficientsforcommonlyusedjointmaterials.
1154H
The joint may be designed with specific features that increase its slip resistance such as those shown in
Figure 8.2.5. However, the calculation of the shear capacity of such joints is beyond the scope of this
document.
136
ECSSEHB3223A
16April2010
Figure99JointDesignFeaturesthatIncreaseSlipResistance
Thestrengthanalysisdocumentationforsuchajointshouldclearlyidentifyitasafrictiongripjointand
themarginofsafetyforslippingshouldbestated.
Themarginofsafetyforslippingiscalculatedtakingintoaccountthelowestpossiblepreload,FV,min,and
thesafetyfactorforultimateloads(see Table54).Itisdefinedas;
F 1 e ,n FA s x
1155H
positive.Ifused,thisfactshouldbeclearlystatedinthestrengthanalysisdocumentation.
The net tension section of each flange should also have sufficient strength to carry the shear load. See
Subsection 9.3.3forthenetsectionevaluationmethod.
1157H
Itcanalsoexistadesignrequirementthat,ifthefrictiongripjointslipintobearing,thejointdoesnotfail
inanyoftheothermodesdescribedinthenextparagraphforbearingjoints.
137
ECSSEHB3223A
16April2010
9.3.1.1 Overview
322B
910.Ajointisalsoconsideredtobeabearingjointwhenatighttoleranceshaft(orpluggauge)isused.
Bearing type joints are mostly used where the necessary clamping force to prevent slip exceeds the
clampedmaterialsunderheadcrushingstrength(seeSection 7.11).Suchjointsarethereforepreloadedto
1159H
a level below the flanges crushing strength, but sufficiently high to lock the fastener under transverse
vibration. The joint is then allowed to slip into bearing, where the shear load is transmitted via the
fastenersintotheflangesviabearingstresswithinholes.
Figure910ShearJointinBearing
Figure911showsalapjointwithalargenumberoffasteners.Insuchjointssomeofthefasteners,dueto
1160H
theirpositionattheendofthefastenerrow,takeahigherloadthanothers.Whentheexternalshearload
is applied, some load redistribution invariably takes place due to local line contact yielding. However,
afterthisredistributiontheouterfastenerscanstillexperienceshearloadssome15%aboveaverage.
138
ECSSEHB3223A
16April2010
FASTENER
SHEAR
STRESS
FASTENER NUMBER
Figure911LongBearingShearStressDistribution
An important design feature of any bearing joint is the number of disengaged threads in the load path
withinthejoint.Thisisillustratedby Figure910,inwhichthethreadedregionsofthefastenersextend
1161H
partiallyintothehole,proportionatelyincreasingtheholebearingstress.Intheextremecasewherethe
interface plane passes through a threaded part of the fastener, its shear area is reduced from the shank
crosssectional area, Asha, to the threads stress area, AS. For these reasons, it is recommended that the
thread be carefully dimensioned to minimise the length of disengaged thread in the joint, while also
ensuringthatthereissufficientthreadtoaccountformanufacturingtolerances.
Another type of bearing joint that is commonly found in mechanisms where relative rotation of the
flangesaboutasinglefastenerpivot(scissorsaction)isadesignrequirement.Inthiscasethenutneedsa
locking mechanism and the fastener preload is negligibly low. The fastener is therefore predominantly
subjecttotransverseshearandbearingloadsbetweenfastenerandhole.Also,dependingonthedegreeof
eccentricityoftheappliedload,thebendingstresseswithinthefastenermaybesignificant.
Thestrengthofabearingdesignedshearjointdependsonitsresistanceinthreefailuremodes;fastener
shear,holebearingandflangeshearoutfailure.Thestrengthcalculationmethodsforeachofthesefailure
modesaregiveninthefollowingSubsections.
139
ECSSEHB3223A
16April2010
9.3.2 Fastener Shear Failure
175B
9.3.2.1 Overview
326B
Forthecalculationoftheshearresistanceofthefastenerscrosssectionatthejointsshearingplanehasto
beused.Ifthethreadislyingintheshearplane,thefastenersstresscrosssectionalareatobeusedisAS.
Otherwise,thefastenersshankisintheshearplaneanditscrosssectionalareaAshashouldbeused.
TheMoSforshearfailureofthefastenerareasfollows:
y As [9.3.1]
MoS Q , y 1 (foryield)
FQ sf y
ult As [9.3.2]
MoS Q , y 1 (forultimate)
FQ sf ult
Inabearingdesignthepreloadischosentoensurethattheassembledjointwillslipintobearingatthe
first external shear load application. It is therefore necessary to calculate the slip load and preload the
fastenerstoalevelbelowthattheonetopreventslip.
Thejointsstrengthdependsnotonlyontheshearandtensioninteractioninthefastenerbutalsoonthe
dimensionsandmaterialpropertiesoftheclampedparts.Therelativemagnitudeoftheaxialandshear
componentsofthecombinedloadingmaymakeitnecessarytoreducetheaxialpreloadinordertoallow
more of the fasteners strength to be used in shear. It is therefore necessary, when designing a bearing
jointexperiencingbothtensionandshearloads,toworkbackwardstoaremainingclampingforcerather
thansizethefastenerbythepreloadrequirements.
Thefastenersshearcapacitydependsnotonlyontheboltsizeanditsmaterialshearstrength,butalsoon
thelevelsofaxialandtorsionalstressinthetriaxialstresssysteminitsshearingplanes.
The equations below assume that all torsional stresses (due to torque tightening) are removed by
embedding,andbendingstressesaresmall.Undertheseassumptionsonlytheinteractionbetweenaxial
andshearstressareconsideredwithinthefastener.
In cases where it cannot be assumed that bearing joints have negligible fastener bending, more
comprehensivefastenerfailuretheoriesshouldbeapplied.
Thus,theverificationofthecombinedloadedbearingjointusesstressinteractionrelations.Ifthethreadis
in the shear plane (not a recommended design practice), the following combined stress interaction
relationsshouldbesatisfied:
[9.3.3]
Rcomb, y R A, y RS , y 1 (foryield)
2 2
[9.3.4]
Rcomb ,ult R A,ult RS ,ult 1 (forultimate)
2 2
140
ECSSEHB3223A
16April2010
Iftheshankisintheshearplane,thefollowingstressinteractionrelationscanbeused:
Rcomb , y R A, y RS , y 1 (foryield)
2 3 [9.3.5]
Theaboveequations[9.3.25]usethefollowingratiosforthefastenersstrengthutilisationratios:
F e ,n FA sf y
R A, y [9.3.7]
V , max
y AS
F e ,n FA sf ult
R A,ult [9.3.8]
V , max
y AS
FQ sf y
RQ , y [9.3.9]
y AS
FQ sf ult
RQ ,ult [9.3.10]
ult AS
Ifgappingistolerable,thestressinteractionrelationsshouldalsobecheckedwithEquations[8.3.6]and
[8.3.8]replacedbythefollowing(overallaxialload)equationsrespectively:
FA sf y
R A, y [9.3.11]
y AS
FA sf ult
R A,ult [9.3.12]
y AS
Themarginsofsafetyforcombinedaxialandshearforcesactingonthefastenerare,
1
MoS comb, y 1 [9.3.13]
Rcomb , y
1
MoS comb,ult 1 [9.3.14]
Rcomb,ult
Thecalculationofthemarginofsafetyoncombinedloadsforequation[9.3.3&4]isgraphicallydepictedin
Figure 912. Point A represents a loading case that results in a combined margin of safety below the
1162H
criticallevel.PointBrepresentsthemarginofsafetyforahigherloadingcaseatthecriticallevelbutwith
the same ratio between the combined loads. This graphical interpretation cannot be used for the cubic
termsinequations[9.3.5&6].
141
ECSSEHB3223A
16April2010
RA
1.0
0.5
0.5 1.0 RQ
Figure912GraphicalDeterminationoftheMarginofSafetyonCombinedLoads
Whenajointhasmorethanonefastener,holetolerancescannotbeperfectandthusaclearancebetween
holeandfastener.Thisclearancerequiresextradeformationstobringthefullsetoffastenersintobearing,
andcreatesareductioninthestrengthofthejoint.
ThetheoryinReference 9.1canbeusedtocalculatethetotalbearingstrengthofafastenergroupwith
1163H
hole clearances at their maximum tolerance bounds. The theory makes the following limiting
assumptions:
Steelfastenersinaluminiumflanges
Fastenershanksareinbearing(i.e.nothreadinholes)
Thetheoryuseaconservativeapproachwherebythefirstfastenertofailisintheholeatthemaximum
tolerance bound, whereas as all other holes are at the centre of their tolerance bounds. This situation is
shownin Figure913.
1164H
142
ECSSEHB3223A
16April2010
CL
2
Figure913FastenerGroupwithholeclearance
Thecombinedstrengthofthefastenergroupistheidealstrengthassumingperfecttolerances(i.e.thetotal
shearstrengthofallfastenershanks),minustheforcecapabletobringallfastenershanksintobearing,i.e.
AS ult [9.3.15]
FastenerGroupBearingStrength= n n 1 FCL
sf ult
where n is the number of fasteners in the bearing joint and FCL is the shear force capable to move the
fastenershankbyadistanceequaltotheholeclearance.
* [9.3.16]
t1 t 2 FCL
FCL
t1 t 2
*
Where FCL isaparametergivenbytheempiricalrelation,
Thevariable,CL,istheholeclearance,whichisdefinedas Dh d .
Equation[8.3.16]isgraphicallypresentedin Figure914.
1165H
143
ECSSEHB3223A
16April2010
0.020
0.018
0.016
0.014
0.012
CL
0.010
(mm)
0.008
0.006
0.004
0.002
0.000
0 200,000 400,000 600,000 800,000 1,000,000
*
FCL (N/mm)
Figure914Shearloadparameterversusholeclearance
9.3.3.1 Overview
330B
ThemethodforevaluatingnettensionsectionfailureistakenfromReference 9.2.Bythistheory,thenet
1166H
tensionsectionisdefinedasthesmallestareathatisobtainedbydeductingthesummedareaofallholes
lying in any potential failure path (straight or zigzag) across a flange from the gross crosssectional
sectionalarea.Thisfailuremodeshouldbeconsideredforbothfrictiongripjointsandbearingjoints.
In order to obtain a sufficiently large net tension section area to transmit the tension load through the
flange,itisnecessarytomeetthefollowingcondition:
FQ
Anet ,min [9.3.18]
K R ult
where ult is the yield strength of the respective flange, and KR is a reduction factor for the flange
materialsultimatestrength(seeSection9.3.3.2below),andAnet,ministhesmallestnettensionsectionarea
calculatedby,
Anet ,min Agross max Aholes ,i [9.3.19]
whereAgrossistheflangeareaignoringthepresenceoftheholes,andAholes,iisthesummationofthehole
areasalongtheithpotentialpathacrosstheflange.
Thesummationoftheholesareas,Aholes,considersthefullareaofthefirstholebutthenonlyfractional
partsofeachsucceedinghole,andiscalculatedby,
144
ECSSEHB3223A
16April2010
m
pi
2
t d 1 1
[9.3.20]
Aholes
i 2 4 si d
wherepiandsiarethelateralandlongitudinaldistancesbetweenthecentresoftheithand(i1)thholes
(seeFigure8.3.3),andtisthethicknessoftheflange,andtheexpressioninsidetheangledbracketsshould
bereplacedwithzeroifitevaluatestoanegativevalue.
Figure915showsthenomenclatureusedtodefinethenettensionsection.
1167H
Figure915NomenclatureforNetTensionSectionCalculation
Themarginofsafetyonnettensionsectionfailureisthereforegivenby,
K R ult Anet ,min
MoS Q ,net 1 [9.3.21]
FQ sf ult
Aluminium alloys develop a net ultimate stress that varies according to the proportion of area lost
through holes. The curves in Figure 916 give reduction factors to be applied in Equation [9.3.20]. The
1168H 1169H
curvescorrespondtothefollowinggeneralgroupsofalloys:
AlCusolutiontreatedandnaturallyagedsheetandplate(e.g.L1632014ACLAD105OAT3)
AlCuAlZnsolutiontreatedandartificiallyagedsheetorplate(e.g.L1652014ACLAD105OAT6,L88
7075CLAD7072T6)
AlCusolutiontreatedandnaturallyagedextrudedbar(e.g.L1022014AT4),
AlCusolutiontreatedandartificiallyagedextrudedbar(e.g.L1682014AT6510,DTD5124AlZn,7075
T6)
Withtheexceptionofcurve3,thedatain Figure916isapplicabletoloadingsthatareeitherlongitudinal
1170H
ortransversetothegrain(orfinaldirectionofrolling).Thereisalsosomelimitedevidencethatcurve1
couldbeapplicabletonaturallyagedextrudedbarloadedinthetransversedirection.
145
ECSSEHB3223A
16April2010
Figure916UltimateStrengthReductionFactorsforNetTensionSectionFailure
Thecurvesin Figure916arebasedondatafromtestsonspecimenswithunfilledholes,howeverspot
1171H
checkshaveshownthattheycanbeappliedtoloadedholes.Thedatawasderivedfromspecimenswhere
theedgedistanceoftheholeswas4d.Fromthegeneralshapeofthecurvesitmaybeassumedthatthey
areapplicableatloweredgedistances,sayto2d.
Afterslippingoneormorefastenersbearagainstthesideoftheirholes.Initially,anelasticbearingstress
is developed in the plate material adjacent to the hole, as indicated in Figure 917, peaking at the line
1172H
918.Inmostcasesalinearstressdistribution,shownin Figure919,cangiveanadequateapproximation,
1174H
wherebythebearingstressis,
FQ
br [9.3.22]
dt
wheretisthethicknessoftheflange,anddisthediameterofthefastener.
If the fasteners thread extends within the hole, it should be assumed that the threaded region of the
fastenerisnotcapableoftransmittinganybearingstress.Insuchcases,thevalueoftinEquation [9.3.22] 1175H
shouldbedecreasedaccordingly.
146
ECSSEHB3223A
16April2010
Inmostcases,thevalueofdinEquation [9.3.22]isequaltothenominalholediameter,Dh.However,if
1176H
thefastenersdiameterissignificantlysmallerthanthatofthehole,thesmaller(fastener)diametershould
be used. This situation typically occurs in multifastener joints where deliberately loose tolerances are
neededtoensureassemblyofpredrilledparts.
SIDE VIEW
PLAN VIEW
Figure917InitialBearingContactElasticStressesandDeformations
147
ECSSEHB3223A
16April2010
SIDE VIEW
PLAN VIEW
Figure918Elastic/PlasticStressesandDeformationsAfterIncreasedLoading
148
ECSSEHB3223A
16April2010
SIDE VIEW
PLAN VIEW
Figure919IdealisedBearingStresses
Thebearingstresscalculatedwithequation [9.3.22]shouldbecomparedtotheallowablevaluesforyield,
1177H
thicknessratios t Dh intherange0.25to0.5.
Foral/Dhlessthan1.5itcanbeassumedthattheshearoutfailuremodewillbemorecritical(seeSection
9.3.5below).Aboveanedgedistanceratioof2.0thebearingstrengthisconstant.Iftheedgedistanceisin
1179H
the intermediate range (i.e. 1.5<al/Dh<2.0), the bearing strength should be interpolated linearly from the
valuesat1.5and2.0.
Themarginsofsafetyonholebearingstrengtharecalculatedasfollows:
br , y d t [9.3.23]
MoS br , y 1 (foryield)
FQ sf y
br ,ult d t [9.3.24]
MoS br ,ult 1 (forultimate)
FQ sf ult
149
ECSSEHB3223A
16April2010
Notethatthevariable,d,intheaboveequationsisthediameterofthefastenerandnotthediameterofthe
hole. This distinction is important for joints with oversized holes, in which case the bearing area is
determinedbythefastenerdiameter.
thefastenersarelocatedclosetotheedgeoftheplate.Thesemodesoffailuredonotneedtobeconsidered
forfastenersthataremorethan2dfromaflangeedge.
Figure920ShearOutFailure Figure921TearOutFailure
To calculate the shear stress in the flange, , the minimum shearout length is used. Referring to the
notationofFigure922,thislengthisgivenby,
2al= ef + gh [9.3.25]
Figure922NomenclatureforPlateShearOutCalculation
150
ECSSEHB3223A
16April2010
Thefollowingconditionistherelationtobesatisfiedbythedesign,foreachfastenernear(within2d)an
edge:
FQ
ult [9.3.26]
2 al t
wheretisthethicknessoftheflange,andFQisthecomponentofthefastenersloadactingtowardsthe
nearestedgeoftheflange.
