ECSS E HB 32 20 - Part2A PDF

ECSS-E-HB-32-20 Part 2A
20 March 2011
Space engineering
Structural materials handbook -
Part 2: Design calculation methods
and general design aspects
NOTE:
Thispdffiledoesnotcontainautomaticcrossreferences.Tomakeuseof
thecrossreferencespleaseusetheMSWordversionofthisdocument.
ECSS Secretariat
ESA-ESTEC
Requirements & Standards Division
Noordwijk, The Netherlands
ECSSEHB3220Part2A
20March2011
Foreword
ThisHandbookisonedocumentoftheseriesofECSSDocumentsintendedtobeusedassupporting
material for ECSS Standards in space projects and applications. ECSS is a cooperative effort of the
EuropeanSpaceAgency,nationalspaceagenciesandEuropeanindustryassociationsforthepurpose
ofdevelopingandmaintainingcommonstandards.
This handbook has been prepared by the ECSSEHB3230 Working Group, reviewed by the ECSS
ExecutiveSecretariatandapprovedbytheECSSTechnicalAuthority.
Disclaimer
ECSSdoesnotprovideanywarrantywhatsoever,whetherexpressed,implied,orstatutory,including,
butnotlimitedto,anywarrantyofmerchantabilityorfitnessforaparticularpurposeoranywarranty
that the contents of the item are errorfree. In no respect shall ECSS incur any liability for any
damages,including,butnotlimitedto,direct,indirect,special,orconsequentialdamagesarisingout
of,resultingfrom,orinanywayconnectedtotheuseofthisdocument,whetherornotbasedupon
warranty,businessagreement,tort,orotherwise;whetherornotinjurywassustainedbypersonsor
propertyorotherwise;andwhetherornotlosswassustainedfrom,oraroseoutof,theresultsof,the
item,oranyservicesthatmaybeprovidedbyECSS.
Publishedby: ESARequirementsandStandardsDivision
ESTEC,P.O.Box299,
2200AGNoordwijk
TheNetherlands
Copyright: 2011bytheEuropeanSpaceAgencyforthemembersofECSS
2
ECSSEHB3220Part2A
20March2011
Change log
ECSSEHB3220Part2A Firstissue
20March2011
3
ECSSEHB3220Part2A
20March2011
Table of contents
Change log .................................................................................................................3
Introduction..............................................................................................................27
10 Stress-strain relationships ................................................................................28

10.1 Introduction ........................................................................................................... 28
10.2 Elastic property prediction for UD ply from constituent properties........................ 28
10.3 Analytical notation for elastic constant methods................................................... 30
10.4 Calculation methods for elastic constants of UD ply............................................. 30
10.5 Longitudinal modulus............................................................................................ 32
10.6 Longitudinal Poisson's ratio .................................................................................. 34
10.7 Transverse modulus ............................................................................................. 35
10.7.1 General....................................................................................................... 35
10.7.2 Jones method ............................................................................................. 35
10.7.3 Frster/Knappe method.............................................................................. 35
10.7.4 Schneider method ...................................................................................... 36
10.7.5 Puck method............................................................................................... 36
10.7.6 Tsai method................................................................................................ 36
10.7.7 HSB method ............................................................................................... 37
10.7.8 Graphs........................................................................................................ 37
10.8 Transverse Poisson's ratio ................................................................................... 40
10.9 Transverse shear modulus ................................................................................... 40
10.9.1 General....................................................................................................... 40
10.9.2 Jones method ............................................................................................. 40
10.9.3 Frster/Knappe method.............................................................................. 40
10.9.4 Schneider method ...................................................................................... 40
10.9.5 Puck method............................................................................................... 41
10.9.6 Tsai method................................................................................................ 41
10.9.7 HSB method ............................................................................................... 41
10.9.8 Graphs........................................................................................................ 42
10.10 In-plane stress calculation methods ..................................................................... 44
4
ECSSEHB3220Part2A
20March2011
10.11 Analytical notation for in-plane methods............................................................... 44
10.12 Stress-strain relations for unidirectional plies ....................................................... 46
10.12.1 Fibre-oriented co-ordinate system.............................................................. 46
10.13 On axis stress strain relations............................................................................... 46
10.13.1 General....................................................................................................... 46
10.13.2 Compliance matrix...................................................................................... 47
10.13.3 Modulus matrix ........................................................................................... 47
10.13.4 Symmetry of compliance and modulus matrices ........................................ 47
10.14 Stress-strain relations for a ply of arbitrary orientation ......................................... 47
10.14.1 General....................................................................................................... 47
10.14.2 Off axis stiffness of a unidirectional ply ...................................................... 49
10.15 Stiffness matrix for a laminate .............................................................................. 50
10.15.1 General laminates ...................................................................................... 50
10.15.2 Symmetric laminates .................................................................................. 52
10.15.3 Flow chart ................................................................................................... 53
10.16 Calculation methods with interlaminar stresses and strains ................................. 54
10.16.1 Calculation with free-edge stresses............................................................ 54
10.17 Qualitative evaluation of interlaminar strength for design purposes ..................... 56
10.17.1 General....................................................................................................... 56
10.17.2 Variation of fibre direction within a [, 0, ] laminate .......................... 58
10.17.3 Variation of the thickness of the 0 layer within the [ 30, 0n, 30]
laminate ...................................................................................................... 58
10.17.4 Variation of the sequence of layers ............................................................ 59
10.18 References ........................................................................................................... 59
10.18.1 General....................................................................................................... 59
11 Strength prediction and failure criteria.............................................................61

11.1 Introduction ........................................................................................................... 61
11.1.1 Micro-mechanical strength models............................................................. 61
11.1.2 Lamina failure models ................................................................................ 61
11.1.3 Failure criteria studies ................................................................................ 62
11.1.4 Summary of World Wide Failure Exercise (WWFE) ................................... 63
11.2 Tensile strength of UD composites in fibre direction ............................................ 65
11.2.1 General....................................................................................................... 65
11.2.2 Weakest-link failure model ......................................................................... 65
11.2.3 Cumulative weakening failure model.......................................................... 65
11.2.4 Fibre break propagation model................................................................... 67
11.2.5 Cumulative group mode failure model........................................................ 67
5
ECSSEHB3220Part2A
20March2011
11.2.6 Status of models......................................................................................... 67
11.3 Compressive strength of UD composites in fibre direction ................................... 67
11.3.1 General....................................................................................................... 67
11.3.2 Extension mode buckling............................................................................ 68
11.3.3 Shear mode buckling.................................................................................. 68
11.3.4 Analysis of compression failure .................................................................. 69
11.4 Transverse tensile strength of UD composites ..................................................... 71
11.4.1 General....................................................................................................... 71
11.4.2 Prediction of transverse tensile strength .................................................... 72
11.4.3 Empirical analysis....................................................................................... 72
11.5 Static strength criteria for composites................................................................... 73
11.6 Analytical notation for static strength criteria for composites................................ 73
11.6.1 Co-ordinate system .................................................................................... 73
11.6.2 Formulae .................................................................................................... 74
11.7 Different types of failure criteria ............................................................................ 75
11.7.1 General....................................................................................................... 75
11.7.2 Evaluation studies ...................................................................................... 76
11.8 Overview - Failure criteria..................................................................................... 76
11.8.1 Introduction................................................................................................. 76
11.8.2 Independent conditions .............................................................................. 77
11.8.3 Interactive conditions Pure interpolative conditions................................. 78
11.8.4 Interactive conditions - Physical considerations ......................................... 80
11.9 Comparison between test data and various failure criteria................................... 88
11.9.1 Effects on failure mode............................................................................... 88
11.10 Description of failure modes ................................................................................. 92
11.10.1 Laminates ................................................................................................... 92
11.10.2 Failure......................................................................................................... 92
11.11 Fatigue strength of composites............................................................................. 97
11.11.1 Background ................................................................................................ 97
11.11.2 Analytical notation ...................................................................................... 98
11.11.3 Approximation of fatigue life ....................................................................... 98
11.12 References ......................................................................................................... 100
11.12.1 General..................................................................................................... 100
12 Calculation of thermal stress and displacement ...........................................104

12.1 Introduction ......................................................................................................... 104
12.1.1 General..................................................................................................... 104
12.1.2 Longitudinal CTE ...................................................................................... 104
6
ECSSEHB3220Part2A
20March2011
12.1.3 Transverse CTE ....................................................................................... 104
12.2 Analytical notation for thermal stress calculations .............................................. 105
12.3 Calculation of CTE from constituents ................................................................. 106
12.3.1 CTE in fibre direction ................................................................................ 106
12.3.2 CTE perpendicular to fibre direction ......................................................... 106
12.4 CTE for a laminate.............................................................................................. 107
12.5 Thermal stresses within laminate layers............................................................. 109
12.5.1 General..................................................................................................... 109
12.5.2 Residual curing stresses .......................................................................... 109
12.6 Stress strain temperature relation....................................................................... 110
12.6.1 General..................................................................................................... 110
12.6.2 Mechanical strains.................................................................................... 110
12.6.3 Incremental strain theory .......................................................................... 110
12.7 Microstress analysis ........................................................................................... 112
12.7.1 General..................................................................................................... 112
12.7.2 Microstresses on fibre axis ....................................................................... 112
12.7.3 Microstresses normal to fibre axis ............................................................ 112
12.8 References ......................................................................................................... 112
12.8.1 General..................................................................................................... 112
13 Moisture effects on composite properties .....................................................114

13.1 Introduction ......................................................................................................... 114
13.1.1 General..................................................................................................... 114
13.1.2 Moisture penetration................................................................................. 114
13.1.3 Moisture effects ........................................................................................ 114
13.2 Analytical notation for moisture effects ............................................................... 115
13.3 Typical effects of moisture .................................................................................. 116
13.3.1 General..................................................................................................... 116
13.3.2 Sample data: Effects of moisture.............................................................. 116
13.4 Approximate method for calculation of strength and modulus retention of
[0/90] laminates ............................................................................................... 121
13.4.1 General..................................................................................................... 121
13.4.2 Modulus retention ..................................................................................... 121
13.4.3 Strength retention ..................................................................................... 123
13.5 Moisture content ................................................................................................. 124
13.5.1 Fick's law .................................................................................................. 124
13.5.2 Determination of moisture content............................................................ 125
13.5.3 Maximum moisture content ...................................................................... 126
7
ECSSEHB3220Part2A
20March2011
13.5.4 Experimental determination of the diffusion coefficient ............................ 127
13.6 Calculation of swelling coefficient from constituents........................................... 129
13.6.1 General..................................................................................................... 129
13.6.2 Swelling coefficient 1 in fibre direction .................................................... 129
13.6.3 Swelling coefficient 2 transverse to fibres ............................................... 129
13.6.4 Swelling coefficient for a laminate ............................................................ 130
13.7 Coefficient of moisture expansion (CME) ........................................................... 132
13.7.1 Resin behaviour........................................................................................ 132
13.7.2 Composite behaviour................................................................................ 134
13.8 References ......................................................................................................... 134
13.8.1 General..................................................................................................... 134
14 Stress concentrations and fracture ................................................................136

14.1 Introduction ......................................................................................................... 136
14.1.1 General..................................................................................................... 136
14.1.2 Fracture mechanics models ..................................................................... 136
14.2 Analytical notation for stress concentrations ...................................................... 137
14.3 Summary of fracture models............................................................................... 137
14.4 Evaluation of fracture models ............................................................................. 138
14.5 WEK fracture model............................................................................................ 139
14.5.1 General..................................................................................................... 139
14.5.2 Circular holes............................................................................................ 139
14.5.3 Straight crack............................................................................................ 142
14.6 WN fracture model.............................................................................................. 145
14.6.1 General..................................................................................................... 145
14.6.2 Failure criteria........................................................................................... 145
14.6.3 Characteristics of WN fracture model....................................................... 147
14.6.4 Circular holes............................................................................................ 148
14.6.5 Straight cracks.......................................................................................... 152
14.6.6 Point stress criteria ................................................................................... 153
14.6.7 Average stress criterion............................................................................ 154
14.7 Finite plate models.............................................................................................. 158
14.8 Finite width correction (FWC) ............................................................................. 158
14.8.1 General..................................................................................................... 158
14.8.2 Circular holes............................................................................................ 160
14.8.3 Centre crack ............................................................................................. 161
14.9 Calculated stress concentration factor at circular holes ..................................... 161
14.9.1 NASA results ............................................................................................ 161
8
ECSSEHB3220Part2A
20March2011
14.9.2 Finite width correction (FWC) ................................................................... 163
14.9.3 MBB/ERNO study..................................................................................... 163
14.10 Stress distribution around circular holes............................................................. 168
14.10.1 General..................................................................................................... 168
14.10.2 Stress concentration due to tensile load................................................... 168
14.10.3 Stress concentration due to shear load .................................................... 169
14.11 Interlaminar fracture mechanics ......................................................................... 171
14.11.1 Nomenclature ........................................................................................... 171
14.11.2 Delamination and fracture mechanics overview ....................................... 171
14.11.3 Standard test methods (static and fatigue)............................................... 175
14.11.4 Calculation of strain energy release rate in structural analysis ................ 179
14.12 References ......................................................................................................... 185
14.12.1 General..................................................................................................... 185
15 Prediction of dynamic characteristics ............................................................189

15.1 Introduction ......................................................................................................... 189
15.2 Definition of damping terms ................................................................................ 189
15.2.1 General terms........................................................................................... 189
15.2.2 Complex modulus model .......................................................................... 191
15.3 Prediction methods for damping ......................................................................... 192
15.4 Determination of damping characteristics........................................................... 193
15.4.1 Unidirectional characteristics.................................................................... 193
15.4.2 Off axis characteristics ............................................................................. 193
15.4.3 Laminate characteristics........................................................................... 194
15.5 Approximate data on damping............................................................................ 194
15.6 References ......................................................................................................... 195
15.6.1 General..................................................................................................... 195
16 Computer analysis of composites ..................................................................196