Thisleadstothefollowingmarginofsafetyonshearout,
2 ult al t
MoS SO 1 [9.3.27]
FQ sf ult
9.4.1 Overview
179B
WhenthelineofactionoftheappliedshearloadFQdoesnotpassthroughthecentroidofafastenergroup
the joint is said to be eccentrically loaded. A typical eccentrically loadedshear joint is shown in Figure
1182H
923.
Figure923ATypicalEccentricLoadedShearJoint
Thetypeofloadfeltbyindividualfastenersdependswhetherthejointisoperatingasafrictiongriporas
a bearing type joint. The loadrotation curve for a typical joint loaded through friction grip and into
bearingisshownin Figure924.
1183H
Beforethejointslips,theflangesareassumeddeformtogetherasasingleelasticbody.Aftertheonsetof
slippageitshouldbeassumedthatallloadsaretransmittedviathebearingloadpaths.
151
ECSSEHB3223A
16April2010
Figure924LoadvRotationCurveforaTypicalEccentricLoadedShearJoint
ThecentroidofthefastenergroupcanbedeterminedbyassuminganytwoaxesXandYandapplying
theequations,
Asha ,1 y1 Asha , 2 y 2 ... Asha ,n y n
y m
[9.4.1]
A
i 1
sha ,i
A
i 1
sha ,i
wherexiandyiarethecoordinatesoftheithfastener,andAsha,iisitsshankarea.
Insomecasessymmetrycanalsobeusedtodeterminethefastenergroupcentroid.
In Figure923theeccentricloadF,canberesolvedtothecentroidofthefastenergroupasaforceFQand
1184H
amomentMQ=FQe(whereeistheeccentricityofloading).EachfastenerexperiencesaloadduetoF(ina
directionparalleltoit)andaloadduetoMactinginadirectionperpendiculartoalinepassingthrough
thatfastenerscentreandthecentroidofthefastenergroup.
152
ECSSEHB3223A
16April2010
TheexternallyappliedshearforceFQcanthenbedistributedamongstthefastenersproportionallytotheir
shearareasaccordingto,
Asha ,i FQ
FQ ,i m
[9.4.3]
A
j 1
sha , j
whereAsha,iistheshankareaoftheithfastener,andmisthetotalnumberoffasteners.
Similarly, the applied moment MQ can be distributed amongst the fasteners proportionally to their
bearingandshearareasandtotheirdistancefromthecentroidaccordingto,
[9.4.4]
A r
FM ,i
m
sha ,i c ,i
MQ
Asha , j rc , j
2
j 1
rc,iisthedistanceoftheithfastenerfromthecentroid.
The force vectors resulting from Equation [9.4.3] act parallel to the externally applied shear load. The
1185H
force vectors resulting from Equation [9.4.4] act perpendicularly to the moment arm from the fastener
1186H
group centroid to the relevant fastener. The shear load acting on each fastener is determined by the
vectorialadditionofthesetwoforces.
Equation [9.4.4] uses the fastener shank areas (Asha,i) as the weighting factor for the load division.
1187H
However,insomecasesthebearingareaandfastenerstiffnessmayalsobesignificant(e.g.forgroupsof
dissimilar fasteners or when the flange thickness is not constant). In such cases it is recommended to
replacethearea,Asha,i,withthestrengthofthefastenerholepair,whichisgivenby,
FQ ,ult ,i min FQ ,br ,i , FQ , sha ,i [9.4.5]
whereFbr,iistheultimatebearingstrengthoftheithhole(calculatedusingSubsection 9.3.4)andFb,iisthe
1188H
Usingtheshearforceresultingfromequation[9.4.3],therelevantmarginsofsafetyshouldbecalculated
foreachfasteneraccordingthecompletesetofbearingjointanalysismethodsinSection 9.3. 1190H
AworkedexampleofeccentricbearingjointsisgiveninSubsection9.8.
9.4.4.1 Overview
332B
In a friction grip joint the load is transmitted between the flanges via frictional forces at the clamped
interstice.Hence,therotationalslipresistanceisneededtoestimatetheloadingcapacityofthejoint.This
iscalculatedinReference 9.2byassumingthattheslippingmovementisaboutaninstantaneouscentreof
1191H
rotation.Thelocationofthecentreofrotationdependsonthefastenerpatternandthelineofactionofthe
appliedload,asillustratedin Figure925. 1192H
153
ECSSEHB3223A
16April2010
CENTRE OF
ROTATIONAL
SLIPPING
CENTROID OF
FASTENER
GROUP
Figure925ShearResistanceRelativetotheCentreofRotation
Itisassumedthat,attheonsetofslipping,themagnitudesallfastenershearloadsareequal.Thus,theslip
resistingforcetransmittedbyfasteneriisgivenby;
S r ,i S r [9.4.6]
m
whereSristheoverallslipresistanceofthejointascalculatedfromEquations [9.2.1], [9.2.2]or [9.2.3],
1193H 1194H 1195H
(seeSection9.2.).
It is also assumed that, at the onset of slipping, each fastener carries a reaction force in a direction
perpendiculartotheradiallineemanatingfromthecentreofrotation(see Figure925).Theangleofthe
1196H
radiallineisgivenby;
yi y0 [9.4.7]
i tan 1
xi x0
wherexiandyiarethecoordinatesoffasteneriandx0andy0arethecoordinatesofthecentreofrotation,
whichcanbefoundbytheiterativeprocessdescribedbelowinParagraph 9.4.4.2. 1197H
m [9.4.9]
Sr
m
cos( ) F
i 1
i Q 0 (forceequilibriuminYdirection)
m [9.4.10]
Sr
FQ (e r0 )
m
sin( ) 0 (momentequilibriumaboutr )
i 1
i 0
154
ECSSEHB3223A
16April2010
where i isafunctionoftheassumedcentreofrotation.
Asuitableiterativealgorithmshouldthenbeusedtofindacentrepointthatsatisfiesallthreeequations.
Duetosymmetry,thecentreofrotationforthefastenergroupshownin Figure925liesonaline,which
1198H
is perpendicular to the externally applied load and passing through the centroid of the fastener group.
Therefore,inthatcase,thesearchforthecentreofrotationonlyneedsvariationofoneparameter.Ifthe
fastenerpatternisnonsymmetric(withrespecttotheappliedload),thelocationofthecentreofrotation
isunknowninboththeXandYdirections.
Thisexamplecalculationusesthejointdefinedinthejointshownin Figure926.
1199H
FA/2 FA/2
FQ FQ
FA/2 FA/2
Figure926Shearloadedjointexample
Thefrictioncoefficientattheinterfacebetweentheflanges,
s =0.3
Thestrengthofthisstructureisverifiedonlybyanalysis.Thesafetyfactorforyieldsfyandultimateloads
sfultisreadfrom Table54.Thevalueisthenmultipliedwiththejointfittingfactor1.15.:
1200H
sfy=1.25x1.15=1.4375
sfult=2x1.15=2.3
In addition to the axial load of Section 7.14, the joint is also subject to a shear load. Thus, the loads to
whichthisjointissubjectedareasfollows:
FA =1000N
FQ =1000N
155
ECSSEHB3223A
16April2010
Thejointisfirstcheckedforitsfrictiongripcapacity.Thisisdonebycalculatingthemarginofsafetyon
slippingusingEquation [9.2.6].Thereisonlyoneintersticeinthisjointsox=1.
1201H
[9.3.710],
F n FA sf y 12553 0.132 1000 1.4375 0.667
R A, y
V , max
y AS 950 20.12
Thus,theMoSonfastenerfailurecanbecalculatedusingequations[9.3.13&14],
1 1
MoS comb , y 1 1 0.471
Rcomb , y 0.680
1 1
MoS comb,ult 1 1 0.650
Rcomb,ult 0.606
Next,thestrengthofbothflangesshouldbecheckedforholebearingandshearout.Bothflangesconsist
ofAA7075T7351.
Forflange1;
Thicknesst1=2mm
HolediameterDh=6.5mm
Edgedistanceratioal/Dh=16/6.6=2.46
156
ECSSEHB3223A
16April2010
Thatisabove2.0,sothebearingstrengthofthematerialisthevaluefortheedgedistanceratioof2.0.The
marginsofsafetyonholebearingisanalysedforflange1arecalculatedusingequations[9.3.23&24],
brg , y , 2.0 d t1 613 6 2
MoS brg , y 1 1 4.12
FQ sf y 1000 1.4375
524
1.85 1.5 613 524
0.5
586.3 MPa
al Dh 1.5
brg ,ult ,1.85 brg ,ult ,1.5 brg ,ult , 2.0 brg ,ult ,1.5
0.5
689
1.85 1.5 882 689
0.5
824.1 MPa
Now,themarginsofsafetyonholebearingforflange2are,
157
ECSSEHB3223A
16April2010
Forflange2;
2 ult al t 2 2 262 12 3
MoS SO 1 7.20
FQ sf ult 1000 2.3
stressis427MPa.
Figure927ShearJointExampletoShowtheMethodofNetTensionSection
Calculation
Inspectionofthejointshowsthreepotentialfracturepathsacrossthejoint;ABDF,ABCDFandABCEG.
The summed hole areas for each potential failure path across a flange are calculated using equation
[9.3.19],
ABDF:
12 2
Aholes 12.5 6 1 1 25 6 (1.00 0.88) 141mm
2
4 50 6
158
ECSSEHB3223A
16April2010
ABCDF:
0 0
0 0
32 2
20 2
12.5 6 1 1 .00 0.33) 100mm 2 AB
4 25 6 4 25 6
Aholes 1 25 6 (1 . 00 0
CEG:
0 0
0 0
32 2
20 2
12.5 6 1 1 .00 0.55) 116.5mm 2
4 25 6 4 (25 12) 6
Aholes 1 25 6 (1 . 00 0
ThelargestvalueofAholesis141mm2,whichoccursforfailurepathABDF.
Thesmallestnettensionsectionisnowcalculatedwithequation[9.3.19],
Anet ,min Agross max(Aholes)= 100 12.5 141 =1109mm2
Theproportionofnetsectionlostbyholesiscalculatedwithequation[9.3.20],
d 2 6 0.12
W 100
The reduction factor for ultimate strength in the net tension section is now found using the graph in
Figure914.Forthematerialtype1.Paragraph 9.3.3.2showsthatthereductionfactorKRisfoundtobe
1204H 1205H
approximately0.9.
Themarginofsafetyonnettensionsectionfailurecannowbecalculatedusingequation[9.3.21],
K R ult Anet ,min 0.9 427e 6 1109e 6
MoS Q ,net 1 1 6.09
FQ sf ult 50000 1.4
ThisexampleistakenfromReference 9.1.Thisexampleconsidersaneccentricshearloadactingonthe
1206H
159
ECSSEHB3223A
16April2010
Figure928OffsetLoadsonFastenerGroups
The centroid of the group (marked CG in Figure 928) is determined by symmetry. The fasteners all
1208H
havethesamebearingandshearareas,therefore,relativelysimplecalculationscanbemadetodetermine
theshearloadsonthefasteners,
20000
Horizontalloadoneachfastener: N
6
10000
Verticalloadoneachfastener: N
6
MomentatCG: 10000 0.075 =750Nm
Loadduetomomentonfastenersb,c,eandd:
0.04272 AB 750
=4134N
(4 AB 0.04272 2 ) (2 0.015 2 AB )
Loadduetomomentonfandg:
0.015 AB 750
=1452N
(4 AB 0.04272 2 ) (2 0.015 2 AB )
160
ECSSEHB3223A
16April2010
The force components for each fastener are shown in Figure 929 For more complex fastener group
1209H
geometryequation[9.4.3]wouldbeneeded.
Figure929Vectoradditionofthefastenerloads
Itcanbeseenbyinspectionthatfastenercisthecriticalfastener,sinceallcomponentsofforceactinthe
samequadrant(downwardstotheright).Nowthatthemosthighlyloadedfastenerhasbeendetermined,
themarginsofsafetyforbearingjointsshouldbecalculatedaccordingtoSection8.3.
9.6 References
63B
9.1
1210H LEEH.M. ShearJointCapabilityVersusBoltClearance,NASATM108378
9.2
1211H BAe Structural Design Data Handbook, British Aerospace Dynamics
Group,Filton,1985
9.3
1212H DOT/FAA/ARMMPDS Metallic Materials Properties, Development and Standardization,
01 Office of Aviation Research, Washington, D.C. 20591 (supersedes
MILHDBK5)
161
ECSSEHB3223A
16April2010
10
Low Duty Joints
11B
10.1 Introduction
64B
Previoussectionshavedealtwithjointsthatdesignedprimarilyforstrengthcriteria.Therearehowever
cases where the joint has to be modified to meet other criteria. These cases usually results in low duty
(lowstressed)joints.Thesejointsmaybeassumedthattobestrongenoughprovidedtheexternalloadon
the joint does not exceed a threshold value. Low duty joints do not normally need detailed strength
checksaswereshowninprevioussections.Section 10.4describesthethresholdloadingthatdetermines
1213H
whetherornotajointislowduty.
10.2.1 Overview
186B
Lowdutyjointscanresultfromconsiderationofthefollowingdesignrequirements:
Insertpulloutstrength(sandwichpanelconstructions)
Jointthermalconductivity
Jointelectricalconductivity
Handlingsize
Stiffness
Tolerancing
Redundancy(failsafe)
Eachoftheseisexpandedinthefollowingsubsections.
Threaded inserts (Reference 10.1) in sandwich panels are normally potted into the panel (see Figure
1214H 1215H
101).Generallytheshearstrengthoftheinsertishighbutitspulloutstrengthislow.ForanM4insert,
typicalshearandpulloutstrengthsare2kNand0.4kNrespectively.Thislimitstheloadstransmittedby
the joint to be substantially less than the tensilestrength of a steel M4 fastener, whichisapproximately
11kN.Therefore,itcanbeassumedthatthecriticalfailuremodeofthejointisinsertpullout.
162
ECSSEHB3223A
16April2010
Figure101TypesofInsertsUsedinHoneycombPanels
163
ECSSEHB3223A
16April2010
10.2.3 Joint Thermal Conductivity
188B
Jointthermalconductivityisadesignrequirementforcertainspaceapplications.Acommonexampleis
electronicsboxesforwhichitisnecessarytoconductmostoftheheatgeneratedinternallyawayintothe
mountingstructure.Thisneedsagoodthermalconductivityatthebaseofthebox,whichisbestachieved
with large stiff mounting surfaces, good surface finishes, high interface pressures and, if necessary,
conductive gaskets. Additionally, mountings can be necessary for extra thermal paths to conduct heat
awayfromlocalhotspots.
The interstices between clamped parts have lower thermal conductivity than the surrounding material.
Therefore,whendesigningitemssuchasequipmentboxes,itiscommonpracticetomakethemounting
feetintegralwiththebody,inordertominimizethenumberofintersticesinthethermalpath.
Table 101 lists the thermal conductivity across joints with a variety of materials, surface finishes and
1216H
interfacepressures.
Table101ThermalConductivityofMaterials
Material Surfacefinish Contactpressure Contactconductivity
[m] [kPa] [W/(mK)]
Aluminium2024T3 0.1521.65 13.82.41 114738
Aluminium2024T4 0.305 3176719 66434600
Aluminium2024T4 0.2031.27 1.248531 118021400
Aluminium2024T6 0.2031.52 677870 15603230
Aluminium2024T6 0.0760.228 1.38138 2841830
Aluminium2024T6 0.381343 1386615 34017000
BerylliumCRGrade 0.0510.102 3657939 26114800
Brass 0.2540.862 1386547 5687260
CopperETP(Hard) 0.2540.762 13.8152 2272270
CopperOFHC 0.1520.203 4488118 166090500
MgAlloyAZ31 0.8891.14 13.8241 114709
Stainlesssteel302 0.3050.508 13.8138 74415
Stainlesssteel416 1.573.81 551396475 425068100
Stainlesssteel174PH 0.4323.43 2766891 1707380
Anexamplecalculationincludingamethodtoestimatethethermalconductivityofajointisgivenatthe
endofthissection(seeSection 12.6).
1217H
The requirements for good electrical conductivity are similar to those for thermal conductivity. The
necessaryinterfaceareaisnormallysmall.Thus,asinglesmallfastener,properlypreloaded,andstandard
washer generally provide adequate clamping force and sufficiently low electrical contact resistance.
Typicalexamplesofthistypeofjointaregivenin Figure102.
1218H
164
ECSSEHB3223A
16April2010
Figure102TypicalEarthingArrangements
Thereareclearadvantagesinhandlingfastenersthathavereasonablesize.Verysmallfastenersaremore
easily dropped or damaged and assembly tends to be more time consuming when they are used.
Additionally, very small fasteners cannot be reliably torque tightened. M4 is commonly held to be the
smallest useful fastener size for general use, although other constraints may dictate the use of smaller
fasteners.Thus,lowdutyjointoftenoccurforeaseofhandlingreasons,typicalexamplesbeingarecable
andpipeclamps,see Figure103.
1219H
165
ECSSEHB3223A
16April2010
Figure103TypicalLowDutyClampConnections
10.2.6 Stiffness
191B
The fastener size or number of fasteners used has a significant effect on the joint stiffness, it may also
affectthestiffnessofthemountedassembly.Forexample,byincreasingthenumberofmountingpoints
for an equipment box, its stiffness when mounted may increase, eliminating the need for complex
stiffeningstructurewithinthebox(thussavingweightandreducingcost).
10.2.7 Tolerances
192B
In order to minimise costs the tolerances should be minimised. The larger the fastener, the larger the
acceptableholeclearancewhilstmaintainingsufficientunderheadbearingarea,andhencethegreaterthe
misalignmentthatcanbeaccommodated.Thisisparticularlyimportantforlargelightweightassemblies
where the distance between the mounting hole centres may be large, hence requiring relatively large
positionaltolerancetoavoidtheneedforexcessivemanufacturingprecision.
10.2.8 Redundancy
193B
For some structures it may be deemed prudent to add additional fasteners to maintain the structural
integrityshouldafastenerfail.Thisismoreimportantinjointswithfewfastenersandhencemayleadtoa
substantialreductioninduty.
Low duty connections may also be found in wire wrap mounts and thermal blanket anchorages. These
connections can be done without high clamping loads and, in the case of thermal blankets, metals are
undesirableduetotheirhighconductivity.Asaresult,plasticjointmembersandfastenersinconjunction
withadhesivesareoftenused.Thesetypesofconnectionsareconsideredtobebeyondthescopeofthis
guidelinesdocument.
166
ECSSEHB3223A
16April2010
Intheabsenceofotherinformationitmaybeassumedthatajointisofthelowdutytypeiftheapplied
joint load, either shear or axial, is less than 20% of the 0.2% proof load of the fastener, in the
correspondingloadingdirection.
Thisassumesthatthefastenerwillbepreloadedto75%ofyieldstressalthoughinsomecircumstancesthe
preloadwilldiffer,inwhichcasetheexternalloadlimitmaybechanged.
10.5.1 Overview
194B
Thisexampleconsiderstheequipmentboxshownin Figure104.
1220H
167
ECSSEHB3223A
16April2010
Figure104EquipmentMountingBoxGeometry
Theboxistooperateinaspaceenvironmentasfollows:
Massincludingcontents:5kg
Fastenersizetobeused:M4
Maximumacceleration:20gintothebaseofthebox.
Heattobedissipated:5W
Maximumallowableequipmenttemp:70C
Maximumambienttemp:450C
Pulloutstrengthofinsertjoints:400N
Shearoutstrengthofinsertjoints:2kN
Boxandpanelmaterial:Aluminium2024T4
Feetdimensions:30mmx20mmx3mmthick
168
ECSSEHB3223A
16April2010
10.5.2 Initial Assumptions
195B
Thefeetoftheboxeachhaveasinglefastenerloadedto65%ofyield.Thefourfastenersexperiencethe
followingtensileforcesduetotheboxsinertiaunderacceleration:
Inertiaforce=massxacceleration=5x20x9.81=981N
Therefore,theloadperfasteneris981/4=245N
AnM4x0.7fastenerwith8.8gradesteelhasthefollowingproperties:
Nominaldiameter(d) 4mm
Pitch(p) 0.7mm
Yieldstrength( y ) 640MPa
d2(Equation[4.4.?]) 3.545mm
d3(Equation[4.4.?]) 3.141mm
ds(Equation[4.4.?]) 3.343mm
As(Equation[4.4.?]) 8.779mm2
Theyieldloadofthefasteneriscalculatedwiththestressarea,
Fy y d s 640 8.779 5619 N,
Theappliedloadtothefastenerisonly4.3%ofthisyieldstrength.Therefore,lowdutycriteriaapply(see
Section 10.4)sonofurtherstrengthchecksneedtobeperformedforthefastener.
1221H
Theloadactingontheinsertsisbelowtheirpulloutstrength.
The thermal conductivity of the joint is assessed using the theory in Subsection 10.2.3, and with an
1222H
empiricalmethoddiscussedbelow.
Thepreloadinthefastenersis65%oftheyieldload,
169
ECSSEHB3223A
16April2010
The conductance calculated by the first method is considered more accurate since it is based on the
contactpressure.
The lateral conductance, C2, is 0 assumed to be 0.3 W/K, which is the recommended value for internal
honeycombpanels.
Thetotalconductancebetweentheequipmentboxandthepaneliscalculatedby,
1 1
C
1 1
C C 2
12.834 10.3 0.271 W/K
1
Theconductionfromequipmenttothenominalenvironmentisnowcalculatedby,
Thisishigherthanthe5Wthatisrequired.Therefore,thejointdesignissatisfactory.
10.6 References
69B
10.1.1 ECSSEHB3222
1223H InsertDesignHandbook
170
ECSSEHB3223A
16April2010
11
Fatigue and Fracture Mechanics of
12B
Fasteners
11.1 Introduction
70B
This section is intended to give the designer a broad understanding of Fatigue and Fracture Mechanics
analysisoffasteners,presentinghandcalculationmethodssuitableforpreliminarydesignpurposesonly.
The areas of fatigue and fracture mechanics are very much interrelated, in many cases, the relevant
analysis tasks are complementary but usually a fracture critical item is designed to only one of these
methods.
The differences between fatigueand fracture mechanicsanalyses arein theirbasic philosophies;fatigue
estimates the life to failure of an uncracked item while fracture mechanics calculates whether or not a
crackofagivensizewillpropagateinacatastrophicmannerunderserviceloading.
Ingeneral,threadedfastenerjointsshouldbeverifiedforadequatefatigue.Fastenersthatareofaerospace
quality and design that are used on a single mission (i.e. nonreusable items) are generally not fatigue
criticalsincetheyarenotsubjecttosignificantcyclicloadingotherthanlaunchandacceptancetesting.
Potentially fracture critical joints should be designed in accordance with ECSS_E_ST3201 Fracture
Control (Reference 11.1). The current issue states that shear bearing joints are not fracture critical,
1224H
whereas joints where the fasteners in tension or combined tension and bending are potential fracture
criticalitemsandrequiresthattheyarethesubjectsofeitherFailSafeorSafeLifeanalyses.
WherepossibleitisrecommendedthatthelatestversionoftheapprovedESA/NASAcomputerprogram
be used. The current version is ESACRACK, which incorporates ESALOAD, NASGRO and ESAFATIG
(seeReferences 11.2, 11.3and 12.4).
1225H 1226H 1227H
11.2.1 Fundamentals
197B
Fatigue failure occurs in a material under the cumulative effect of a number of cycles of alternating,
repeatedorvaryingstresses(usuallytensile)ofalevellowerthanthemaximumstaticfailure(oryield)
stress.Suchfailuresoccurduetoprogressiveextensionofamicrocrackinitiatingatthepointofhighest
stressinalocalstressfield.
171
ECSSEHB3223A
16April2010
The initial speed of extension of the crack is dependent on the crack propagation properties of the
material and the applied loading cycle. Eventually a point of rapid and unstable crack growth occurs
resultinginultimatefailureofthefastener.
Thenumberofcyclestofailuredecreasesastherangeofalternatingstressincreases.Forsomematerials
an indefinitely long crack free life may be expected provided that the range of alternating stress is
sufficientlylow.However,somematerials(suchasAluminiumalloy)havenofatiguelimit,althoughthe
slopeathighvaluesofNbecomesverylow.Atypicalalternatingstressversesnumberofcyclestofailure
(SN)curveisshownin Figure111.
1228H
Figure111TypicalFatigueCurveatConstantMeanStress
Anincreaseofthemeanstresslevelreducesthealternatingstressthatcausesfailure. Figure112showsa
1229H
typicalconstantamplitudeloadingand Figure113showstheeffectofchangedmeanstress.
1230H
172
ECSSEHB3223A
16April2010
Figure112TypicalConstantAmplitudeLoading
Figure113TypicalStressRatioMeanStressDiagram
Provided the combination of alternating and mean stresses does not exceed the fastener yield stress, a
high initial preload is beneficial to the fatigue resistance of the joint. This is illustrated in Figure 114.
1231H
Even very high tensile stresses, which cause plastic deformation at local stress concentrations, can be
beneficial.Plasticdeformationresultsinresidualcompressivestresses,whichreducetheeffectivestress.
TheESAFATIGsoftware(Reference 11.5)containsspecificfastenerdata.
1232H
173
ECSSEHB3223A
16April2010
Figure114TypicalPreloadEffectsonFatigueLifeofFasteners
(takenfromreference 11.4)
1233H
All cyclic stresses above the fatigue threshold level give rise to permanent fatigue damage. Damage
accumulates whether there is continuous load variation or whether there are single cycles separated by
longperiodsoftime.Equaldamagemayarisefromasmallnumberofhighstresscyclesoralargenumber
oflowstresscycles.ThisisbestenvisagedusingtheSNcurve Figure111.
1234H
Itisnormallyassumedthatwhenthesumtotalofthefractionaldamageatallstresslevelsreachesunity,
failurewilloccur.
ThemathematicalexpressionoftheaboveconceptisknownasthePalmgrenMinerLinearcumulative
damage rule. A scatter factor of 4 is typically applied to fatigue damage to allow for uncertainty in the
fatigue analysis (see Reference 12.4). Thus, for fatigue resistant design, it is necessary to satisfy the
1235H
followinginequality,
f 4 1 [11.2.1]
wherefisthetotalfatiguedamageand isthefatiguedamage.
Thefatiguedamage isgivenby,
m
ni
[11.2.2]
i 1 N f ,i
whereniisthenumberofcyclesthathaveoccurredataparticularmeanandalternatingstress,Nf,iisthe
numberofcyclestofailureatthesamestresscondition,andmisthenumberofstressconditionsthatthe
jointissubjectedto.
174
ECSSEHB3223A
16April2010
11.2.3 Fatigue Design Principles
199B
11.2.3.1 Overview
334B
Anyfactorsthataffectstresswillaffectfatiguelife.Suchfactorsgiverisetothebasicdesignprinciplesas
listedinthefollowingparagraphs.
Appropriatejointdesigncangreatlyreducetheexternalloadsfeltbythefastenersandincreasethelifeof
the joint. Any features of the design that increase the fastener compliance or decrease clamped parts
compliance will have a similar effect, e.g. smaller diameter, longer, or reduced shank bolts. For more
informationon fastener compliance refer to Section 7.5. The choice of fastener material can also have a
1236H
significantimpact.Thejointdiagramin Figure115illustratesthatatitaniumfastenerwithhalftheelastic
1237H
performanceofM8fastenersmadefromthetwomaterialsoperatingatameanstressof0.75 y .
2 Fb , A (STEEL)
2 Fb , A
(TITANIUM)
FV
FA
AXIAL FA
LOAD
AXIAL DEFLECTION
Figure115JointDiagramComparingSteelandTitaniumFasteners
175
ECSSEHB3223A
16April2010
Figure116TypicalFatigueComparison
Preferably, fasteners that are designed to reduce the stress concentrations should be specified. Features
such as large under head radius, large thread root radius and close tolerance threads should all be
incorporated.
Rolledthreadsshouldbespecified,andinparticular,threadrollingandcoldworkingoftheunderhead
radius should be carried out as the final manufacturing operation to guarantee beneficial compressive
stresses.
Preloadisthemostimportantjointparameterinresistingexternalloads(seeSection 6).Hencethemost
1239H
accuratepreloadingmethod(appropriatetotheassembly)shouldbeused,togetherwithadrylubrication
systemthatgivesthemostconsistent(ratherthanthelowest)frictionconditions.
Prevailingtorqueshouldbemeasuredduringeachassemblyoperationandthattorqueshouldbeadded
tothetorquespecifiedforpreloading.
ItisimportanttopayspecialattentiontosurfacefinishandNonDestructiveInspection(NDI)procedures
forfracturecriticalfasteners.Fatiguelifecanbegreatlyreducedbyflaws,scratchesorabrasionsarising
eitherduringmanufacturingorwhileinservice.
176
ECSSEHB3223A
16April2010
Other effects that should be considered are differential thermal stresses, stress corrosion and fretting
underload.
FractureMechanicstheoryisstillunderdevelopmentandanumberofdifferentapproachescanbefound
intheliterature.Themostwellestablishedandusefulconceptisthestressintensityfactor,aparameter
usedtocharacterisethestressdistributionaroundacracktip.
ThestressintensityisthebasisofLinearElasticFractureMechanics(LEFM)theory,whichappliesincases
ofplanestrainsuchasaradialcrackpropagatingfromafastenerthreadroot.
FromReference 11.6,thestressintensityfactorisgivenby,
1240H
KI a F [11.3.1]
whereKIisthestressintensityfactorforthetensileopeningmode(asopposedtoKIIandKIIIfortheshear
and torsion crack opening modes), is the gross axial fastener stress (calculated using the fasteners
nominalcrosssectionalarea,A),aisthecracklengthmeasuredradiallyintothebolt,andFisthestress
intensitycorrectionfactor(accountsforthegeometryandlocationofthecrack).
TodevelopanLEFManalysisusingEquation [11.3.1]itisassumedaninitialcracklength,a0,eitherfrom
1241H
knowledgeoftheinitialcracksizeorfromanestimationofthelargestcrackthatcouldescapedetection
with the relevant NDI techniques. The equation can be used to calculate the stress intensity factor for a
cracked component, and allows estimation of either the critical static stress corresponding to the crack
length,ortheinstantaneousstressintensityfordynamicloadingcrackgrowthanalysis.
11.3.2.1 Overview
339B
In a fastener subjected to both tensile and bending loads the correction factor for stress intensity factor
shouldbecalculatedby,
K I a 0 F0 1 F1 [11.3.2]
bendingstressfieldsrespectively.
Thefollowingthreetypesofcracksaremostlikelytooccurinfasteners:
circumferentialcrackatthethread(mostlikelyinfastenerswithcutthreads),
thumbnail crack at the thread (most likely in fasteners with rolled threads loaded in bending), and
thumbnailcracksatthefilletunderthefastenerhead
177
ECSSEHB3223A
16April2010
The ESCRACK/NASGRO software calculates stress intensity factors for these and other crack types,
including those mentioned above. The following paragraphs briefly describe the theory adopted by
ESCRACK/NASGRO.
by,
4 Fb
0 [11.3.3]
d local 2
and,
32 M b
1 [11.3.4]
d local 3
where Fb is the tensile force in the fastener, Mb is the bending moment in the fastener, and dlocal is the
assumedlocaldiameterofthefasteneratthecracklocation,whichiseither:
dforcircumferentialcracks,or
d3forthumbnailcracks
ESACRACK/NASGROmodelforcircumferentialfastenercracks.
0 1
dd =Majordiameter
= Major diameter
h=Threaddepth
h = Thread depth
a = d + crack depth
a=d+crackdepth
and
00 and 11 basedond
based on d
h
d 32M bb
11 3
d3
Figure117ESACRACKModelofaCircumferentialCrack
The following equations are used by ESACRACK/NASGRO to determine the stress intensity correction
factorsforcircumferentialcracksinthefastenerthread:
178
ECSSEHB3223A
16April2010
G0
F0 3
[11.3.5]
2
k
where;
G0 =0.5+0.25k+0.1875k0.1815k+0.3655k4,
k=12a/d,
And,
r [11.3.6]
F1 G1 3
r
where;
G1 =0.375(1+0.5r+0.375r2+0.3125r3+0.2734r4+0.531r5)
r=(d2a)/d
Figure 10.3.2 shows the assumed stress distributions (left) and geometry (right) of the
ESACRACK/NASGROmodelforthumbnailfastenercracks.
0 1
Figure118ESACRACKModelofaThumbnailCrack
Forthistypeofcrack,thestressintensitycorrectionfactorsfortension(F0)andbending(F1)areshownin
Table111(fromReference 11.3).