16.1 Introduction ......................................................................................................... 196
16.2 Computer programs: Analysis of composites ..................................................... 196
16.2.1 General..................................................................................................... 196
16.2.2 Finite element programs........................................................................... 198
16.2.3 Laminate analysis programs..................................................................... 199
16.2.4 Special applications programs.................................................................. 201
16.3 ESDU data for composite analysis ..................................................................... 203
16.3.1 General..................................................................................................... 203
16.3.2 ESDU data items ...................................................................................... 203
16.3.3 ESDUpac.................................................................................................. 205
9
ECSSEHB3220Part2A
20March2011
16.4 Buckling of orthotropic plates.............................................................................. 205
16.4.1 Title........................................................................................................... 205
16.4.2 Usage and scope...................................................................................... 205
16.4.3 Analysis .................................................................................................... 205
16.4.4 ESDUpac A7303 ...................................................................................... 205
16.4.5 Notes ........................................................................................................ 206
16.5 Flexural stiffness of flat strips ............................................................................. 206
16.5.1 Title........................................................................................................... 206
16.5.2 Usage and scope...................................................................................... 206
16.5.3 Analysis .................................................................................................... 206
16.6 Metallic skin stiffeners reinforced by composite - local buckling......................... 207
16.6.1 Title........................................................................................................... 207
16.6.2 Usage and scope...................................................................................... 207
16.6.3 Analysis and data ..................................................................................... 207
16.7 Laminate stress analysis .................................................................................... 208
16.7.1 Title........................................................................................................... 208
16.7.2 Usage and scope...................................................................................... 208
16.7.3 Analysis .................................................................................................... 208
16.8 Plate stiffnesses (In-plane) ................................................................................. 209
16.8.1 Title........................................................................................................... 209
16.8.2 Usage and scope...................................................................................... 209
16.8.3 Analysis and methods .............................................................................. 209
16.9 Bonded joints - 1................................................................................................. 210
16.9.1 Title........................................................................................................... 210
16.9.2 Usage and scope...................................................................................... 210
16.10 Bonded joints - 2................................................................................................. 211
16.10.1 Title........................................................................................................... 211
16.10.2 Usage and scope...................................................................................... 211
16.10.4 ESDUpac A7916 ...................................................................................... 212
16.11 Bonded joints - 3................................................................................................. 212
16.11.1 Title........................................................................................................... 212
16.11.2 Usage and scope...................................................................................... 212
16.12 Buckling of rectangular specially orthotropic plates............................................ 213
16.12.1 Title........................................................................................................... 213
10
ECSSEHB3220Part2A
20March2011
16.12.2 Usage and scope...................................................................................... 213
16.13 Bonded joints - 4................................................................................................. 214
16.13.1 Title........................................................................................................... 214
16.13.2 Usage and scope...................................................................................... 214
16.13.4 ESDUpac A8039 ...................................................................................... 215
16.14 Bonded joints - 5................................................................................................. 215
16.14.1 Title........................................................................................................... 215
16.14.2 Usage and scope...................................................................................... 215
16.14.3 Information and guidance ......................................................................... 215
16.15 Buckling of orthotropic plates.............................................................................. 216
16.15.1 Title........................................................................................................... 216
16.15.2 Usage and scope...................................................................................... 216
16.15.4 ESDUpac A8147 ...................................................................................... 216
16.16 Lay-up arrangements for special orthotropy ....................................................... 217
16.16.1 Title........................................................................................................... 217
16.16.2 Usage and scope...................................................................................... 217
16.17 Failure modes of laminated composites ............................................................. 218
16.17.1 Title........................................................................................................... 218
16.17.2 Usage and scope...................................................................................... 218
16.17.3 Analysis and failure modes....................................................................... 218
16.18 Failure criteria for layers of a laminated composite ............................................ 219
16.18.1 Title........................................................................................................... 219
16.18.2 Usage and scope...................................................................................... 219
16.19 Plate stiffnesses and apparent elastic properties ............................................... 219
16.19.1 Title........................................................................................................... 219
16.19.2 Usage and scope...................................................................................... 219
16.19.3 Analysis for stiffnesses and elastic properties.......................................... 220
16.19.4 ESDUpac A8335 ...................................................................................... 220
16.20 Natural frequencies of laminated flat plates........................................................ 220
16.20.1 Title........................................................................................................... 220
16.20.2 Usage and scope...................................................................................... 221
16.20.3 Calculation of natural frequencies ............................................................ 221
11
ECSSEHB3220Part2A
20March2011
16.20.4 ESDUpac A8336 ...................................................................................... 221
16.21 Strain in skin panels under acoustic loading....................................................... 222
16.21.1 Title........................................................................................................... 222
16.21.2 Usage and scope...................................................................................... 222
16.21.3 Calculation of surface strains ................................................................... 222
16.21.4 ESDUpac A8408 ...................................................................................... 222
16.22 Failure analysis................................................................................................... 223
16.22.1 Title........................................................................................................... 223
16.22.2 Usage and scope...................................................................................... 223
16.22.3 Analysis .................................................................................................... 223
16.22.4 ESDUpac A8418 ...................................................................................... 224
16.23 Endurance under acoustic loading ..................................................................... 225
16.23.1 Title........................................................................................................... 225
16.23.2 Usage and scope...................................................................................... 225
16.24 Stress and strain around circular holes .............................................................. 226
16.24.1 Title........................................................................................................... 226
16.24.2 Usage and scope...................................................................................... 226
16.24.3 Analysis .................................................................................................... 226
16.24.4 ESDUpac A8501 ...................................................................................... 226
16.25 Damping in composite plates.............................................................................. 227
16.25.1 Title........................................................................................................... 227
16.25.2 Usage and scope...................................................................................... 227
16.25.3 Calculation of damping ............................................................................. 227
16.25.4 ESDUpac 8512......................................................................................... 228
16.26 Sandwich panel natural frequencies................................................................... 228
16.26.1 Title........................................................................................................... 228
16.26.2 Usage and scope...................................................................................... 228
16.26.4 ESDUpac A8537 ...................................................................................... 229
16.27 Selection of reinforcement around circular holes................................................ 229
16.28 Buckling of unbalanced composite plates........................................................... 230
16.28.1 Title........................................................................................................... 230
16.28.2 Usage and scope...................................................................................... 230
16.28.4 ESDUpac A8620 ...................................................................................... 231
16.29 Sandwich panel response to acoustic loading.................................................... 231
16.29.1 Title........................................................................................................... 231
12
ECSSEHB3220Part2A
20March2011
16.29.2 Usage and scope...................................................................................... 232
16.29.3 Calculation of natural frequencies and surface strains............................. 232
16.29.4 ESDUpac A8624 ...................................................................................... 232
16.30 Sandwich column and beam face plate wrinkling ............................................... 233
16.30.1 Title........................................................................................................... 233
16.30.2 Usage and scope...................................................................................... 233
16.30.3 Analysis .................................................................................................... 233
16.30.4 ESDUpac A8713 ...................................................................................... 234
16.31 Buckling of curved composite panels ................................................................. 234
16.31.1 Title........................................................................................................... 234
16.31.2 Usage and scope...................................................................................... 234
16.31.4 ESDUpac A8725 ...................................................................................... 235
16.32 Sandwich panel face plate wrinkling................................................................... 235
16.32.1 Title........................................................................................................... 235
16.32.2 Usage and scope...................................................................................... 236
16.32.3 Analysis .................................................................................................... 236
16.32.4 ESDUpac A8815 ...................................................................................... 236
16.33 Vibration of singly-curved laminated plates ........................................................ 237
16.33.1 Title........................................................................................................... 237
16.33.2 Usage and scope...................................................................................... 237
16.33.4 ESDUpac A8911 ...................................................................................... 237
16.34 Plate through-the-thickness shear stiffnesses .................................................... 238
16.34.1 Title........................................................................................................... 238
16.34.2 Usage and scope...................................................................................... 238
16.34.3 Analysis .................................................................................................... 238
16.34.4 ESDUpac A8913 ...................................................................................... 239
16.34.5 Notes ........................................................................................................ 239
16.35 Vibration of plates with in-plane loading ............................................................. 239
16.35.1 Title........................................................................................................... 239
16.35.2 Usage and scope...................................................................................... 239
16.35.4 ESDUpac A9016 ...................................................................................... 240
16.36 Delamination and free edge stresses ................................................................. 240
16.36.1 Title........................................................................................................... 240
16.36.2 Usage and scope...................................................................................... 241
13
ECSSEHB3220Part2A
20March2011
16.36.3 Analysis .................................................................................................... 241
16.36.4 ESDUpac A9021 ...................................................................................... 241
16.36.5 Notes ........................................................................................................ 242
16.37 Delamination at termination of plies.................................................................... 242
16.37.1 Title........................................................................................................... 242
16.37.2 Usage and scope...................................................................................... 242
16.37.3 Analysis .................................................................................................... 242
16.37.4 ESDUpac A9103 ...................................................................................... 243
16.38 Thickness selection to meet a loading combination ........................................... 243
16.38.1 Title........................................................................................................... 243
16.38.2 Usage and scope...................................................................................... 243
16.38.3 Analysis .................................................................................................... 244
16.38.4 ESDUpac A9233 ...................................................................................... 244
16.39 ESAComp ........................................................................................................... 245
17 Composite adequate design............................................................................246

17.1 Introduction ......................................................................................................... 246
17.2 Anisotropy of composites.................................................................................... 246
17.3 Stress-strain relationships .................................................................................. 248
17.3.1 Reinforcing fibres...................................................................................... 248
17.3.2 Stress concentrations ............................................................................... 248
17.4 Fibre strength and stiffness ................................................................................ 250
17.4.1 General..................................................................................................... 250
17.4.2 High stiffness applications ........................................................................ 251
17.5 Basic design rules............................................................................................... 252
17.5.1 General..................................................................................................... 252
17.5.2 Aspects of construction ............................................................................ 252
17.5.3 Aspects of laminate lay up........................................................................ 255
17.5.4 Fabrication aspects .................................................................................. 258
17.6 First steps in designing a composite................................................................... 261
17.6.1 General..................................................................................................... 261
17.6.2 Carpet plots .............................................................................................. 261
17.6.3 Use of carpet plots.................................................................................... 263
17.7 References ......................................................................................................... 264
17.7.1 General..................................................................................................... 264
18 Curing stresses: Effects and prediction.........................................................266