1245H 1246H
179
ECSSEHB3223A
16April2010
Table111StressIntensityFactorsforThumbnailCrackinaFastenerThread
TensionCorrectionFactor(F0) BendingCorrectionFactor(F1)
a/d a/c=0.645 a/c=1.0 a/c=0.645 a/c=1.0
0.0 Kt f x 1.00 Kt f x 0.60
Fora/d=0.05interpolationshouldbeused.
For the case of no crack (i.e. a/d = 0) the stress intensity correction factor is determined by the stress
concentrationfactorKt,whichcanbeinterpolatedfrom Table112,andanextrafactor,fx,whichisgiven
1247H
by,
1.65 2
1
[11.3.7]
1 1.464
a
for
a
1
c c
fx
1.65 2
1
a a
1 1.464 for 1
c c
Table112StressConcentrationFactorsataFastenerMinorDiameter
r/d .005 .01 .015 .02 .025 .03 .035 .04 .045 .05.
Kt 10.8 7.89 6.55 5.73 5.17 4.77 4.48 4.19 3.97 3.79
r/d .055 .06 .065 .07 .075 .08 .085 .09 .095 .10
Kt 3.63 3.49 3.37 3.26 3.16 3.07 2.97 2.91 2.84 2.78
Figure119showsthistypeofcrack.
1248H
180
ECSSEHB3223A
16April2010
Figure119ESACRACKModelofaFilletCrackunderaFastenerHead
Thestressintensityfactorsforthistypeofcrackarecalculatedasforathumbnailcrackatthethreadroot
inParagraph 11.3.2.4aboveexceptthefollowingadditionalassumptionsaremade:
1249H
Foramachinedfilleta/c=0.645
Forarolledfilleta/c=1.0
r/d=0.1
Fora/d=0thestressconcentrationfactorof2.78isused
Fora/d=0.05smoothinterpolationisused
Topredictsafelifeforaparticularfracturecriticalfasteneritisnecessarytoassumeaninitialcracksize,
calculatethestressintensityatthefirstloadcycle,calculatethechangeincracklengthduetothatcycle
andthencalculateanewstressintensity.Comparisonofthiswiththecriticalvalueforthematerial,KICor
KISCC,willdetermineiffailurewilloccur.Ifnot,thenextcyclecanbecalculatedandtheprocessrepeated
untilthecriticalcracklengthisreached.
Asthisprocessmayextendoverthousandsofcyclesitisgenerallybeyondthescopeofhandcalculations
andtheuseofcomputerbasedmethodssuchasESACRACK/NASGROareadvised.
181
ECSSEHB3223A
16April2010
11.3.4 Corrosion Considerations
203B
The principal means by which corrosion affects the fatigue life is the formation of pits, which create
geometricstressconcentrationsinthesameareasasachemicallydegradedmetallicsurface.Manyofthe
samefeaturesthataffectbasicfatiguelifearealsocorrosive.Therefore,gooddesignagainstfatiguealso
eliminatesmanypossiblecorrosionhazards.
Good fatigue performances need avoidance of stress corrosion cracking at locations in the joint where
corrosionislikely.Theeliminationofresidualtensilestress,aswellasadequatecorrosionprotectionand
thecorrectchoiceofmaterialisessential.
To give guidance on material selection References 11.3 and 11.4 classify fastener materials by their
1250H 1251H
susceptibilitytostresscorrosioncracking.Ithasbeenshownthatwithsomematerialsthereisathreshold
value, below which stress corrosion cracking will not occur. This threshold is denoted KISCC and is a
materialproperty,whichcanbemeasured.However,somematerialsdonotexhibitathresholdstressfor
stresscorrosioncracking.
It is important to note that KISCC is dependent upon the environment. Therefore, comparison between
differentmaterialsand/ordifferentenvironmentsmaynotbevalid.
Thecombinedeffectofcorrosionandfatigueisamorerapidfailurethaneitheractingalone.Thisiscalled
corrosion fatigue. The calculation of the effect of corrosion on fatigue life is not possible in the original
designcalculations.Tomakethedesigncalculationsrealisticitisessentialtoensurethatthepossibilityof
corrosionisminimisedandlatertoassesstheeffectofsalvageschemesiftheybecomenecessary.
MoreinformationoncorrosioncanbefoundinSection 13. 1252H
11.4.1.1 Overview
344B
AcriticalthreadedfastenerjointdesignusesanM8diameterTitaniumfastenertothefollowingloading
spectrumduringeachmission:
50cyclesatastressamplitudeof160MPa
4000cyclesatastressamplitudeof50MPa
880cyclesatastressamplitudeof60MPa
300cyclesatastressamplitudeof130MPa
100cyclesatastressamplitudeof90MPa
Whatisthelifeofthefastenerinnumbersofmissions?
11.4.1.2 Analysis
345B
FromtheSNcurvein Figure116thecyclestofailureforeachspectrumblockareasfollows:
1253H
182
ECSSEHB3223A
16April2010
Table113Analysisofcyclestofailure
AlternatingStress, a (MPa) CyclesperMission,n CyclestoFailure,N
160 50 103
50 4000 belowendurancelimit
60 880 5x105
130 300 7x103
90 100 8x104
FromSubsection 11.2.2,usingthePalmgrenMinerRule,thefatiguedamagepermissionisgivenby,
1254H
f 4
m
ni
4
i 1 N f ,i
50 880 300 100
4 3 5 3 4
1e 5e 7e 8e
0.37
Thefatiguelifeofthefasteneristherefore1/0.37=2.7missions.
AM8Titaniumfastener(TA28)isusedtoclamptogethertwo10mmthickflangesmadefromAluminium.
Theexternalaxialloadappliedtothejointis18kNandthefastenerhasa0.5mmdeepthumbnailcrackat
thefirstthreadintothenut.
Determinethestaticstrengthofthefastener.
Thefollowingassumptionsandpropertiesapply:
TheyieldstressforaTA28fasteneris1100N/mm
ThestressareaofanM8fastener=32.8mm
Thefastenerispreloadedtobetween60and65%ofyield
Thefastenerseesnobendingloads
Theflangesaresufficientlywidetoallowfullydevelopedcompressioncones
Nothermaleffectsapply
183
ECSSEHB3223A
16April2010
11.4.2.2 Solution
347B
Theyieldloadofthefasteneris,
Fb , y AS y 1100 32.8 36.1 kN
Thus,themaximumfastenerpreloadis
FV ,max 0.65 36.1 23.5 kN
Nowtheforceratioiscalculated.SinceLj>2d,
Dlim 3d uh ,brg
Itisassumedforthesakeofthisexamplethattheforceratiois,
0.47
Theloadingplaneisassumedtobeatthemidpointofthejointmembers,i.e.n=0.5.
Thus,
Fb , A
n n 0.235
FA
Therefore,sinceFA= 18kN,theloadincrementinthefastenerduetotheexternalloadisgivenby,
Fb,A=18kN 0.235=4.2kN
Themaximumfastenertensileloadis,
Fb,max=FV+Fb,A=21.3kN+4.2kN=25.5kN
Thegrossaxialstressis,
Fb ,max 25.5
0 1000 777 MPa
AS 32.8
For the thumbnail crack with an initial assumed length a = 0.5mm and c = 0.5mm the stress intensity
correctionfactorfortensilestress,F0,canbereadfrom Table111tobe0.84.
1255H
Thus,thestressintensityfactorcanbecalculatedwithEquation [11.3.1],
1256H
ForTA28fastenersatypicalvalueforthecriticalstressintensityfactor,K1C,is2780Nm 2 (seeReference
11.3),thereforetheMarginofSafetyonfastenerfractureis,
1257H
K 1C 2780
MoS 1 1 2.40
K1 818
Hence, given the assumptions in this calculation, it can be concluded that a thumbnail crack of 0.5mm
length(ordepth)willnotrepresentacriticaldefectandmaybetolerated.
Tofindthecriticaldefectsizeforthisjointitwouldbenecessarytorepeatthiscalculationforanumberof
cracklengthsuntilKICisexceededandinterpolateforthecriticallength.
184
ECSSEHB3223A
16April2010
11.1
1258H ECSSEST3201C SpaceengineeringFracturecontrol
11.2
1259H ESA ESACRACKUsersManualVersion4.2.0,2009
11.3
1260H NASA FatigueCrackGrowthComputerProgramNASGRO
2000.JSC22267B(includedinESACRACKmanual)
11.4
1261H NASA FatigueCrackGrowthComputerProgram
NASA/FLAGRO,August1986.CSC22267
11.5
1262H BROEKD. ElementaryEngineeringFractureMechanics
Norrdhof(Netherlands)1974.
185
ECSSEHB3223A
16April2010
12
Preloded Fastener Installation
13B
12.1 Overview75B
This section discusses a variety of methods for controlling the preload in fasteners. For each method,
descriptions are provided for the theoretical basis of each method. Also, recommendations are given
regardingpracticallimitationsandtheexpectedlevelsofaccuracy.
Thissectionalsopresentssomeguidelinesforthepreloadeffectsthatresultfromreuseofafastenerby
subjectingtomultipletensioningcycles.
12.2.1 Introduction
206B
Yieldloadorgradientcontrolledtighteningtechniquesarebasedonthefactthatwhentheyieldloadof
the fastener material is reached, the tightening torque, Mapp ceases to increase linearly with the angle of
rotation.Thatiswhen,
dM app
ceasestobeconstant,
d
orequivalently,
d 2 M app [12.2.1]
0
d 2
Whenthisconditionisdetected,acontrolsystemautomaticallystopsthetighteningprocess.Proprietary
fastenertighteningsystemsbasedonthisyieldcontrolledtighteningareavailableandtheyareclaimedto
reducethepreloadscattertolessthan8%.
Sinceeachfasteneristreatedasuniquebythecontrolsystem,thistechniqueislargelyindependentofthe
factors that contribute to high values of preload scatter when using other assembly methods. However,
themethodissusceptibletovariationsinthefastenersyieldstressandstresssectionarea.Variationsin
the proportions of these stresses, which might as arise due to effects such as variations in friction
conditions, will change the value of axial tensile stress at which yield occurs and hence the observed
fastenerpreload.Thisisexaminedtheoreticallybelow.
186
ECSSEHB3223A
16April2010
12.2.2 Method of Operation
207B
The basis of the method is an algorithm that accurately detects the occurrence of yield in the fastener.
Typicalcurvesillustratingthesystemcharacteristicsareshownin Figure121and Figure122.Itisseen
1263H 1264H
thatthegradientofthetorquerotationcurvehasacharacteristicshapewithapeakorplateaufollowedby
a marked reduction in magnitude, to about 20 30% of the maximum value. This feature is observed
under varying friction conditions and for different joint materials and hardness. The rapid drop in
gradientcorrespondstotheonsetofyieldandprovidesthebasisonwhichyieldisdetected.
From Figure 121 it is apparent that the complete torquerotation curve includes an initial zone of low
1265H
PRELOAD
APPLIED SCATTER OF FV
TIGHTENING FASTENER
TORQUE YIELD LOAD
Mtot
dM app dM app
n
d d max
dM app
d max
Figure121YieldLoadControlledTightening
187
ECSSEHB3223A
16April2010
Mapp
dM app
d
Figure122TorqueandGradientValuesasaFunctionofe
Fromthispointthecontrolalgorithmisasfollows:
dM app
a. Determinethetorquerotationgradient:
d
dM app
b. Detectandstorethemaximumvalueof:
d
dM app dM app
c. Comparecurrentvalueof with
d d max
dM app dM app
d. Ceasetighteningwhen: n
d d max
Atypicalrangeofvaluesfornis0.5to0.7.
ForyieldcontrolledtighteningtheVonMisesYieldcriterionisadopted,
v.m. V 2 3 2 y [12.2.2]
where; V and are the axial and torsional stresses at the smallest crosssection of the fastener
respectively,and y istheaxialyieldstressofthefastener.
In Equation [12.2.2] the crosssection at which the stresses are computed usually corresponds to the
1266H
threadminordiameter,d3,unlessreduceddiametershankfastenersareused,inwhichcased0shouldbe
used.
RearrangingEquation [12.2.2]yields,
1267H
188
ECSSEHB3223A
16April2010
0.2 3 2 [12.2.3]
1 2
V V
Thus,
0.2 48 M th 2 [12.2.4]
1
V FV 2 d 2
whereMthisthethreadtorque(transmittedbythefastenershank),whichfromSubsection 6.4.2isgiven 1268H
by,
FV p typ [12.2.5]
M th d2
2 cos
FromEquation [12.2.4]itisclearthattheratio 0.2
1269H V isalwaysgreaterthan1,andthusthepreload
stress developed in a fastener tightened to yield, V , is less than that corresponding to the axial yield
stress, 0.2 .
SinceMthisfunctionofthethreadfrictioncoefficientvariationsinthefrictionconditionsleadtopreload
scatter.However,assuminglubricationandotherfactorsarecontrolled,thevariationisconsiderablyless
thanthatobservedfortorquetightenedfasteners.
Significant differences in preload are expected between lubricated and unlubricated fasteners. This is
illustratedin Table121where y V iscalculatedforarangeofmetricfastenersizesandlubrication
1270H
conditions.
Table121 y V foraRangeofFastenerSizesandTypicalFrictionConditions
Fastener p d d2 y V
size [mm] [mm] [mm] typ =0.10 typ =0.15 typ =0.20
M3 0.5 2.459 2.675 1.198 1.330 1.483
M3.5 0.6 2.850 3.110 1.203 1.337 1.490
M4 0.7 3.242 3.545 1.207 1.341 1.496
M5 0.8 4.134 4.480 1.191 1.321 1.472
M6 1.0 4.917 5.350 1.192 1.330 1.483
M8 1.25[2] 6.647 7.188 1.187 1.316 1.466
M10 1.5[2] 8.376 9.026 1.181 1.309 1.457
1.Assumes =30
2.ForMJfasteners,M8shouldhavep=1.0mmandM10shouldhavep=1.25mm
189
ECSSEHB3223A
16April2010
12.3.1 Introduction
209B
Thismethodoftightening(alsocalledturnorthenutcontrolledtightening)isbasedonthefactthatitis
possible to stretch the fastener a prescribed amount by turning the nut or fastener a certain number of
degrees.Therefore,controlbyangleofrotationisindirectlycontrolbymeasuringtheelongation.
Themethod,althoughsimpleinconcept,isdependentonanumberofparametersspecifictoaparticular
joint.Therefore,itshouldbeconsideredanempiricalmethod.
Therearetwovariationsonthemethod:
a. Elasticrangetightening,wherethefastenerisnotplasticallydeformed.
b. Plasticrangetightening,wherethefasteneristakenbeyonditsyieldpoint.
Eachoftheseisdiscussedinthefollowingsubsections.
This variant is based on the measurement of extension occurring in the grip length of the fastener. The
effect of frictional variations is eliminated except for the effect they have on the snug torque, which
determines the starting point of angle measurement. However, this method is subject to a number of
shortcomingsrelatedtovariationsintheeffectivestiffnessofthefastener,andthejointclampedparts.
Since the part of the fasteners length is not under tension, an allowance needs to be made for the
diffusionofstressandstraininthethreadsthatareengagedtothenut.Theregionofunengagedthread
regionbelowthenutneedsalsotobeconsideredsincethishasalowereffectivecrosssectionalarea,and
hencestiffness,thantheunthreadedshank.Withoutclosedimensionalcontrolovertheshankandthread
lengthsitisimpossibletobecertainhowmanythreadswillbeinthetensilestrainregion.Furthermore,
some deformation of the nut takes place near the first few threads, adding to the uncertainty in the
elongation.
Someofthesepointsareillustratedin Figure123,whichshowsasectionthroughatypicaljointwiththe
1271H
variationofstrainalongthelengthofthefastener.Thereductioninstrainintheshankisapparentand
indicatestheimportanceofcontrollingthelengthsoftheshankandthethreadtoareasonabledegreeof
accuracyinordertopredicttheelongationofthefastenerunderload.
190
ECSSEHB3223A
16April2010
Figure123VariationofStrainalongFastener
The secondimportant element in thismethod is control of the clamped parts stiffness. Compression of
the joint members and crushing of the surfaces under the fastener headand nut have to be considered.
These factors can cause significant uncertainty and may be alleviated by the use of hardened steel
washers.
Associatedwiththisaspectisnonalignmentoftheflangespriortoassembly.Themethodassumesthat
the flanges are perfectly parallel, and in close uniform contact and that there is no deflection of the
flanges. The first assumption can be ensured by specifying a snug torque to which the fastener is
tightened prior to commencing measurement of the rotation angle. This ensures that the joint faces are
effectivelyincloseuniformcontactbeforeassembly.Thesecondassumptionmaybeensuredbycareful
design,andbyconsiderationoftheeffectsofanygasketmaterialsthatmaybeincludedinthejoint.
The accuracy with which the operator can set the rotation angle depends on the accessibility of the
fastenerandhavingspecificationsthatinvolvesimplefractionalrotations(e.g.ofarevolution).
12.4.1 Introduction
211B
Asthefasteneristighteneditslengthincreases,changingboththetransittime(ortimeofflight)ofan
ultrasonicsignalandtheresonantfrequencyofaxialvibration.Thechangeineitherofthesecanbeused
todeterminefastenerextension.Thus,twomainultrasonicmethodsexist;
a. thetransittimesystem,inwhichthetimetakenforapulseofultrasoundtotravelfromoneendof
thefastenertotheotherandbackagainismeasured,and
b. the resonant oscillator system in which the ultrasonic resonant frequency of the fastener is
measured.
Oncethelengthofthefastenerisreliablyobtained,thepreloadcanthenbecalculatedfromitsstrain.
191
ECSSEHB3223A
16April2010
12.4.2 Ultrasonic Extensometers
212B
12.4.2.1 Overview
348B
Extensometers employ a technique based on the transit time of an ultrasonic pulse traveling along the
fastenersaxis,butwithallowanceforthereductionofsonicvelocitywithincreasingstress.
Theprincipleofoperationisasfollows:
a. Apulserwithintheinstrumentshocksatransducerthatthendeliversabrief,highlydampedpulse
ofultrasoundtooneendofthefastener.
b. Thispulsetravelsdownthefastener,echoesoffthefarend,andreturnstotheoriginaltransducer.
c. Theinstrumentmeasuresandrecordsthetransittimeofthepulse.
d. The fastener is then tightened, increasing the length of the path to be covered by the signal and
decreasingthesonicvelocity.
e. Thenewincreasedtransittimeisrecorded.
f. Using this data, the instrument (typically incorporating a microprocessor) computes the total
changeintransittime,computesanddiscardsthatportionofthechangewhichresultedfromthe
changeinvelocity,andthendisplaystheactualchangeinfastenerlength.
Afurtherimportantfactor,whichhastobeconsidered,istheeffectoftemperature.Atemperaturechange
willaffectthelengthofthefastenerandthesonicvelocity.Iftemperatureofthefastenerchangesbetween
the unstressed and stressed readings its influence on length and velocity should be factored out by the
measuringequipment.
Typically, therefore, before commencing preload measurement, three factors will be set into the
instrument. These are velocity and stress factors for the fastener material and a temperature factor.
These factors should be experimentally determined, since theoretical calculation of the relationship
betweenpreloadandchangeintransittimeisnotpossible.Extensometermanufacturershavedetermined
factorsformanymaterialsandtheseareavailabletotheuser.Incaseswherethedataisnotavailable,the
instrumentcanbeusercalibratedonasampleofthefastener.Subsequentrecalibrationisusuallydoneby
meansofacalibrationblockprovidedwiththeinstrument.
Aftercalibrationatypicalprocedure(Reference 12.1)fortighteningajointisasfollows:
1272H
a. Theinstrumentissetinthelengthmode.
b. Thetemperatureofthefastenertobetightenedismeasuredandtheappropriatetemperaturefactor
inputintotheinstrument.
c. An acoustic couplant such as glycerine or oil is applied to the fastener and the transducer set in
place.Thismaybeattachedtoeitherendofthefastener.
d. Theinitiallengthofthefastenerisultrasonicallymeasuredandrecordedforfuturereference.
e. Steps3and4arerepeatedforotherfastenersinthejoint.
f. Theinstrumentisswitchedtothechangeoflengthmodeandtheextensionnecessarytogivethe
desiredpreloadiscalculated.
192
ECSSEHB3223A
16April2010
g. The joint is pulled together by tightening each fastener to approximately 30% of the required
preload. Torque control is commonly used for this pass. The length of the fasteners may be
monitoredattheendofthepasstoconfirmthatresultsarereasonable.
h. Additionalpassesaremade,tighteningthefastenertoahigherpercentageoftherequiredpreload.
Asmallsampleoffastenersismeasuredwiththeextensometerattheendofeachpass,buttorqueis
stillusedforcontrol.
i. The extensometer is used to control the tightening process for values of preload above, typically,
70%oftherequiredpreload.
j. It is recommended to remeasure every fastener in the joint after the final pass to monitor the
relaxationeffectssuchasembedding.
Inlaboratoryconditionsinstrumentsarecapableofresolvingtobetween2.5mmand0.25mm.However,
inthefieldtheperformancemaybesignificantlyworse,byanorderofmagnitudeormore.
The accuracy to which preload can be controlled is a function of fastener length. The fastener diameter
shouldalsobetakenintoaccountifitislessthantentimesthesonicwavelength.
Asfortheanglecontrolledtighteningmethod,variationsinfastenerandjointstiffnesswillbereflectedin
preloadscatterandshouldbeconsidered.
Transducersvaryinfrequency,diameter,signalstrength,acousticdamping,etc,andatpresent,selection
isstilllargelyamatteroftrialanderror.
Goodandconsistentacousticcouplingofthetransducerisnecessary.However,itisnotpossibletocouple
allfastenerconfigurationssinceasufficientlylargemountingareashouldbeavailableforthetransducer
and the opposite end should provide a substantial signal reflection. Hexagon headed fasteners usually
workwellbutsocketheadcapscrews,maynotpermitthismethodunlesstheyarelargeenoughtoallow
the transducer to be mounted in the wrench socket, or beside it. If the transducer is too large, it may
receiveunwantedreflectionsfromtheunderheadinterstice,etc.
Fasteneridentificationmarkingscanpreventasatisfactoryacousticcouplingandmachiningofthehead
may be necessary to eliminate air, which creates acoustic impedance. Even with perfectly flat surfaces,
consistentcouplingcanbedifficulttoachieve,sincevaryingthecontactpressurebetweenthetransducer
and the fastener can significantly alter results. In critical applications, it may be necessary to accurately
controlthepressurebetweenthesurfaces,asdescribedinSection12.3.2.
Differentmaterialshavedifferenteffectsontheacousticsignalwithregardtoabsorptionandscattering.
For example, some stainless steels aregrainyand can distort the signal. The American A286 materialis
believedtocontainTitaniumplateletswithinitthatcausefalseechoesandsevereattenuation.Irregular
hardness,e.g.casehardening,canalsowarpthesignal.
The fact that preload is actually calculated from the strain needs that the length of the portion of the
fastener under tension is accurately known. This is usually not the case and therefore, the standard
methodofcalibrationcanbeinadequateincaseswhereahighdegreeofaccuracyisneeded..
193
ECSSEHB3223A
16April2010
12.4.2.4 Increasing the Accuracy of Preload Measurement
351B
Although the procedures described above enable fastener preload to be controlled far more accurately
thanispossiblebytorquecontroloranglecontrolledtighteningmethods,generallythemethodsleadto
preloadscatterofgreaterthan5%.Higheraccuracycanbeachievedbymakingtwomainmodificationsto
theprocedure.
a. Increasingtherepeatabilityofthecouplingbetweenthetransducerandfastenerhead.Thiscanbe
achievedbyusingtransducersthatcanbescrewedintoatappedrecessinthehead.Bytightening
thetransducertoapredeterminedtorque,moreconsistentcouplingisachieved.Itshouldbenoted,
however,thatthefastenerswillneedtobespeciallymachinedandthetransducersmayhavetobe
speciallymade.
b. Usingadirectmethodforderivingasetofcalibrationdata,specifictothefastenersbeingused.As
mentionedpreviously,theusualmethodofcalibrationinvolvesmeasuringthefastenerextensionto
calculatepreloadwiththeinherentinaccuraciesmentionedinparagraph12.3.2.Theseinaccuracies
canbeeliminatedbymeasuringthefastenerloadandextensionrelationshipasfollows:
1. Thefastenerispassedthroughtwosmallsteelplatessandwichingasuitableloadwasher,the
totalthicknessbeingequaltothethicknessoftheclampedpartsoftherealjoint.
2. Thenutistightenedtogivevariousincrementsofpreload.
3. For each increment the preload and extension readings (given by the load washer and
ultrasonicextensometer)arerecorded.
4. Anydesiredpreloadcannowbeobtained,knowingtherequiredextensionreading.
Approximatevaluesofthecalibrationfactorsareinputintotheinstrumentatthestart.Thesedonotneed
to be exact, since any error in the extension reading is unimportant, as the output has been calibrated
againstloadrequired.However,itisimportanttousethesamefactorsforallfuturemeasurementsforthe
calibratedjointconfiguration.
An appreciable amount of data has been obtained using this method (References 12.2 and 12.3) but it
1273H 1274H
tendstobeveryspecifictoacertaintypeandsizeoffastener.Ifonlyasmallnumberofdifferentfasteners
are involved, it may be practicable to calibrate each type of fastener for its specific required preload.
However,ifthenumberislargethismaynotbepossible.Itispossibletousedataobtainedforalimited
variety of specimens to develop empiricalformulae that hold trueforfasteners of different dimensions,
(seeReference 12.4).Ithasbeenfoundthatthefollowingrelationshipcanbeapplied:
1275H
Lb
a b L1 c L2 [12.4.1]
FV
whereL1istheplainlengthofthefastenerandL2isthethreadedlengthofthefasteneruptothefaceof
thenut,and,a,bandcareconstants,tobeevaluatedfortheparticularfastenersystem.
Examples of these methods include the Reflection Oscillator Ultrasonic Spectrometer (ROUS) and
improvedPseudoContinuousWave(PCW)methods.Thefirststepistoestablisharesonantconditionin
thefastener.Asthefastenerisstressedthewavevelocitythroughitdecreases.Thefrequencyofexcitingis
held constant by the transducer and thus the wavelength therefore decreases (since the velocity is the
product of the wavelength and the frequency). Therefore, to reestablish a resonant condition, it is
necessarytodecreasethetransducerfrequency.
ThereaderisdirectedtoReferences 12.5to 12.9formoredetailedinformation.
1276H 1277H
194
ECSSEHB3223A
16April2010
12.4.2.6 Scatter
353B
It is not possible to give an accurate maximum error figure covering all fastener configurations, since
this depends on many factors, including fastener dimensions, fastener material and the type of nut.
However,asaguidetotheorderofmagnitudeoferrors, Table122showsvaluesofBvaluescatter,for
1278H
extensometer readings for fasteners of six different diameters that were tested as described in Section
12.3.2 (see also Reference 12.4). It should be noted that, for each diameter, different lengths of fastener
1279H
were tested. Since the largest errors are likely to be incurred with the shortest fastener, the minimum
lengthusedisquoted.
Itcanbeseenthattheworstcasesinthissetoftestsproducedscattervaluesintheregionof6%,whileitis
possibleinsomecasestoobtainresultswithin1%.
Table122TypicalScatterofUltrasonicExtensometerPreloadReadingsUsinga
DirectLoadMethodofCalibration
Fastener Fastener Fastener BValue
Diameter Length Material Scatter
(mm) (mm)
8 51.2 Inconel 5.5%
10 40.7 Inconel 6.4%
12 52.6 Inconel 4.9%
14 61.7 Inconel 2.8%
18 70.0 Titanium 6.2%
12.5.1 Overview
213B
Thismethodofpreloadcontrolislongestablishedandisincludedhereforcompleteness.Ithasanumber
ofdisadvantagesthatareoutlinedbelow.
Amicrometerisusedtomeasurethefastenerextensionbytakingreadingsbeforeandaftertightening.To
makeaccuratemeasurements,itisnecessarytoprovidegaugingsurfacesateachendofthefastener.This
isusuallydonebygrindingtheendfacessquareorbyprovidingaconicalrecessintowhichasteelball
canbelocated.
Themethodneedsaccesstobothendsofthefastenerandsufficientclearancetoaccommodateacaliper
typemicrometer.
195
ECSSEHB3223A
16April2010
The method also needs the application of both analytical and practical judgment, and most of the
disadvantages cited for the angle controlled tightening method apply. Tension cannot be rechecked
withoutslackeningthenutandretighteningitunlessalogiskeptofallfastenerdimensions.
The level of accuracy to which the preload can be controlled, assuming a uniform precision in
measurement,isprimarilyafunctionofthefastenerlength.
12.6.1 Overview
215B
Inmanysituationsathreadedfastenerhastobeloosenedandretightenedagain,e.g.MechanicalGround
SupportEquipment(MGSE)hardwareduringserviceonthegroundorpartsofatestmodelthatarelater
reused as flight hardware. Reuse of fasteners is desirable since new fasteners are often expensive,
however additional analysis is recommended. Two major parameters are affected by reuse; the friction
coefficients,andtheprevailingtorquesofthelockingdevice.
During the tightening procedure the surfaces of the threads are at first planated by the force acting
perpendiculartothesurface.Undertheeffectofanincreasedpreloadduringthetightening,thegrinding
of the two moving surfaces creates scratches and grooves on both surfaces, thus resulting in higher
friction coefficients. The friction coefficients of unlubricated fasteners may increase by up to 100%, and
further increases can occur with a higher coefficient of utilisation. This effect depends mainly on the
hardnessoftheengagedmaterialsandtheirsurfacetreatments.Therefore,theuseofexperimentaldatais
recommendedwhendeterminingtheeffectsofreuseonthefrictioncoefficient.
Theresultsofaseriesofreusetestsaresummarisedin Table123,whichshouldbeusedinplaceofTable
1280H
B1iffastenerreuseisexpected.Thevaluesin Table123arethemaximumandminimumvaluesforthe
1281H
2ndto5thapplicationsoftighteningtorque.Asfrictioncoefficientsmayincreasewithfurtherretightening,
this data should be used with care, and extensive reuse should be avoided. Nonetheless, the first
retighteningresultsinthehighestincreaseoffrictioncoefficients.Dependingonthefastenersizeandnut
type, the friction coefficients after the first retightening are 50% to 75% of the value after the 5th
retightening.ThedatainTable11.51isnotapplicableforMJtypethreads;inthatcaseseeISO5858and
DIN65349.
Using a lubricant for the fastener significantly reduces the effects of retightening on the friction
coefficients. Here also, an increase of the friction coefficients occurs, but it is around 10% 15% of the
valuemeasuredatthefirstuse.
196
ECSSEHB3223A
16April2010
Table123FrictionCoefficientsofUnlubricatedReusedFasteners
Fastener TypeofNutorInsert th ,max th ,min Flangematerial
AA7075T7351
h ,max h ,min
LN29949M4(A286) Anchornut(floating)LN29693
HelicoilLN9499madeof
bronzeCuSn6
NutLN9161(selflocking,
hexagonflanked)
LN29949M5(A286) Anchornut(floating)LN29693 0.226 0.129 0.316 0.229
HelicoilLN9499madeof 0.190 0.154 0.330 0.217
bronzeCuSn6
NutLN9161(selflocking, 0.196 0.100 0.171 0.098
hexagonflanked)
LN29949M6(A286) Anchornut(floating)LN29693 0.306 0.144 0.346 0.259
HelicoilLN9499madeof 0.184 0.152 0.333 0.278
bronzeCuSn6
NutLN9161(selflocking, 0.140 0.094 0.257 0.087
hexagonflanked)
LN29949M8(A286) Anchornut(floating)LN29693
HelicoilLN9499madeof 0.399 0.236 0.352 0.294
bronzeCuSn6
NutLN9161(selflocking,
hexagonflanked)
Someselflockingdevicesuseathreadcrosssectionthatisdeformedaftermanufacturinginordertoform
a slightly elliptical shape. When used for the first time, the elliptical shape is deformed to a circle. This
deformation is partially plastic, leaving a permanent deformation for the next use. Thus, the prevailing
torque reduces after the first use. Subsequent usage involves mainly elastic deformation, and thus the
prevailing torque reductions are small. The increase of the prevailing torque due to increased friction
coefficients (as discussed in Section 12.6.2) is negligible compared to the decrease due to plastic
1282H
deformation. Therefore, the choice of lubricant has only a minor effect on the prevailing torque after
multiplereuses.
Someexperimentaldataforprevailingtorquelossisshownin Figure124.Thedataindicatesthat,bythe
1283H
5thretightening,nofurtherplasticdeformationoccursinthelockingdevice(i.e.theprevailingtorqueis
generatedpurelybyelasticdeformations).Therefore,whenanalysingfastenerswithlockingdevicesthat
willbereused,theelasticrangeprevailingtorquesof Table124,shouldbeusedinplaceof Table62.