18.1 Introduction ......................................................................................................... 266
18.2 Cure process ...................................................................................................... 266
14
ECSSEHB3220Part2A
20March2011
18.2.1 Composite materials................................................................................. 266
18.2.2 Cure parameters....................................................................................... 267
18.3 Analytical notation for residual stress ................................................................. 268
18.4 Residual stresses ............................................................................................... 268
18.4.1 General..................................................................................................... 268
18.4.2 Types of residual stresses........................................................................ 268
18.5 Calculation of curing stresses............................................................................. 269
18.5.1 Residual stresses after curing .................................................................. 269
18.6 Reduction of thermal stresses and distortions.................................................... 271
18.6.1 General..................................................................................................... 271
18.6.2 Stress relieving ......................................................................................... 272
18.7 References ......................................................................................................... 272
18.7.1 General..................................................................................................... 272
19 Manufacturing faults and service damage .....................................................274

19.1 Introduction ......................................................................................................... 274
19.2 Manufacturing defects in composite materials.................................................... 276
19.2.1 General..................................................................................................... 276
19.2.2 Description of manufacturing defects ....................................................... 277
19.2.3 Detection of defects.................................................................................. 281
19.3 Service threats for composite structures ............................................................ 281
19.4 Impact behaviour of laminates and sandwich constructions............................... 282
19.4.1 General..................................................................................................... 282
19.4.2 Laminates ................................................................................................. 283
19.4.3 Sandwich panels ...................................................................................... 286
19.5 Impact behaviour ................................................................................................ 286
19.5.1 General..................................................................................................... 286
19.5.2 BVID ......................................................................................................... 287
19.5.3 Impact tests .............................................................................................. 289
19.5.4 Compression after impact (CAI) ............................................................... 290
19.6 Detection of defects ............................................................................................ 290
19.6.1 Damage detection techniques .................................................................. 290
19.6.2 Laboratory and production based NDT .................................................... 290
19.6.3 Other techniques ...................................................................................... 290
19.7 References ......................................................................................................... 291
19.7.1 General..................................................................................................... 291
19.7.2 ASTM standards....................................................................................... 291
20 Environmental aspects of design ...................................................................292
15
ECSSEHB3220Part2A
20March2011
20.1 Introduction ......................................................................................................... 292
20.2 Description of environments ............................................................................... 292
20.2.1 Earth environment .................................................................................... 292
20.2.2 Space environment................................................................................... 293
20.2.3 Composite structures................................................................................ 297
20.3 Low earth orbit (LEO) ......................................................................................... 298
20.4 Geostationary orbit (GEO) .................................................................................. 299
20.5 Deep space exploration ...................................................................................... 299
20.6 Galvanic corrosion .............................................................................................. 300
20.6.1 General..................................................................................................... 300
20.6.2 Physical basis of galvanic corrosion......................................................... 300
20.6.3 Prevention of galvanic corrosion in space structures ............................... 302
20.7 Effects of moisture on composites...................................................................... 302
20.7.1 General..................................................................................................... 302
20.7.2 Modification of CTE by outgassing ........................................................... 302
20.7.3 Low coefficient of moisture expansion (CME) resins................................ 303
20.7.4 Hot/wet performance ................................................................................ 303
20.8 LDEF in LEO....................................................................................................... 304
20.8.1 Mission ..................................................................................................... 304
20.8.2 Materials and experiments ....................................................................... 304
20.8.3 Variations in exposure conditions............................................................. 305
20.8.4 Composite materials aboard LDEF .......................................................... 307
20.8.5 LDEF experiments M0003-9/10................................................................ 309
20.8.6 LDEF experiment AO 171 ........................................................................ 311
20.8.7 LDEF experiment AO 180 ........................................................................ 313
20.8.8 Surface characterisation of eroded composites ....................................... 313
20.8.9 Overall conclusions on LDEF ................................................................... 314
20.8.10 Non-polymeric composites on LDEF ........................................................ 314
20.8.11 Polymer films on LDEF............................................................................. 314
20.8.12 Lubricants, adhesives and seals on LDEF ............................................... 314
20.9 Thermal cycling................................................................................................... 315
20.9.1 Conditions................................................................................................. 315
20.9.2 Damage mechanisms............................................................................... 316
20.10 Vacuum .............................................................................................................. 319
20.10.1 Effects of vacuum ..................................................................................... 319
20.11 Radiation ............................................................................................................ 320
20.11.1 Radiation spectra...................................................................................... 320
16
ECSSEHB3220Part2A
20March2011
20.11.2 T300/934 in GEO...................................................................................... 323
20.12 Damage by combined environmental factors ..................................................... 324
20.12.1 General..................................................................................................... 324
20.12.2 P75/930 in GEO ....................................................................................... 324
20.12.3 UHM CFRP in GEO and LEO................................................................... 328
20.12.4 PEI CFRP in LEO and GEO ..................................................................... 331
20.13 Atomic oxygen .................................................................................................... 333
20.13.1 Effects of atomic oxygen .......................................................................... 333
20.14 Siloxanes and silicon polymers........................................................................... 335
20.14.1 Protection methods against ATOX ........................................................... 335
20.14.2 Protection of polymer films ....................................................................... 336
20.14.3 Protection of composites .......................................................................... 336
20.15 Protective coatings ............................................................................................. 338
20.15.1 Surface coatings....................................................................................... 338
20.16 Debris ................................................................................................................. 339
20.16.1 Classification of debris.............................................................................. 339
20.16.2 Damage to LDEF...................................................................................... 340
20.16.3 Damage to composites............................................................................. 340
20.16.4 Damage to aluminium alloys .................................................................... 341
20.16.5 Damage to thermal control materials........................................................ 341
20.16.6 Significance of impact events ................................................................... 342
20.16.7 Protective shielding .................................................................................. 342
20.17 References ......................................................................................................... 343
20.17.1 General..................................................................................................... 343
20.17.2 ECSS documents ..................................................................................... 347
21 Bonded joints....................................................................................................349
21.1 Introduction ......................................................................................................... 349
21.2 Adhesives ........................................................................................................... 349
21.2.1 General..................................................................................................... 349
21.2.2 Types of adhesives................................................................................... 349
21.2.3 Adhesives for joining different materials................................................... 352
21.3 Design of bonded joints ...................................................................................... 354
21.3.1 Basic considerations................................................................................. 354
21.3.2 Basic guidelines........................................................................................ 354
21.3.3 Failure modes........................................................................................... 354
21.3.4 Features ................................................................................................... 355
21.4 Joint configuration............................................................................................... 357
17
ECSSEHB3220Part2A
20March2011
21.4.1 Basic configurations ................................................................................. 357
21.4.2 Orientation of surface fibres ..................................................................... 360
21.5 Environmental factors for bonded joints ............................................................. 361
21.5.1 General..................................................................................................... 361
21.5.2 Effect of moisture...................................................................................... 361
21.5.3 Effect of temperature ................................................................................ 364
21.5.4 Combined moisture and temperature ....................................................... 364
21.6 Bonding defects .................................................................................................. 365
21.6.1 General..................................................................................................... 365
21.6.2 Description of bonding defects ................................................................. 365
21.6.3 Inspection of bonded joints....................................................................... 372
21.7 Bonded joint failure modes ................................................................................. 374
21.7.1 Typical failure modes................................................................................ 374
21.7.2 Loading modes ......................................................................................... 374
21.8 Calculation of bonded joint strength ................................................................... 375
21.9 Analysis of joint configurations ........................................................................... 376
21.9.1 Analytical notation .................................................................................... 376
21.9.2 Single lap shear joint ................................................................................ 376
21.9.3 Double lap shear joint............................................................................... 380
21.9.4 Double-lap shear joint under mechanical and temperature loads ............ 382
21.9.5 Single taper scarf joint .............................................................................. 390
21.9.6 Double taper scarf joint............................................................................. 391
21.9.7 Stepped lap joint....................................................................................... 392
21.10 Bonded joint design curves and test data........................................................... 394
21.10.1 General..................................................................................................... 394
21.11 Acoustic fatigue of bonded configurations .......................................................... 405
21.12 References ......................................................................................................... 407
21.12.1 General..................................................................................................... 407
21.12.2 ECSS documents ..................................................................................... 408
22 Mechanically fastened joints ...........................................................................409