1284H 1285H
197
ECSSEHB3223A
16April2010
Figure124Prevailingtorquesduringaseriesofloosingandretighteningcyclesofa
LN29949M6fastenerandwithHelicoilLN9499madeofCuSn6
Table124PrevailingTorquesofLockingDevicesafterSeveralLooseningand
RetighteningCyclesforUnlubricatedFasteners
Fastener LockingDeviceType Size MP(Nm)
Type Min. Max. Mean
LN29949 Helicoil(screwlock)LN M4
9499bronze(not M5 0.20 0.90 0.69
cadmiumplated)
M6 1.20 2.00 1.63
M8 1.60 3.60 2.28
AnchornutLN29693 M4
M5
M6 0.40 1.60 0.87
M8
NutLN9161M M4
M5 0.90 1.60 1.15
M6 0.20 0.70 0.33
198
ECSSEHB3223A
16April2010
12.6.4 Recommended Practice for Fastener Resuse
218B
Twoapproachestodeterminingthenumberoftimesthatafastenercanbereusedarerecommended.The
first approach is to rely on experimental data in order to determine limits. The second approach, much
more accurate, is to measure the prevailing (also known as running or selflocking) torque during the
installationofeveryfastenerandaddittothespecifiedseatingtorquetocalculatetheinstallationtorque
that is finally applied. In this approach, the fastener is considered unusable if the prevailing torque is
outsidetherangespecifiedbythemanufacturer.
12.7 References
81B
12.1
1286H BICKFORDJ.H. UltrasonicControlofBoltPreload,HydrocarbonProcessing.,Jan
1982.
12.2
1287H BENNET J.K. and D. de Procedure for Deriving and Verifying Calibration Data for bolt
VILLIERS preload extensions using Raymond Bolt Gauge PDX 934 (BAe.
Doc.No.PRBAe0072)
12.3
1288H HUTCHINSA.R. Test Report Deriving and Verifying Calibration Data for bolt
preload extensions using Raymond Bolt Gauge PDX 934 (BAe.
DocNo.RPBAe0120)
12.4
1289H MORGANE. AssessmentofCalibrationTestsofRayhmondExtensometerPDX
934(BAe.Doc.No.TNBAe50oo882)
12.5
1290H YAMAMOTOE. Direct Stress Measurement by Ultrasound Proc. Ninth World
Conf.onHonDestnuctineTesting1979.
12.6
1291H DEPUTATJ. UltrasonicTechniqueforMeasuringStressinScrews.Proc.Ninth
WorldConf.onNonDestructualTesting,1979.
12.7
1292H BOBRENKOV.M. Ultrasonic Method of Measuring et a1 Stresses in Parts of
ThreadedJoints:SovietJournalofNonDestructualTesting1974.
12.8
1293H HEYMANJ.S.and Ultrasonic Measurement of Axial Stress: Journal for Testing and
CHERNE.J. Evaluation,September1982.
12.9
1294H SMITHJ.F. StressMeasurementandBoltTensioningbyUltrasonicMethods:
GREINERJ.D. JournalofMetals,July1980
12.10
1295H BAereport Examination of Factors Affecting Torque Tightening and the
199
ECSSEHB3223A
16April2010
GHFM/TST/SDec85 ProductionofDesignRecommendations.
12.11
1296H FISHERJ.W. GuidetoDesignCriteriaforBoltedandRivetedJoints
STRUIKJ.H.A. JohnWileyandSons1974
12.12
1297H BICKFORDJ.H. ThatInitialPreloadWhatHappenstoit?
MechanicalEngineering1983
12.13
1298H VDIRICHTLINIEN Systematiccalculationofhighdutyboltedjoints,VDIHandbuch
Konstruktion, December 1974 (1979 English translation by R.M.
Durham)
12.14
1299H BICKFORDJ.H. An introduction to the Design and Behaviour of Bolted Joints,
MarcelDekker
12.15
1300H Pfaff,H. Calculation of the preload at the yielding point of hyperelastic
tightenedscrews,Konstruktion47(1995Nr.7/8S.237/240)
200
ECSSEHB3223A
16April2010
13
Corrosion
14B
13.1 Overview
82B
Corrosion may be defined as the chemical reaction of a metallic material with its environment. The
products of this reaction maybe solid, liquid or gaseous and the physical and chemical natures of the
productsareimportantastheyofteninfluencethesubsequentrateofreaction.
Handbooks and reference charts are available, providing information on the behaviour of metals and
othermaterialsincertainwelldefinedenvironments.Unfortunatelythepreciseoperationalenvironment
is,inpractice,difficulttopredictandmayvaryconsiderably,e.g.forreusablespacecomponents.
Additionally, the method of fabrication and fastening can affect the susceptibility to, and rate of,
corrosion.
Thustoavoidcorrosionproblems,thefactorstobeconsideredbyadesignerare:
Theworkingenvironmentandexposuretoit
Theprobablereactivityofthematerialsinvolvedinfabrication
Thefeaturesofthedesignthatmaymodifythatreactivity
Figure131showsthetypesofcorrosionthathavebeenidentifiedasmostrelevanttospaceapplications.
1301H
The following sections briefly describe the various types of corrosion and some recommendations are
givenforovercomingcorrosionproblems.
201
ECSSEHB3223A
16April2010
Figure131FormsofCorrosion
202
ECSSEHB3223A
16April2010
13.2.1 Introduction
219B
Aqueouscorrosionprocessesgenerallyneedoxygenandwaterinintimatecontacttoachievemeasurable
corrosion rates. In space environments, both these substances are virtually nonexistent and hence
aqueouscorrosionprocessesmaybeessentiallyignored.
However during manufacture, test and storage the structure will experience earth environments with
highlevelsofoxygenandwater,thereforethecorrosionprocesscannotbeignored.
When dissimilar metals are joined, accelerated corrosion can occur at the interface due to galvanic
corrosion.Thiscorrosionoccursonametal(theanode)asaresultofcurrentflowfromtheanodetoaless
reactive (more noble) metal (the cathode), when they are in electrical contact and in the presence of an
electrolyte. The anode corrodes due to dissolution of the metal and the process is balanced (in most
practicalapplications)bythereductionofdissolvedoxygenintheelectrolyte.
13.2.2.1 Electrolyte
354B
Electrolytefactorsthathavemajorimplicationsforgalvaniccorrosionareasfollows:
a. Electrolytecomposition
b. Degreeofacidityoralkalinity
c. Electricalconductivity
Ingeneraltheseverityofcorrosionoftenincreaseswithincreasingelectricalconductivitysinceinpractical
situationshighconductivityisoftencausedbythepresenceofaggressiveionssuchaschlorideorbyacid
oralkali.
Formanysituationstherateofcorrosionisdependentupontherateofdiffusionofdissolvedoxygento
thecathodeandisthusproportionaltotheareaofthecathodemetalsurface.Thus,foraconstantareaof
cathode material, the amount of corrosion of the anode area is constant but the intensity of corrosion
increases as the area of the anodic metal is decreased. Since adverse area ratios may occur in threaded
fastener joints, the fastener metal should be of the same polarity as the clamped parts, or be relatively
cathodictothem.
Differences in corrosion potential can exist between coupled metals or alloys of nominally the same
composition.Thecoldworkingofanalloy/metaltendstomakeitmoreanodic.Bimetalliccorrosioncan
alsooccurbetweengenericalloyswithslightlydifferingcompositions.
203
ECSSEHB3223A
16April2010
13.2.3 Prevention of Bimetallic Corrosion
221B
The prevention or minimisation of galvanic corrosion is best addressed during the design stage (see
Reference 13.1).Whendissimilarmetalsarejoined,careshouldbetakentoprotecttheanodicmemberby
1302H
properelectricalinsulationofthejointorbyexcludingtheelectrolyteifthisisfeasible.
Table 131 lists metalsinthe order of their relativeactivity inan aqueous environment. The list begins
1303H
withthemoreactive(anodic)metalandproceedsdowntotheleastactive(cathodic)metalofthegalvanic
series.A galvanicseries pertains to a particularelectrolyte solution and hencefor each specific solution
encountered, a different series will apply. Usually a seawater environment is chosen as the basis for
comparisonsinceitisoneofthemostcorrosiveenvironmentsandisreadilyencountered.Galvanicseries
relationshipsareusefulasaguideforselectingthemetalsusedforajoint.Generallythecloseronemetal
istoanotherintheseries,themorecompatibletheywillbe.Inagalvaniccouplethemetalhigherinthe
seriesrepresentstheanodeandwillcorrodepreferentiallyintheenvironment.Itisthereforenecessaryto
protectmetalswidelyseparatedinthegalvanicseries,iftheyareintendedtobejoined.
Methodsofpreventingorminimisinggalvanicattackarebaseduponbreakingtheelectricalpathinthe
metallic or electrolytic parts of the system, by excluding oxygen from the electrolyte, or by sacrificial
corrosion. This strategy cannot be used if the galvanic couple needs to act as an electrical connection.
Electrical insulation can be achieved by interposing a gasket or washer between the dissimilar metals
priortoconnectingthem.Ifthecouplehastoactasanelectricalconnection,acompatiblemetallicgasket
orwashercanbeplacedbetweenthetwometalspriortofastening,orthecathodicmembercanbeplated
with a metal compatible with the anode. Jointing compounds applied to the contacting surfaces of
fasteners and lap joints do not normally provide electrical insulation, they are however very useful in
preventingwaterpenetratingthejointandhencepreventinggalvanicandcrevicecorrosion.
204
ECSSEHB3223A
16April2010
Table131GalvanicSeriesofMetalsinanAqueousEnvironment
MaterialCategory Variant
Mgalloy AZ31B
Beryllium(hotpressed)
Alalloy 2014T3
Alalloy 1160H14
Alalloy 7079T6
Alalloy 50520
Alalloy 5052Hl2
Alalloy 54560,H353
Alalloy 5052H32
Alalloy 11000
Alalloy 6061T6
Alalloy 7075T6 Anodi
Alalloy 1160H14
Alalloy 60610
Alalloy 20140
Alalloy 2024T4
Alalloy 5052H16
Stainlesssteel 430(active)
Stainlesssteel 410(active)
Copper(plated,castorwrought)
Nickel(plated)
Tantalum
StainlessSteel AM350(active)
Cathodi
StainlessSteel 301(active)
StainlessSteel 304(active)
StainlessSteel 177PH(active)
Tungsten
Copper110
Carpenter20
StainlessSteel (active)
StainlessSteel 321(active)
StainlessSteel 316(active)
205
ECSSEHB3223A
16April2010
MaterialCategory Variant
StainlessSteel 309(passive)
StainlessSteel 177PH(passive)
StainlessSteel 304(passive)
StainlessSteel 301(passive)
StainlessSteel 321(passive)
StainlessSteel 201(passive)
StainlessSteel AM355(active)
Carpenter20
StainlessSteel (passive)
StainlessSteel AM355(passive)
StainlessSteel A286(passive)
Titanium(annealed) 13V,11Cr,3Al
Titanium(solutiontreatedand
aged) 6Al,4v
Titanium(annealed) 6Al,4V
Titanium(solutiontreatedand
aged) 13V,11Cr,3Al
StainlessSteel Am350(passive)
Silver
Gold
Graphite
Surfacecoatingsareveryeffectiveinprotectingmetals.Bothmembersofthecoupleshouldbecoated,but
wherethisisnotpossiblethenthecathodicmetalshouldbecoatedasthisincreasestheeffectiveanode
arearatioandreducesthecorrosionintensity.
Nonmetallicmaterialsthatarejoinedtometalsshouldbetreatedwithcaution.Forexample,composite
materialsmaycontaincarbonorgraphite,whichbeingverynoble,mayresultincorrosionofthemetal.
Graphitepencilmarkingsonaluminiumaircraftstructureshavegivenrisetogalvaniccorrosionproblems
inthepast.Itisimportanttoensurethatthenonmetallicmaterialdoesnotcontainanyothercorrosive
agents.Inparticulartheyshouldbefreeofthefollowing:
a. Ionicsalts
b. Acidoralkalinematerials
c. Carbonormetallicparticles
d. Anabilitytowickorabsorbwater.
206
ECSSEHB3223A
16April2010
Gaskets used in RF shielding (typically a metallic loaded or metallic mesh supported in an elastomeric
medium) should be carefully screened to achieve compatibility with the mating surfaces, otherwise the
matingsurfacesandfastenersmayneedtobetreatedinsomeway.
13.3.1 Introduction
222B
StressCorrosionCracking(SCC)occurswherecertainalloysaresubjectedtoacontinuoustensilestressin
thepresenceofaspecificcorrosiveenvironment.Thefailuresappeartobeofabrittletypeevenwherethe
materialisrecognisedtobeductile.Surfacesintheregionofthecrackmaynotappearaffectedbygeneral
corrosionthatoftenmakesSCCmoredifficulttodiagnose.Timetofailureisamatterofminutesunder
severeconditions,oryearswhenconditionsarelesssevere.
SCCisoftenassociatedwithaerospacestructuresconstructedfromhighstrengthmaterialsandoperating
athighstresses.TheApolloProgramsufferedmanySCCfailuresandmuchwaslearnedaboutthedesign
of space structures as a result (see Reference 13.2). Subsequently NASA issued design criteria for
1304H
controllingSCC(Reference 13.3),tobemetbyallcontractorsforNASAflighthardware.Itcontainsalist
1305H
of materials that are resistant to SCC and also those which are most susceptible. A similar document
calledECSSQST7036MaterialSelectorforControllingStressCorrosionCrackingexistswithinECSS.
(Reference 13.4).
1306H
ThefollowingsubsectionsaimtohighlighttheaspectsofSCCrelevanttothreadedfastenerjointsandto
makedesignersawareofthemethodsavailabletoavoidorreducetheoccurrenceofSCC.
13.3.2.1 Overview
357B
TheprerequisitesforSCCareasfollows:
a. asusceptiblematerial
b. aspecificcorrosiveenvironment
c. atensilestress
Eachoftheseisdiscussedbelow.
The effect of alloy composition and metallurgical condition (grain size and orientation, distribution of
precipitates,etc.)onSCCissufficientlyunderstoodtospecifycertainpreferredalloysandtreatmentsto
minimisetheeffectofSCC(seeReference 13.3).
1307H
In high strength aluminium alloys the effects of heat treatment and cold work are significant in
determining resistance to SCC. For example when wrought aluminium products are manufactured by
rolling, extrusion or drawing the alloy microstructure becomes distorted into elongated grains in the
directionofworking.Thisisdesignatedthelongitudinaldirection,thewidthofthesheetisthetransverse
207
ECSSEHB3223A
16April2010
direction, and the thickness is designated the short transverse direction (as shown in Figure 132 for a
1308H
numberofsectionshapes).
Figure132GrainOrientationsinStandardWroughtForms
Material properties vary depending on the direction of testing with the longitudinal direction being
strongest.Theshorttransversedirectionisusuallytheweakestandasaresultismostpronetotheeffects
ofSCC.Itisthereforedesirabletodesigncomponentsthatarenotheavilystressedintheshorttransverse
direction. Figure 133 and Figure 134 illustrate examples of poor joint design that would lead to high
1309H 1310H
stressesintheshorttransversedirection.
208
ECSSEHB3223A
16April2010
Figure133AssemblyStressResultingfromMismatch
Figure134HighAssemblyStressesinShortTransverseDirection
The short transverse direction of a forged component is perpendicular to its parting plane as shown in
Figure 135. In this case any bush or fastener with an interference fit in the bore indicated is likely to
1311H
acceleratefailurebySCC.
209
ECSSEHB3223A
16April2010
Figure135ExamplesofTensileStressesinShortTransverseDirectionResulting
fromAssembly
Heat treatments have been developed to reduce the susceptibility of high strength aluminium alloys to
SCCandtheseshouldbespecifiedwherepossible.
When specifying alloys for use on space systems ECSSQST7036 on prevention of SCC is applicable
(Reference 13.4).ThisdocumentcollectsinatablethealloyswithhighresistancetoSCCthatcanbeused,
1312H
and prescribes their conditions of use. Where an alloy is not listed and SCC data is not available, it is
necessary to conduct suitable tests for SCC resistance, such as those given by ECSSQST7037 (see
Reference 13.6).TheresultsshouldbesubmittedforapprovalonaStressCorrosionEvaluationForm.A
1313H
similarsituationexistsformaterialapprovaltoNASAstandards.
EnvironmentsthatgiverisetoSCCareoftenspecifictoparticularalloys,thusmakingSCCverydifficult
to predict unless tests have been conducted simulating the operating conditions or failures have been
notedpreviouslyundersimilarconditions.
As the corrosive environment is essential for SCC to occur, the designer should ensure that approved
protectiveanticorrosioncoatingsarespecified.
The effect of applied tensile stress on SCC varies with materials. In some cases a threshold stress
concentration factor Kiscc can be specified, below which failure by SCC will not occur (see Subsection
11.3.4). In other cases, SCC occurs at all values of applied tensile stress, the time to failure being
1314H
dependentonthestresslevel.
Tensile stresses in joint assemblies that can give rise to SCC may be imposed by several sources, in
additiontothenormalworkingstressesinthesystem.Misalignmentorgapsbetweenjointfacescancause
highsurfacetensilestresseswhichwhencombinedwithanunfavourablegrainorientationcanresultin
rapid SCC failures. Such failures can be avoided by ensuring that gaps are properly shimmed, and the
alignmentofassembliesischecked.
210
ECSSEHB3223A
16April2010
Fastenerswithsubstantialinterferencefitscanimposehightensilestressesintheflangestransversegrain
direction,leadingtolongitudinalcracks.Careshouldbetakentospecifyalloysthatarehighlyresistantto
SCCintheseregions.
In some cases, manufacturing processes may leave components with residual surface tensile stresses.
Examples of such processes are bending, tube drawing, stretch forming, electro discharge machining
(EDM)andsomecuttingoperations.Anindicationofwhatlevelofresidualstressescanbeimpartedby
these operations is given in Reference 13.5. Heat treatment processes may also leave residual stresses,
1315H
especially for components with complex shapes. All residual tensile stresses should be reduced or
eliminatedbythermaltreatmentsorbymechanicalmeans.
Compressive surface stresses can be introduced in many components by shot peening. This process is
veryusefulforcontrollingSCCincriticalparts.
13.4.1 Introduction
224B
Acceleratedcorrosiveattackisoftenobservedincrevicesthatareexposedtowetconditions.Assemblies
joined together by fasteners are particularly prone to this type of corrosion due to the difficulty in
avoiding crevices between the joint elements. However, it is possible to minimise the problem by
preventingmoisturefrompenetratingthecrevice.
Itisrecommendedtoavoid,wherepossible,crevicesindesign.Iftheycannotbeavoided,theyshouldbe
coveredwithacontinuouspaintfilm.Threadsealantsareveryusefulincombatingcrevicecorrosionand
inmostcasesagoodpaintfilmwillpreventtheingressofwater.
Joints should be carefully cleaned and dried before assembly since even small amounts of debris can
prevent mating faces from contacting, resulting in crevice conditions. Similarly any moisture trapped
betweenjointfacesislikelytoinitiatecorrosion.
13.5.1 Introduction
226B
Pitting corrosion may be thought of as a form of localised corrosion that occurs on one part of a metal
surface at a much higher rate than over the rest of the surface. Pitting corrosion is associated with the
breakdownofasurfacefilmandoftenoccursoncompletelyflatsurfaces.Ifthesurfacefilmiscathodic,an
areawithoutthefilmwillactasasmallanode,thussufferingintensecorrosionthatleadstopitting.
Pits may also be nucleated at points determined not only by faults in the surface film, but also at sites
determinedbytheunderlyingmetal.Suchsitesmayarisefromalloyheterogeneitiesinthesurface,orbe
associated with grain or phase boundaries.Solid nonmetallicinclusions fromprocessing procedures or
fromimpurities,e.g.sulphidesinstainlesssteels,mayalsoprovideinitiationsites.
211
ECSSEHB3223A
16April2010
13.5.2 Alloy Susceptibility
227B
Stainlesssteelspossessexcellentresistancetopittingcorrosioninoxidisingatmospheres.However,inthe
absenceofoxygentheprotectiveoxidefilmmaybreakdownandpittingcorrosioncanoccurveryrapidly.
13.5.2.2 Aluminium
362B
Inmostnaturalenvironments,aluminiumanditsalloysgivesatisfactoryresistancetopittingcorrosion.
However,athighandlowpHvalues,intheabsenceofoxygen,orinthepresenceoffilmdestructiveions,
theprotectingoxidefilmmaybreakdownwitharesultantpittingattack.InAlZnMgalloysthismaylead
tointergranularcorrosionandasignificantlossofstrength.
13.5.2.3 Titanium
363B
Thismetalanditsalloysareextremelyresistanttopittingcorrosion.
Theincidenceofpittingcorrosioncanbereducedbyselectionofmaterialsthatexhibitahighresistanceto
thisformofcorrosion.
13.6 References
87B
13.1
1316H T.KRoss Metal Corrosion, published by the Design Council, the British
Standards Institution and the Council of Engineering
InstitutionsbyOxfordUniversityPress.
13.2
1317H R.E.Johnson Apollo Experience Report The Problem of Stress Corrosion
Cracking,NASATND711,March1973.
13.3
1318H NASA MSFC DesignCriteriaforControllingStressCorrosionCracking.
SPEC 522A, Nov
1977
13.4
1319H ECSSQST7036 Space product assurance Material Selection for Controlling
StressCorrosionCracking
13.5
1320H ATBainbridge Residual Stresses Arising from Machining and Fabrication,
AGARDCPNo53,1970.
13.6
1321H ECSSQST7037 SpaceproductassuranceDeterminationofthesusceptibilityof
metalstoStressCorrosionCracking
212
ECSSEHB3223A
16April2010
14
Lubricants for Space Use
15B
14.1 Introduction
88B
Thenearlytotalabsenceofairaroundandwithinspacevehicleshastwomajorimplicationsfortribology.
Firstly, oxygen and water vapour are not on hand to repair damaged surfaces, so strong adhesion
(welding and galling) between clean degreased materials becomes inevitable if they are unprotected.
Secondly,aliquidphaselubricant(e.g.oilorgrease)wouldquicklyvolatiliseorboilaway,eitherdirectly
tospaceorcontaminatingthespacecraft,unlessthelubricanthassufficientlylowvapourpressureand/or
is suitably sealed to retard molecular effusion. Thus, the space environment impose the use of
lubricantsthatneithervolatilisenorcreep,andwhicharenotinfluencedbytemperatureextremes.
Reference 14.1 contains guidelines for space materials selection. It deals with lubricants in general and
1322H
listsfourlubricantsthataresuitableforspaceapplications.
Table141providesalistofrecommendedlubricants.
1323H
213
ECSSEHB3223A
16April2010
Table141RecommendedLubricantsforSpaceApplications
TradeName Chemical TypeofProduct Manufacturer
Composition
FOMBLINYVAC3 Perflouropolyether Greaselubricant Montedison,Milan,
Italy
BRAYCO Perflourinatedpoly Oil BrayOilcompany,
MICRONIC815Z ether Californis,USA
MoS2 Molybdenum Lubricant Several(Dow
MolycoteZ,Moly disulphide CorningBelgium,
PaulITC,Lubri K.T.PaulProducts
Bond U.K.,ElectrofilemInc.
USA.etc.)
BRAYCOTE3L Perflourinatedpoly Greaselubricant BrayOilcompany,
38RP ether Californis,USA
DICRONITE ModifiedTungsten Lubricant RotaryComponents
Disulphide(WS2) Inc.CaliforniaUSA
APIEZONL Hydrocarbon Greaselubricant ShellChemie
DenHaagNL
KINEL5518 Polymide/PTEE Thermosetting,self RhonePoulene
lubricatingresin ParisF
FOMBLINZ25 Perflouralkylether Oil Montedison,Milan,
Italy
P.T.F.E Polytetrafluorethylene Thermoplastic Several(DuPontUS,
(Teflon,Halon, HoechstD,
Fluon, MontecatiniI,etc.)
Hostafluon)
Silver is primarily a lubricant for high temperature applications on corrosion resistant fasteners in steel
andotherstructureswheregalvaniccorrosionisnotaproblem.
Itisalsousedbecauseitdoesnotsublimate.ThespecificationisusuallyAMS2410.Othercoatingssuchas
Electrofilmarereallylubricantsandprovideminimalcorrosionprotection.SoftmetalslikeLead(Pb)can
beshearedeasily,sothatincertaincircumstancestheycanalsobeusedassolidlubricants.However,they
will not support heavy loads without deforming, so they are used as thin coatings on stronger metal
surfaces.Softmetalfilmsaremainlyusedaslubricantsonlyinavacuum.
Fastener qualification data has been generated on the use of Ion Vapour Deposited (IVD) aluminium
(IVADISE)asareplacementforcadmiumingeneralengineeringuse(Reference 144.2). 1324H
Theconclusionscanbesummarisedasfollows:
214
ECSSEHB3223A
16April2010
a. IVDaluminiumdoesnotproduceanydetrimentaleffectsonthemechanicalproperties.
b. Thecoefficientoffrictionofaluminiumishigherthancadmium,thereforehigherinstallationforces
are necessary. These higher values, however, are within the working ranges presently used for
cadmium in most cases. A closer attention to the type of lubricants used can be necessary for
interferencefitfasteners.
With a silverplated A286 nut, an IVD coated titanium bolt, and using EVERLUBE 812 as a lubricant, a
consistently low coefficient of friction can be obtained for multiple installations. Kalgard 2240 (an
aluminiumrichpaint)canbealsousedinsteadoftheIVADIZEcoating(Reference4.3).
ProbablythemostsignificantadditiontospacelubricantsisapolyfluoralkylethercalledFomblinZ25.The
fluid has the very high viscosity index of 345 and a vapour pressure that is suitably low at room
temperature.
TheBrayOilCointheUSAhavedistilledtheItalianrawstock(FomblinOil)tomakeBray815Z,which
has been flown on a number of USA satellites mostly in the form of a PTFE thickened grease, Bray
3L38RP.
Molybdenum disulphide (MoS2) in various formulations is the most extensively used lubricant. It
possessesexcellentantigallingpropertiesandlowfriction.InvacuumMoS2performsevenbetterthanin
airbecausethereisnooxidation.Also,MoS2willwithstandover1000Cinvacuumbeforeitsublimes.
ApplicationofMoS2canbebyanumberofmeans;
a. asaspraybondedfilm,
b. usingapolymericorinorganiccarrier,
c. asaburnishedfilmwherethehardedgedplateletsareforcedintothesurface,or
asasputteredfilmappliedbyradiofrequency(RF).
TheRFsputteringmethodoffersahighlyreproduciblewayofobtainingadherentandthinfilms.
Processes
The behaviour of any tribological system is governed by a large number of factors, many of which are
difficult to control; material properties, both macro and micro, surface condition, presence of micro
quantities of contaminants, system geometry, speed, load, duty cycle are only some of the variables. In
consequencetherulestobemetbyanylubricationprocesstobeacceptableforspaceareasfollows:
a. In order to ensure consistent repeatability of performance, oOnly fully codified and documented
lubricationprocessesareusedforspaceapplication.
215
ECSSEHB3223A
16April2010
b. Onlylubricantsthatareapprovedandvalidatedforspacetoarecognisedspecification,andsource
traceable,areused.
c. The test programme to determine the performance of the lubrication system reproduces all the
operationalconditionsofthedutycycle,theenvironmentandthelifethatitwillexperienceinthe
application.
Theuseofacommercialsystemforaspaceapplicationisacceptableonlyifitfulfilsthesethreerules..4.
14.7 References94B
14.1
1325H ANON Guidelinesforspacematerialsselection,ESAPSS07(QRM07)Issue
5,July1979.
14.2
1326H J. Newnham, Evaluation of Titanium bolts and coated with Ivadize Aluminium
B. Hassell, P. andKalgard2240Cintorquetensiontests.
Betiam
SPSTechnologies,Inc.,NAASLab,Ref.,No.52541979.
14.3
1327H MCAIR M & R&DReportNo.118,6thFebruary,1975.
PD
216
ECSSEHB3223A
16April2010
15
Manufacturing Quality Control
16B
15.1 Introduction
95B
It is essential that quality is builtin at the design stage. If it is not, no amount of subsequent quality
activitywillbeentirelysatisfactory.
The overall quality of a joint is a function of its design quality and manufacturing quality. The former
includestheanalyticalaspectsofjointdesignandisdealtwithinsomedetailintheearlierthisguidelines
document. The latter includes performance variations, which occur as a result of production tolerances
andiscoveredinthissection.
Specific overall requirements are imposed by agencies (such as NASA, ESA, CNES and MILITARY).
These requirements relate to the need to address such aspects as safety factors for the Space
Transportation System. An example of this is the requirement to use materials which have been
previouslyqualifiedasnonstresscorrosive.Theinformationconcerningsuchaspectsistobefoundinthe
respectiveSpaceAgencysupportingdocumentation,processes,specificationorprojectrequirements.
ForprojectsconductedwithintheEuropeanSpaceAgencyasystemofECSSdocumentsexistswhichare
mandatorywhencalledforinthecontractrequirements.
Quality Assurance, or as it is more commonly known in space projects, Product Assurance, has the
responsibilitytoensurethatdesignershaverecognisedrequirementsofthetypeindicatedaboveandthat
the necessary controls and verification have been conducted to ensure compliance. This ensures the
traceability of materials from their source and that the specified testing and inspection is carried out.
Testingcouldincludehardness,nondestructivetestsandpreloadingoffasteners.Inspectionincludesall
dimensionaltolerances.Additionally,allmaterialsusedinspacecrafthavetobelistedandreviewedfor
acceptability.
The controls imposed to achieve these objectives are primarily associated with the detail of planning,
review,inspectionandthatrecordingofdatathatgivesfinaltraceabilityofalltheinformationintheevent
ofanyfailure.
217
ECSSEHB3223A
16April2010
15.3.1 Overview
229B
When designing a threaded fastener joint, the designer should consider all production variations that
mightinfluencethejointsperformance.Thesevariationscomeunderthefollowingbroadcategories:
ProcessVariations
MaterialVariations
Tolerancing
The designer should specify the acceptable variation of the above parameters in the engineering
drawings,alsocitingtherelevantcompany,nationalorinternationalstandards.
Production testing and inspection is an important part of Quality Control and it provides feedback on
achievement and limits variations. The designer should therefore consider the necessary testing and
inspectionandfacilitatethis(asfaraspossible)withinthedesign.
Theperformanceofthethreadedfastenerjointsisaffectedbythefollowing:
a. Dimensionaltolerances,e.g.fastener/holefits,holepositionaltolerances.
b. Surfacefinishonthematingpartofthedesign.
c. Preloadvariations.
d. Threadprofileanddimensionaltolerances.
e. Lubricationconditions.
The above variations are limited by the designer specifying an acceptable level on the drawings or by
invokingstandards.
Performanceofthejointwillbeaffectedbythefollowingmaterialvariations:
a. Ultimatestrength(compressiveandtensile)
b. Proofstrength(compressiveandtensile)
c. Moduliofthematerial
d. PoissonRatio
e. Frictioncharacteristics
f. Corrosioncharacteristics
218
ECSSEHB3223A
16April2010
15.3.4 Tolerancing
232B
Tolerancesmayaffectthejointcharacteristicssuchasitsloadcarryingability,stiffnessandbacklash.Itis
likelythatthelargerthetolerances,thegreaterwillbethejointstrengthvariation.
When considering the effect of manufacturing tolerances it is necessary to assume some statistical
distributionofthevariations.Ifalltoleranceswereassumedtobeattheirworstpossiblevalue,thenthe
design would be far too pessimistic. However, there may be cases where this has been designed in
deliberatelyandsoshouldnotbeoverlookedcompletely.
Whenusingstatisticalanalysisitisessentialtorecognisethedistributionofthevariation.Manufacturing
variations can generally be classified as Gaussian. However, SkewedGaussian or Rectangular
distributionsmaybemoresuitableforcertainmanufacturingmethods.Inspectionandtestingwillaffect
thedistributionbytruncatingtheendsofthedistribution.
The designer should obtain information on the variation of important joint characteristics and examine
testresultsinorderthathemayhaveconfidenceinhisstatisticalanalysis.Thetolerancesspecifiedshould
take account of the variation in the controlled characteristic. In order to optimise production yield and
minimisecosts,thetolerancesshouldbeincreasedtothemaximumconsistentwiththeperformancetobe
achieved.