22.1 Introduction ......................................................................................................... 409
22.2 Basic considerations for design .......................................................................... 409
22.2.1 Advantages............................................................................................... 409
22.2.2 Disadvantages.......................................................................................... 409
22.3 Factors affecting design...................................................................................... 410
22.3.1 General..................................................................................................... 410
22.3.2 Material parameters.................................................................................. 411
18
ECSSEHB3220Part2A
20March2011
22.3.3 Fastener parameters ................................................................................ 413
22.3.4 Design parameters ................................................................................... 413
22.4 Bolted joints ........................................................................................................ 418
22.4.1 Material parameters.................................................................................. 418
22.4.2 Fastener parameters ................................................................................ 418
22.4.3 Design parameters ................................................................................... 420
22.5 Bolted joint: Analysis........................................................................................... 420
22.5.1 General..................................................................................................... 420
22.5.2 Analytical methods ................................................................................... 420
22.5.3 Stress distribution ..................................................................................... 420
22.5.4 Failure prediction ...................................................................................... 423
22.5.5 Experimental data..................................................................................... 425
22.6 Riveted joints ...................................................................................................... 429
22.6.1 General..................................................................................................... 429
22.6.2 Installation damage of riveted joints ......................................................... 429
22.6.3 Pull through strength ................................................................................ 430
22.6.4 Fasteners used in composite structures................................................... 431
22.7 References ......................................................................................................... 433
22.7.1 General..................................................................................................... 433
22.7.2 ECSS standards ....................................................................................... 434
Figures
Figure 10.2-1 - Derivation of macro-mechanical properties for the analysis of laminates...... 29
Figure 10.5-1 - Longitudinal modulus (E1) of glass/epoxy...................................................... 32
Figure 10.5-2 - Longitudinal modulus (E1) of carbon (HT)/epoxy ........................................... 33
Figure 10.5-3 - Longitudinal modulus (E1) of carbon (HM)/epoxy .......................................... 33
Figure 10.5-4 - Longitudinal modulus (E1) of aramid/epoxy ................................................... 34
Figure 10.7-1 - Transverse modulus (E2) of glass/epoxy composites .................................... 38
Figure 10.7-2 - Transverse modulus (E2) of carbon (HT)/epoxy composites ......................... 38
Figure 10.7-3 - Transverse modulus (E2) of carbon (HM)/epoxy composites ........................ 39
Figure 10.7-4 - Transverse modulus (E2) of aramid/epoxy composites ................................. 39
Figure 10.9-1 - Shear modulus (G12) of glass/epoxy composites........................................... 42
Figure 10.9-2 - Shear modulus (G12) of carbon (HT)/epoxy composites ................................ 43
Figure 10.9-3 - Shear modulus (G12) of carbon (HM)/epoxy composites ............................... 43
Figure 10.9-4 - Shear modulus (G12) of aramid/epoxy composites ........................................ 44
Figure 10.12-1 - Definition of the co-ordinate system for the analysis of unidirectional
plies ..................................................................................................................... 46
19
ECSSEHB3220Part2A
20March2011
Figure 10.14-1 - Definition of axes ......................................................................................... 48
Figure 10.14-2 - Transformation of stiffness and compliance matrix ..................................... 50
Figure 10.15-1 - In plane forces and moments ...................................................................... 51
Figure 10.15-2 - Geometry of an n-layered laminate ............................................................. 52
Figure 10.15-3 - Calculation of a laminate ............................................................................. 53
Figure 10.16-1 - Uniaxial loading of a composite ................................................................... 54
Figure 10.16-2 - Typical decrease of interlaminar stresses with high peaks near the
edges ................................................................................................................... 55
Figure 10.17-1 - External load on a composite plate.............................................................. 56
Figure 10.17-2 - Determination of the critical fibre direction................................................... 58
Figure 11.2-1 - Fibre reinforced composite: Tensile failure model ......................................... 66
Figure 11.3-1 - Buckling modes of fibres under compression ................................................ 68
Figure 11.3-2 - Comparison of predicted and experimental values of longitudinal
compressive strength........................................................................................... 70
Figure 11.6-1 - Analytical notation: Co-ordinate system ........................................................ 74
Figure 11.8-1 Cuntze failure criteria - Schematic diagram of failure modes in
transversely-isotropic UD material ....................................................................... 84
Figure 11.8-2 Cuntze failure criteria Notation and formulae for fabric failure
conditions............................................................................................................. 87
Figure 11.8-3 Comparison between predicted failure curves for C-C/SiC composites
with experimental data......................................................................................... 88
Figure 11.9-1 - Test results: Effect of thickness variation on static failure ............................. 89
Figure 11.9-2 - Test results: Effect of static and dynamic loading.......................................... 90
Figure 11.9-3 - Comparison: Failure theories applied to static strength result-...................... 91
Figure 12.3-1 - CTE of unidirectional composites ................................................................ 106
Figure 12.3-2 - CTE for T300/914 0/45/90 laminates..................................................... 107
Figure 13.3-1 - Stress/strain curves of Fibredux 914 (neat resin) for different moisture
contents tested at RT......................................................................................... 117
Figure 13.3-2 - Swelling strain, transverse (UD laminate).................................................... 118
Figure 13.3-3 - Influence of temperature and time on weight gain....................................... 118
Figure 13.3-4 - Test results: Moisture absorption versus time ............................................. 119
Figure 13.3-5 - Transverse tensile strength ......................................................................... 119
Figure 13.3-6 - Transverse modulus of elasticity ................................................................. 120
Figure 13.3-7 - Influence of temperature and moisture on fatigue strength of
unidirectional 0 laminates................................................................................. 120
Figure 13.4-1 - Moisture: Comparison of calculated and experimental results of
strength and stiffness of [0/90] GFRP laminates ............................................ 122
Figure 13.4-2 - Moisture: Comparison of calculated and experimental results of
strength and stiffness [0/90] CFRP laminates................................................. 123
Figure 13.5-1 - Notation: Definition of axes.......................................................................... 124
20
ECSSEHB3220Part2A
20March2011
Figure 13.5-2 - Illustration of change of moisture content with square root of time.............. 127
Figure 13.5-3 - Variation of maximum moisture content with relative humidity for a
T300/1034 composite ........................................................................................ 128
Figure 13.6-1 - Moisture: Coefficient of swelling for unidirectional composites.................... 130
Figure 13.6-2 - Moisture: Definition of swelling coefficient 2 for orthotropic material
(unidirectional laminate)..................................................................................... 131
Figure 13.6-3 - Moisture: Swelling coefficients for T300/914 [0/ 45/90]s laminates....... 132
Figure 14.4-1 - Strength ratio versus hole diameter: Crack length for WN fracture model
- 'Average stress criterion' ................................................................................. 139
Figure 14.5-1 - Notation: Infinite isotropic plate with circular hole........................................ 140
Figure 14.5-2 - Comparison between experiment and prediction by WEK model................ 142
Figure 14.5-3 - Notation: Infinite plate with straight crack .................................................... 143
Figure 14.5-4 - Comparison between experiment and prediction by WEK model................ 144
Figure 14.5-5 - Determination of ac for various laminates ................................................... 145
Figure 14.6-1 - Notation: Point stress criterion diagram....................................................... 146
Figure 14.6-2 - Notation: Average stress criterion diagram.................................................. 147
Figure 14.6-3 - Notation: Infinite orthotropic plate with circular hole .................................... 148
Figure 14.6-4 - Comparison between experiments and predictions of WN model ............... 150
Figure 14.6-5 - Effects of stacking sequence on notch sensitivity........................................ 151
Figure 14.6-6 - Effects of fibre angle on notch sensitivity..................................................... 151
Figure 14.6-7 - Infinite orthotropic plate with straight crack.................................................. 152
Figure 14.6-8 - Comparison between experiments and predictions of WN model ............... 155
Figure 14.6-9 - Critical stress intensity factor KQ................................................................. 155
Figure 14.6-10 - Determination of a0 for various laminates, applying 'Average stress
criterion' ............................................................................................................. 156
Figure 14.6-11 - Characteristic dimensions as a function of half crack length ..................... 157
Figure 14.6-12 - Effect of stacking sequence....................................................................... 157
Figure 14.8-1 - Anisotropy factor versus net section reduction for centre cracked angle
ply CFRP specimens ......................................................................................... 159
Figure 14.8-2 - Anisotropy factor versus net section reduction for centre cracked CFRP
specimens.......................................................................................................... 160
Figure 14.9-1 - Stress concentration factor K for orthotropic CFRP laminates with a
Tn
circular hole, 2 L W 10 .......................................................................... 162