15.4 References
98B
15.1
1328H ECSSQST70 Space product assurance Materials, mechanical parts and
processes
15.2
1329H ECSSQST7036 Space product assurance Material selection for controlling
stresscorrosioncracking
15.3
1330H ECSSQST7037 SpaceproductassuranceDeterminationofthesusceptibilityof
metalstostresscorrosioncracking
15.4
1331H ECSSQST7046 SpaceproductassuranceRequirementsformanufacturingand
procurementofthreadedfasteners
219
ECSSEHB3223A
16April2010
16
Joint Validation by Testing
17B
16.1 Introduction
99B
Successful design is usually achieved through the combination of theoretical analysis and practical
testing.Thelevelofconfidenceintheoreticalworkaloneislimitedbytheaccuracyofthedatabeingused
and the accuracy of the theoretical method. Physical testing overcomes this problem but it is usually
impossible to explore the limits of all potential variations. Close agreement between theoretical and
practical results gives the designer high confidence that the joints behaviour is understood, that the
theoryiscorrectandthatasuccessfuldesignwillresult.
Thissectionaimstogivethedesignerguidelinesforthetestingofthreadedfastenerjoints.Itisnotspecific
since the type of validation will depend on the type of joint and the circumstances in which it is being
designed.
Therearetwobroadcategoriesoftesting:
a. Developmenttesting,and
b. Productiontesting.
During development testing an exhaustive set of tests is recommended, dependent on the type of joint
anditsapplication.Thegeneralaimofdevelopmenttestingistoconfirmtheunderstandingofhowthe
joint works (e.g. to confirm the theory applied for its design). Development testing also includes long
termeffectssuchascorrosionandfatigue.
Production testing is much more restrictive and may include every joint or be limited to testing of
samples.Theobjectiveistodeterminewhetherthejointhasbeenmanufacturedandassembledcorrectly,
andmadefromthecorrectmaterials.Theamountofproductiontestingshoulddependonhowgreatthe
riskoffailureis.
220
ECSSEHB3223A
16April2010
16.3.1 Overview
233B
Developmenttestingaimstoconfirmthetheoreticalanalysiscarriedoutonthejoint.Thisalsoappliesto
longterm characteristics such as fatigue. All joints show a statistical variation of performance, and it is
importanttotakethisintoaccountwhenobtainingapprovalbasedontestresults.Developmenttesting
canincludeanyorallofthefollowingapproaches:
a. Alargenumberofsamplessimplytestedundernormalloadingconditions.
b. A small number of samples simply tested but at loads enhanced in relation to the number of
samples.
c. Asmallnumberofsamplestestedingreatdetailsuchthattheinternalfunctioningofthejointmay
beunderstood.
Theaboveshouldincludebothstaticanddynamictests(ifappropriate)andalsolongtermeffects.
Whenatestismadeonasinglespecimenoranumberofrandomlychosenspecimens,theresultsshould
showanextramarginofstrength,ortestfactor,overthedesignvaluetoallowforthepossibilitythatthe
chosenspecimenswerestrongerthanaverageandalsotoensurethattheweakestspecimenproducedhas
acceptablestrength.Pastexperiencehasshownthatforconventionalmetallicstructuresthestrengthcan
be expressed as a normal distribution with a specified coefficient of variation (the ratio of standard
deviation to the mean). This distribution can be used to deduce the magnitude of the test factor that
shouldbeusedforagivennumberoftestspecimens,inordertoestablishwithanacceptableprobability
thattheweakestspecimenwillhavetherequiredstrength.
A coefficient of variation of 0.03 has been established for joints made to aircraft standards from the
followingmaterials:
AluminiumAlloy
Titanium
Steel
Ifthecoefficientofvariationisnotknown,testswillbenecessarytoestablishavalue.
Theprobabilitythatthemeanstrengthrealisedontestdividedbythetestfactorislessthantherequired
designstrengthgivenby,
S [16.3.1]
P P mt S req
Ft
where P is, Smt is the mean strength found during testing, Ft is the test factor, and Sreq is the required
strength.
Table15.3.1showsthetestfactorstobeusedtogivearangeofvaluesofPunderforarangeofvariation
coefficientsusingEquation [16.3.1].
1332H
221
ECSSEHB3223A
16April2010
Table161TestFactorstobeusedforGivenProbabilityofFailure
Number Probability CoefficientofVariation(V)
ofTests ofFailure .03 .05 .07 .10 .20
(P)
1Test 1in10 1.056 1.095 1.136 1.200 1.452
1in100 1.104 1.180 1.262 1.400 2.069
1in1000 1.141 1.247 1.366 1.577 2.890
1in10000 1.172 1.276 1.461 1.749 4.241
3Tests 1in10 1.046 1.078 1.112 1.167 1.385
1in100 1.086 1.150 1.220 1.339 1.943
1in1000 1.117 1.207 1.310 1.495 2.716
1in10000 1.143 1.257 1.392 1.650 4.023
Table161indicatesthatifonlyonetestcanbedoneonajointwithacoefficientofvariationof0.10and
1333H
theprobabilityofjointfailureinserviceistobebetterthan1in1000thenthetestfactortobeusedis1.577.
However,ifthreetestsweredoneagainstthesamerequirementthenthetestfactortobeusedwouldbe
1.495.
Thealternativemethodtotherelativelystraightforwardtestsdescribedaboveisundertakingtestingthat
providesdetailedknowledgeofhowthejointworks.Inthisapproachtestsareconductedwhichprovide
detailedknowledgeoftheloads,stressesandstrainswithinthejointasafunctionoftheexternalapplied
loads.Theseresultsarecomparedtoanequallydetailedtheoreticalanalysisofthejointusingasuitable
method (such as the Finite Element Method) to ensure that measurements are being made in the most
relevantareas.Withthisinformationthemaximumloadsappliedinsidethejointcanbecomparedtothe
basicmaterialcharacteristics.
16.3.3.1 Overview
364B
The environment to be withstood by a joint should be considered very carefully before specifying the
developmenttestprogramme.Ingeneral,higherloadsareappliedtothestructurewhileitisinabenign
environment,whereaslowerloadsareappliedwhilstthereareextremesoftemperatureandtemperature
gradient.Theeffectsofsuchtemperaturesandtemperaturegradientsshouldbeinvestigatedduringthe
testing, especially if dissimilar materials are involved. Care should be taken to ensure that the joint is
loadedinamannercompatiblewiththerealapplication.
Thefollowingparagraphsdescribethemostimportantcategoriesofbenignenvironmenttests.
Aloaddeflectioncurvewillprovideknowledgeonthefollowing:
Stiffnessofthejoint
Hysteresischaracteristics
222
ECSSEHB3223A
16April2010
Backlash
Loadsharing
Failuremechanism
Ultimate strength tests are used to prove the suitability of the joint. After failure, the joint should be
examinedcarefullytodeterminethemodeoffailure.
Fatiguetestingisnormallyonlyberelevantonreuseablespaceequipment.Itprovidesinformationinthe
followingareas:
FatigueLife
VibrationLoosening
Crackgrowth
Fretting
Corrosion may considerately alter the characteristics of a joint both in terms of its strength and life.
Corrosion testing in accordance with the relevant environmental conditions should be considered. The
effectsofthecorrosionshouldbeobservedandmeasured.
Oneoftheobjectivesoftestingatextremetemperatureconditionsistoreproducetheappropriatematerial
characteristicsaswellasthermalstrainsandstress.Hence,notonlyshouldtemperaturesbesimulatedbut
alsothermalgradients.Duetothedifficultiesandexpense,thistestingshouldbeminimised.Underthis
condition,informationshouldbederivedonjointstrengthandstiffnessandvariationfromsimilartests
carriedoutatnormalambientshouldbenoted.
It is unlikely that individual joints will be production tested. It is far more likely that the joints will be
tested as part of an overall spacecraft structural test. Obviously, normal quality control procedures will
aim to ensure that the production joint resembles the joint development tested within expected
productiontolerances.
Whenaproductiontestprocedureisused,thereisnoneedtouseatestfactor.Itensuresfreedomforeach
partfrombothproofandultimatefailuresuptotheleveloftheproductiontestload.Theproductiontest
load specified should be set at the specified proof load where the component is designed by the proof
criterion.Wherethecomponentisdesignedbytheultimatecriteriontheproductiontestlevelshouldbe
setatavaluehigherthanthespecifiedproofload.Thisshouldonlybedonewhentheproof/ultimateratio
ishigherthantheratioofthespecifiedproofandultimatefactors.
Theproposedmethodofproductiontestingshouldbetakenintoaccountwhendesigningthejoint.
223
ECSSEHB3223A
16April2010
Annex A
Recommended Starting Values for Seating
Torque
224
ECSSEHB3223A
16April2010
A.2 References
A.1 MBBERNO EURECAProject,Doc:ERC12124HBER001,1986.
A.2 DaimlerBenz COLUMBUSProject,Doc:HB1213800002,1997.
Aerospace
225
ECSSEHB3223A
16April2010
Annex B
Measured Friction Coefficients of Fasteners
TableB1FrictionCoefficientsofFastenersusedinEuropeanSpaceIndustry
UnlubricatedFasteners
Fastener Typeofnutorthreadedhole th ,max th ,min uh,max uh,min [
[4] 4]
M4, Anchornut(floating),LN29693
LN29949 Helicoil,bronzeCuSn6,LN9499 [1]
(A286) Nut,LN9161
M5, Anchornut(floating),LN29693 0.199 0.129 0.294 0.226
LN29949 Helicoil,bronzeCuSn6,LN9499[1] 0.176 0.118 0.335 0.222
(A286) Nut,LN9161 0.112 0.046 0.120 0.081
M6, Anchornut(floating),LN29693 0.176 0.086 0.296 0.179
LN29949 Helicoil,bronzeCuSn6,LN9499 [1] 0.119 0.062 0.302 0.222
(A286) Nut,LN9161 0.140 0.074 0.140 0.069
M8, Anchornut(floating),LN29693
LN29949 Helicoil,bronzeCuSn6,LN9499[1] 0.162 0.131 0.335 0.268
(A286) Nut,LN9161
Ti(3.7164) SteelNut(1.4944.4),oxsilverplated,LN 0.26 0.10 [3]
LN29950,ox 9161
blank[2] SteelInsert(1.4301),LN9499 0.34 0.18
SteelCaptiveNut(1.4301),LN9499 0.34 0.18
NOTE1:LN9499isonlyapplicableforCREShelicoils.ForaerospacejointsDIN655361shouldbeused
NOTE2:Steel(1.4944.4)ENN399washerwasusedwiththisfastener
NOTE3:Nodataavailable
NOTE4:FlangematerialAA7075T7351
226
ECSSEHB3223A
16April2010
Annex C
Typical Friction Coefficients for Joint
Materials
TableC1TypicalthreadandUnderheadFrictionCoefficients
Friction Rangefor Selectionoftypicalexamples
coefficient thanduh
class Materialsurfaces Lubricants
A 0.04to0.10 metallicallybright solidlubricants,suchas
blackoxide MoS2,graphite,PTFE,PA,PE,
phosphated PI
galvaniccoatingssuchas inlubricatingvarnishes,astop
coats
Zn,Zn/Fe,Zn/Ni
orinpastes;
Zinclaminatedcoatings
liquefiedwax
waxdispersions
B 0.08to0.16 metallicallybright solidlubricants,suchas
blackoxide MoS2,graphite,PTFE,PA,PE,
phosphated PI
galvaniccoatingssuchas inlubricatingvarnishes,astop
coats
Zn,Zn/Fe,Zn/Ni
orinpastes;
Zinclaminatedcoatings
LIQUEFIED WAX;
AlandMgalloys
waxdispersions,greases;
oils;deliverystate
hotgalvanized MoS2;graphite;
waxdispersions
organiccoatings withintegratedsolidlubricant
orwaxdispersion
austeniticsteel solidlubricantsorwaxes;pastes
227
ECSSEHB3223A
16April2010
TableC2FrictionCoefficientsforCommonJointMaterials(fromReference8.3)
Materialcombination Staticfrictioncoefficient
Dry Lubricated
Steelsteel/caststeel 0,1to0,23 0,07to0,12
Steelgraycastiron 0,12to0,24 0,06to0,1
Graycastirongraycastiron 0,15to0,3 0,2
Bronzesteel 0,12to0,28 0,18
Graycastironbronze 0,28 0,15to0,2
Steelcopperalloy 0,07
Steelaluminumalloy 0,1to0,28 0,05to0,18
Aluminumaluminum 0,21
NOTE:TabletakenfromReferenceVDI
228
ECSSEHB3223A
16April2010
Annex D
Material Data of Fasteners Typically Used in
Aerospace Industry
D.1 Overview
Ashortcompendiumofmaterialdataneededforboltanalysesofthemostcommonfastenermaterialsin
spaceindustryisgiveninthetablesbelow.
229
ECSSEHB3223A
16April2010
TableD2StainlesssteelA270andA280
MaterialSpecification A2
Alternativenotationsforthismaterial 1.4301,1.4541 AISI301304,AISI321
Temperature Roomtemperature
PropertyReferenceSource DIN267Teil11
Condition A270 A280
MaterialBasis A B S A B S
tensileyield [N/mm] 450 600
strengthy
tensileultimate [N/mm] 700 800
strengthu
shearultimate [N/mm] 420 480
strengthu
E(Youngs [N/mm] 193000200000
modulus)
(Poissonratio) [] 0.29
(density[g/cm]) [g/cm] 8.00
Fastenertypesmadeofthismaterial DIN912,DIN7991
TableD3Stainlesssteelcustom455
MaterialSpecification custom455
Alternativenotationsforthismaterial
Temperature Roomtemperature
PropertyReferenceSource MILHDBK5H
Condition H1000
MaterialBasis A B S A B S
tensileyield [N/mm] 1345
strengthy
tensileultimate [N/mm] 1413
strengthu
shearultimate [N/mm] 850
strengthu
E(Youngsmodulus) [N/mm] 200000
(Poissonratio) [] 0.30
(density[g/cm]) [g/cm] 7.76
Fastenertypesmadeofthismaterial
230
ECSSEHB3223A
16April2010
TableD4StainlesssteelPH138Mo
MaterialSpecification PH138Mo
lternativenotationsforthismaterial 1.4534
Temperature Roomtemperature
PropertyReferenceSource MILHDBK5H
Condition H950 H1000
MaterialBasis A B S A B S
(Poissonratio) [] 0.28
231
ECSSEHB3223A
16April2010
232
ECSSEHB3223A
16April2010
TableD6MP35N
MaterialSpecification MP35N
Temperature Roomtemperature
PropertyReferenceSource MILHDBK5H
Condition Solutiontreated,colddrawnandaged
MaterialBasis A B S A B S
tensileyieldstrength [N/mm] 1586
y
tensileultimate [N/mm] 1793
strengthu
shearultimate [N/mm] 1000
strengthu
E(Youngs [N/mm] 234500
modulus)
(Poissonratio) []
(density[g/cm]) [g/cm] 8.26
Fastenertypesmadeofthismaterial
233
ECSSEHB3223A
16April2010
D.5 References
D.1 GermanAviationMaterial
Handbook
D.2 DIN267Teil11
D.3 MILHDBK5H
234
ECSSEHB3223A
16April2010
The documents listed below are of general interest for the field of threaded fasteners and/or were
referencesbypreviousversionsofthishandbook.
235
ECSSEHB3223A
16April2010
JUNKERO.H. TheBoltedJoint:EconomyofDesignThroughImprovedAnalysis
andAssemblyMethodsProc.I.Mech.E.Vol.198B,No.14.,1984
WALLACEP.W.
BROBERGH. SkrurfrbandDimensioneringMonteringIufresultat82611,
July1983
CHERRYJ. A Finite Element Analysis of the Pressure Distribution in Bolted
Joints,BritishAerospacereportBT21927,June1987.
HERTELW,PAULW, InsertDesignHandbook
WAGNERD ERN3442/7728IssA.
RAYMONDL. Infinite Life High Strength Bolts. Assembly Engineering. April
1975.
KELLERMANR Bolts for light weight construction made of ultra high strength
steelsandtitaniumalloys.
TURLACHG.
BROEKD. ElementryEngineeringFractureMechanics
Norrdhof(Netherlands)1974.
DEKONINGA.U., Assessment of 3D Stress Intensity Factor Distributions for Nut
SupportedThreadedRodsandBolt/NutAssemblies
LOF,C.J.,
NLRCR96692L,NationalAerospaceLaboratory(NLR),1996.
SCHRA,L
MCICREPORTS CracksatStructuralHolesMarch1975.
MCICHB01 Damage Tolerant Design Handbook A compilation of Fracture
and CrackGrowth Data for HighStrength Alloys. Metals and
CeramicsInformationCentre,Battelle.
ROOKED.P. CompendiumofStressIntensityFactors
&CARTWRIGHTD.J. HerMajestysStationaryOffice
1976.
AvP32,AvP932,AvP970andtheirsuccessors.
DesignCriteriaforControllingStress
NASA CorrosionCrackingMSFCSPEC522A
November1977.
236