Figure 14.9-2 - Stress distribution at a hole for 0/0 laminates ........................................... 164
Figure 14.9-3 - Stress distribution at a hole for 32.5/0/32.5 laminates......................... 165
Figure 14.9-4 - Stress distribution at a hole for laminates.................................................... 166
Figure 14.9-5 - Stress distribution at a hole for laminates.................................................... 167
Figure 14.10-1 - Stress concentration due to tensile load.................................................... 168
Figure 14.10-2 - Stress concentration due to shear load ..................................................... 170
21
ECSSEHB3220Part2A
20March2011
Figure 14.11-1 - Fracture modes: opening, shearing, and tearing ....................................... 172
Figure 14.11-2 - Outline of no-damage-growth methodology............................................... 173
Figure 14.11-3 - Fracture mechanics life prediction methodology ....................................... 175
Figure 14.11-4 - DCB and ENF specimens.......................................................................... 177
Figure 14.11-5 - Representative G-Nonset data ..................................................................... 179
Figure 14.11-6 - Node definitions for 2D VCCT ................................................................... 181
Figure 14.11-7 - Node definitions for 3D VCCT ................................................................... 183
Figure 14.11-8 - G-a curve for a stiffener debond ................................................................ 184
Figure 14.11-9 - Schematics of possible G-a curves ........................................................... 184
Figure 15.5-1 - Summary of loss factors for various fibres / epoxy composites................... 194
Figure 17.2-1 - Anisotropic behaviour of carbon fibre reinforced plastic .............................. 247
Figure 17.3-1 - Stress-strain response of CFRP (in tension and compression) ................... 249
Figure 17.3-2 - Fatigue strength of notched and un-notched aluminium and carbon fibre
reinforced plastic................................................................................................ 250
Figure 17.4-1 - Specific tensile strength and modulus properties of reinforcing fibres......... 251
Figure 17.5-1 - Shaping changes in thickness ..................................................................... 252
Figure 17.5-2 - Design of shapes ......................................................................................... 253
Figure 17.5-3 - Need for radii on corners ............................................................................. 253
Figure 17.5-4 - Cut outs: Bad and good practice ................................................................. 254
Figure 17.5-5 - Directionality of tensile strength of various semi-product forms:
Preferred loading direction................................................................................. 256
Figure 17.5-6 - Symmetrical lay-up: Avoids warpage and minimises loading the resin
matrix ................................................................................................................. 257
Figure 17.5-7 - Split design technique.................................................................................. 257
Figure 17.5-8 - Dimensional change for various materials due to a temperature change
of 100 ............................................................................................................... 258
Figure 17.5-9 - Fabrication costs for a CFRP structure........................................................ 259
Figure 17.5-10 - Comparison of costs: Metallic and CFRP .................................................. 260
Figure 17.6-1 - Carpet plot curves of modulus for 0/45/90 laminates ............................ 262
Figure 17.6-2 - Carpet plot curves of strength for 0/45/90 laminates............................. 262
Figure 18.2-1 - Typical cure cycle for CFRP ........................................................................ 267
Figure 18.5-1 - Residual stresses after curing ..................................................................... 270
Figure 18.6-1 - Normalised maximum curing stresses in []s laminates........................... 271
Figure 18.6-2 - Through thickness tensor polynomial distributions for curing stresses
and stresses at first failure in []s laminates................................................... 272
Figure 19.1-1 - Examples of defects in composite materials................................................ 275
Figure 19.4-1 - The primary failure modes of composite laminates under impact loading ... 284
Figure 19.5-1 - Multiple delaminations in a [(02, 45)2]s CFRP laminate caused by drop
weight impact..................................................................................................... 287
22
ECSSEHB3220Part2A
20March2011
Figure 19.5-2 - Impact damage in a 64-ply CFRP laminate caused by high velocity
impact ................................................................................................................ 288
Figure 20.2-1 - Space environment effects: Vacuum ........................................................... 294
Figure 20.2-2 - Space environment effects: Radiation ......................................................... 294
Figure 20.2-3 - Space environment effects: Temperature.................................................... 295
Figure 20.2-4 - Space environment effects: Micrometeoroids and debris............................ 295
Figure 20.2-5 - Space environment effects: Atomic oxygen (ATOX) ................................... 296
Figure 20.2-6 - Space environment effects: Re-entry .......................................................... 296
Figure 20.6-1 - Galvanic corrosion: Potential in 3.5 % NaCl solution................................... 301
Figure 20.8-1 LDEF: Experiment locations (M0003-10).................................................... 307
Figure 20.8-2 - LDEF: Estimated AO erosion depth versus fibre content for various
epoxy CFRP materials....................................................................................... 309
Figure 20.9-1 - Ratio of residual tensile strength to the initial strength versus number of
thermal cycles (-160 to +95C): [45]2S specimen tensile tested at RT and
100C................................................................................................................. 317
Figure 20.9-2 - Degradation of stiffness of various materials due to thermal cycling (-
160C to + 95C): [45]2S specimen tensile tested at RT and 100C ............... 318
Figure 20.11-1 - UV radiation: Effects on tensile moduli of carbon/epoxy composites ........ 321
Figure 20.11-2 - Effects of total radiation dose on microcrack formation in composite
specimens subjected to 100 thermal cycles (-150C to +80C)......................... 322
Figure 20.11-3 - Effects of total radiation dose on the coefficient of thermal expansion
of a toughened composite system (-150C to -73C) ........................................ 323
Figure 20.12-1 - Effect of cure temperature and electron radiation exposure on the Tg
of P75/930 in GEO............................................................................................. 326
Figure 20.12-2 - Effects of cure temperature and electron radiation exposure on the
shear strength of P75/930 in GEO..................................................................... 327
Figure 20.12-3 - Step simulation of 7 years in GEO............................................................. 329
Figure 20.12-4 - Effect of ATOX erosion on different CFRP composites ............................. 331
Figure 20.12-5 - Irradiation and thermal cycling: Changes in mechanical properties for
C6000/PEI ......................................................................................................... 333
Figure 21.2-1 - Adhesives for bonding different materials.................................................... 353
Figure 21.4-1 - Bonded joints: Basic configurations............................................................. 359
Figure 21.4-2 - Bonded joints: Orientation of surface fibre................................................... 360
Figure 21.5-1 - Effect of moisture on stress distributions in adhesive bonded joints ........... 362
Figure 21.5-2 - Influence of moisture absorption/desorption on peak adhesive shear
strains ................................................................................................................ 363
Figure 21.5-3 - Effect of progressive moisture absorption on bond strain............................ 364
Figure 21.6-1 - Bonded joints: Examples of acceptable bond flaw sizes ............................. 366
Figure 21.6-2 - Typical quality zoning for bonded joints....................................................... 367
Figure 21.6-3 - Adhesive shear stresses in bonded joints ................................................... 368
Figure 21.6-4 - Adhesive stresses in flawed bonded joints .................................................. 368
23
ECSSEHB3220Part2A
20March2011
Figure 21.6-7 - Adhesive bonded joints: Effect of debond flaws on flexibility....................... 370
Figure 21.6-8 - Variation of peak induced adhesive stress with thickness of adhesive at
ends of overlap .................................................................................................. 371
Figure 21.7-1 - Components of bonded joint strength.......................................................... 374
Figure 21.7-2 - Bonded joints: Loading modes or types of stresses .................................... 375
Figure 21.9-1 - Analysis: Notation for symmetrical single-lap shear joint............................. 377
Figure 21.9-2 - Analysis: Notation for single-lap joint R-degree peeling .............................. 378
Figure 21.9-3 - -degree peeling strength ............................................................................ 379
Figure 21.9-4 - Analysis: Notation for double-lap shear joint ............................................... 380
Figure 21.9-5 - Analysis: Notation for double lap joint (standard overlap length)................. 382
Figure 21.9-6 - Shear stress distribution versus the adhesive length (E1t1 = E2t2) for
single lap joint without eccentricity..................................................................... 382
Figure 21.9-7 - Stress concentration factor f2 as a function of K2 with parameter 2 ........... 384
Figure 21.9-8 - Shear stress distribution for large overlap lengths....................................... 386
Figure 21.9-9 - Stress concentration factor for large overlap lengths .................................. 387
Figure 21.9-10 - Shear stress distribution versus the normalised bonding length x/L for
a bonded joint between CFRP-HT unidirectional (60 vol. %) and aluminium,
titanium, GFRP quasi-isotropic, and GFRP unidirectional materials ................. 388
Figure 21.9-11 - Shear stress distribution versus the normalised bonding length x L
for a bonded joint between CFRP HT quasi isotropic (60 vol. %) and
aluminium, titanium, GFRP quasi isotropic, and GFRP unidirectional
materials ............................................................................................................ 389
Figure 21.9-12 - Analysis: Notation for single taper scarf joint............................................. 390
Figure 21.9-13 - Analysis: Notation for symmetrical double tapered scarf joint ................... 391
Figure 21.9-14 - Analysis: Notation for stepped lap joint (recessed and simple) ................. 392
Figure 21.9-15 - Analysis: Notation for recessed and simple scarf joints (four or more
steps) ................................................................................................................. 393
Figure 21.10-1 - Design curve: Single lap shear joint .......................................................... 395
Figure 21.10-4 - Design curve: Lower bond tension single lap shear joint strength
versus lap length................................................................................................ 397
Figure 21.10-5 - Design curve: Lower bond tension single lap shear joint strength
versus lap length................................................................................................ 398
Figure 21.10-6 - Design curve: Lower bond tension single lap compression loaded joint
strength versus lap length.................................................................................. 398
Figure 21.10-7 - Design curve: Single lap tension fatigue S-N curve................................... 399
Figure 21.10-8 - Design curve: Single-lap shear joint tension fatigue S-N curve................. 399
24
ECSSEHB3220Part2A
20March2011
Figure 21.10-9 - Design curve: Double lap shear joint ......................................................... 400
Figure 21.10-10 - Design curve: Symmetrical double scarf joint strength............................ 401
Figure 21.10-11 - Design curve: Symmetrical double scarf joint strength............................ 402
Figure 21.10-12 - Design curve: Symmetrical double scarf joint, tension fatigue S-N
curve .................................................................................................................. 403
Figure 21.10-13 - Design curve: Symmetrical double scarf joint tension fatigue S-N
curve .................................................................................................................. 404
Figure 21.10-14 - Design curve: Symmetrical double scarf joint tension fatigue S-N
curve .................................................................................................................. 405
Figure 21.11-1 - Bonded joint configurations in Hercules AS 4/3501-6 ............................... 406
Figure 22.3-1 - Bolted joints: Bearing stresses for various CFRP ply configurations ........... 412
Figure 22.3-2 - Mechanically fastened joints: Examples of configurations........................... 413
Figure 22.3-3 - Mechanically fastened joints: Definition of joint geometry ........................... 415
Figure 22.3-4 - Bolted joints: Failure modes ........................................................................ 417
Figure 22.3-5 - Local reinforcements for bolted joints .......................................................... 418
Figure 22.4-1 - Bearing strength of bolted carbon fibre reinforced laminates ...................... 419
Figure 22.5-1 Stress 2 along the x1-axis in an isotropic infinite plate containing an
unloaded circular hole........................................................................................ 421
Figure 22.5-2 - Stress 2 along the x1-axis in an orthotropic infinite plate [0/90]
containing an unloaded circular hole ................................................................. 422
Figure 22.5-3 - Stress 2 along the x1-axis in an isotropic plate of finite width containing
a loaded hole ..................................................................................................... 423
Figure 22.5-4 - Whitney/Nuismer failure hypothesis: Characteristic curve........................... 424
Figure 22.5-5 - Bolted joints: Test device and configuration for inplane face sheet
loading ............................................................................................................... 426
Figure 22.5-6 - Bolted joints: Bearing stress at failure point of CFRP facings with
different lay-ups ................................................................................................. 428
Figure 22.6-1 - Rivets: Improvement in pull-through strength by larger flush head
diameter and footprint........................................................................................ 430
Figure 22.6-2 - Types of rivets for aerospace materials....................................................... 432
Tables
Table 10.4-1 - Elastic property data for comparison .............................................................. 31
Table 10.4-2 - Basis of calculation methods for different authors .......................................... 31
Table 10.17-1 - Intralaminar stresses and strains: value for various stacking
sequences ........................................................................................................... 59
Table 10.17-2 - Intralaminar stresses and strains: value at interface for various
stacking sequences ............................................................................................. 59
Table 11.10-1 - Failure modes of unidirectional laminates..................................................... 93
Table 11.10-2 - Failure modes of multidirectional laminates.................................................. 95
Table 11.10-3 - Examples of some failure modes.................................................................. 96
25
ECSSEHB3220Part2A
20March2011
Table 13.5-1 - Summary of constants for the determination of the maximum moisture
content, M ........................................................................................................ 126
Table 13.7-1 - Moisture absorption: Cyanate ester Fiberite 954 composites ....................... 133
Table 14.1-1 - Intrinsic and extrinsic variables relevant for stress intensity factors ............. 136
Table 14.3-1 - Some fracture models and their authors....................................................... 137
Table 14.9-1 - HT carbon/epoxy........................................................................................... 161

Table 14.9-2 - KT for various laminates, w/d .............................................................. 162
Table 14.9-3 - Material data ................................................................................................. 163

Table 14.11-1 - Standard test methods................................................................................ 177
Table 16.2-1 - Commonly available software for the analysis of composites....................... 197
Table 16.2-2 - Computer programs for finite element analysis ............................................ 198
Table 16.2-3 - Computer programs for laminate analysis .................................................... 199
Table 16.2-4 - Computer programs for special composite applications ............................... 201
Table 16.3-1 - ESDU data sheets for composite analysis.................................................... 204
Table 17.5-1 - Minimum radii for fibres and laminates ......................................................... 254
Table 17.5-2 - Basic and semi-product forms ...................................................................... 255
Table 17.5-3 - Composite manufacturing methods .............................................................. 261
Table 19.2-1 - Types of defects in composite materials....................................................... 277
Table 20.8-1 LDEF: Polymer composites.......................................................................... 306
Table 20.8-2 - LDEF: Exposure levels ................................................................................. 307
Table 20.8-3 - LDEF: Composite material classifications..................................................... 308
Table 20.8-4 - LDEF experiment AO 171: Tensile properties of composites ....................... 312
Table 20.9-1 - Coefficient of thermal expansion for 45 ..................................................... 316
Table 20.12-1 - Microcrack densities in [02/902]S CFRP laminates ...................................... 330
Table 20.12-2 - GEO and LEO effects: Mechanical properties of C6000/PEI composites .. 332
Table 20.13-1 - Reaction efficiencies for some selected composite polymers and
organic films: Determined by ~40 hours of LEO exposure in space.................. 334
Table 20.13-2 - Effects of short-term LEO exposure on coefficient of thermal expansion
for T300/934 carbon/epoxy laminates (RT to +82C) ........................................ 335
Table 21.2-1 - Adhesives: Common types ........................................................................... 350
Table 21.2-2 - Adhesives: Typical characteristics ................................................................ 351
Table 21.6-1 - Bonded joints: Description of NDT methods ................................................. 372
Table 21.6-2 - Bonded joints: Detection methods for various defects.................................. 373
Table 21.10-1 - Bonded joints: Summary of data curves ..................................................... 394
Table 22.5-1 - Bolted joints: Results of static test on material 914C T300........................... 427
Table 22.5-2 - Bolted joints: Results of static test on material 914C HM .......................... 428
Table 22.6-1 - Types of fasteners used to join aerospace materials.................................... 431
26
ECSSEHB3220Part2A
20March2011
Introduction
TheStructuralmaterialshandbook,ECSSEHB3220,ispublishedin8Parts.
Aglossaryofterms,definitionsandabbreviatedtermsforthesehandbooksiscontainedinPart8.
Thepartsareasfollows:
Part1 Overviewandmaterialpropertiesandapplications Clauses19
Part2 Designcalculationmethodsandgeneraldesignaspects Clauses1022
Part3 Loadtransferanddesignofjointsanddesignofstructures Clauses2332
Part4 Integritycontrol,verificationguidelinesandmanufacturing Clauses3345
Part5 Newadvancedmaterials,advancedmetallicmaterials,general Clauses4663
designaspectsandloadtransferanddesignofjoints
Part6 Fractureandmaterialmodelling,casestudiesanddesignand Clauses6481
integritycontrolandinspection
Part7 Thermalandenvironmentalintegrity,manufacturingaspects, Clauses82107
inorbitandhealthmonitoring,softmaterials,hybrid
materialsandnanotechnoligies
Part8 Glossary
27
ECSSEHB3220Part2A
20March2011
10
Stress-strain relationships
10.1 Introduction
The many analytical and practically derived methods developed for predicting the stressstrain
behaviourofcompositematerialsarereviewed.
Calculationmethodsaregivenfor:
Behaviourpredictionfromconstituentproperties,[See:10.2].
Thecalculationofintraply,[See:10.10].
Interlaminarstressstrains,[See:16.16].
10.2 Elastic property prediction for UD ply from

constituent properties
UD ply properties can be predicted by the procedures of micromechanics and be measured by
physicalmeans.Thesecanthenbeusedinamacromechanicalanalysisofthestructure.
Micromechanical analysis has inherent limitations. Fibre properties cannot easily be measured, so
they are determined from measurements of UD by applying inverse homoginisation formulae. The
transverseandshearpropertiesareparticularlyinaccurate.
Inaddition,aperfectbondbetweenfibresandmatrixisacommonassumptionforanalysis,whichis
not always true for some composites. Imperfect bonds give a material with poorer properties than
those of the micromechanical analysis. So, the analysis should be proven by careful experimental
work.ThederivationofmacromechanicalpropertiesissummarisedinFigure10.21.
28
ECSSEHB3220Part2A
20March2011
Prediction of Ply Properties

Micro-mechanical behaviour of ply
Macro-mechanical
Analysis of the Laminate
Measurement of Ply Properties
Physical means
Figure10.21Derivationofmacromechanicalpropertiesfortheanalysisof
laminates
The objective of the micromechanical approach is to determine the elastic moduli of a composite
material in terms of the moduli of the constituent materials. This is also often referred to as the
stiffnessorcompliance.
The additional and complementary objective of the micromechanical approach is to determine the
strengthsofthecompositematerialintermsofthestrengthsoftheconstituentmaterials.
Thebasicassumptionsfortheanalyticaltreatmentofacompositematerialare:
UDply:
Macroscopicallyhomogeneous
Linearlyelastic
Macroscopicallyorthotropic
Initiallystressfree
Perfectbondbetweenfibresandmatrix
Fibre:
Homogeneous
Linearlyelastic
Orthotropic
Regularlyspaced
Perfectlyaligned
Matrix:
Homogeneous
Linearlyelastic
Isotropic
Materialcharacteristicsofmatricesandfibres,alongwiththefibrevolumefraction,cansignificantly
affectthemacromechanicalpropertiesandthereforeneedcarefulevaluation.
29
ECSSEHB3220Part2A
20March2011
10.3 Analytical notation for elastic constant methods
Formulae:
E Youngsmodulus
Poissonsratio
G Shearmodulus
V Volumeratio
E1 Youngsmodulusofaplyinthefibredirection
E2 Youngsmodulusofaplyinthedirectiontransversetothe
fibre
x Poissonsratiooftheplywhenloadedinxdirection
G12 Inplaneshearmodulusoftheply
Stresspartitioningparameter

Subscripts:
f Fibre
m Matrix
1 Fibredirection
2 Transversetofibre
12 Shear(inplane)
10.4 Calculation methods for elastic constants of UD ply

Various authors have expressed analytically the elastic constants of an orthotropic layer dependent
upontheYoungsmoduli,Poissonsratioandshearmoduliofthefibreandmatrixaccordingtotheir
volumeratio.
The equations presented have been developed in each case for one specific material system, i.e.
glass/epoxy or carbon/epoxy. To verify each equation with respect to commonly used composites,
theyaredepictedtogethergraphically,withtestresults.Wherenotestresultsaregiven,thediagrams
canbeusedforpreliminarydesignpurposesonly.
DatausedforcomparisonaregiveninTable 10.4-1.Thebasisoftheapproachusedbyeachauthor,as
referenced,isshowninTable10.42.
30
ECSSEHB3220Part2A
20March2011
Table 10.4-1 - Elastic property data for comparison
E1 E2 G12 12 Ref. [X]
(kN/mm) (kN/mm) (kN/mm) (%) (Key)
Epoxy (1) 3.5 3.5 1.3 0.35 [10-3]
Cyanate - - - - -
Ester (4)
Glass, E 73 73 30 0.22 [10-2]
Carbon, HT 230 24 50 9.4(2) 0.28 [10-3]
Carbon, 550 6.0 18.0 2.0(2) 0.36 (3)
HM
Aramid 130 7.0 12.0 2.5(2) 0.38 (3)
Key: (1) Epoxy: CY209/HT 972, Fibredux 914 (3) Estimated
(2) G12 defined, Ref. [10-5] (4) ESA-study
Table10.42Basisofcalculationmethodsfordifferentauthors
Author Ref.[X] Basis
Basedonrepresentativevolumeelement
Jones [101]
withisotropicfibres.
Frster/Knapp Semiempiricalapproachforglass/epoxy.
[102]
e
Semiempiricalapproachforcarbon
Schneider [103]
(HT)/epoxy.
Puck [104] Semiempiricalapproachforglass/epoxy.
Semiempiricalapproachincluding
Tsai [105]
transverseandshearmodulus.
Modelofcircularfibreswithsquare
HSB [116]
packing.
Theapproachusedbyeachreferencedauthorispresentedfor:
Longitudinalmodulus,[See:10.5].
LongitudinalPoissonsratio,[See:10.6].
Transversemodulus,[See:10.7].
TransversePoissonsratio,[See:10.8].
Transverseshearmodulus,[See:10.9].
31
ECSSEHB3220Part2A
20March2011
10.5 Longitudinal modulus

Eachofthereferencedauthors,[See:Table10.4.2],usesthesameapproachtoderivearelationshipfor
longitudinalmodulus.Thisisthemethodofmixtures,i.e.:
E E V
1 f f
E m (1 V f ) [10.5-1]
[See:10.3fornotation]
The predicted relationship between fibre volume fraction and longitudinal modulus along with test
results,whereavailable,areshownfor:
glass/epoxyinFigure10.51.
carbonhighstrengthHT/epoxyinFigure10.52.
carbonhighmodulusHM/epoxyinFigure10.53.
aramid/epoxyinFigure10.54.
Figure 10.5-1 - Longitudinalmodulus(E1)ofglass/epoxy
32
ECSSEHB3220Part2A
20March2011
Figure 10.5-2 - Longitudinalmodulus(E1)ofcarbon(HT)/epoxy
Figure10.53Longitudinalmodulus(E1)ofcarbon(HM)/epoxy
33
ECSSEHB3220Part2A
20March2011
Figure10.54Longitudinalmodulus(E1)ofaramid/epoxy
10.6 Longitudinal Poisson's ratio

Each of the referenced authors, [See: Table 10.42], uses the same method of mixtures to derive a
relationshipforlongitudinalPoissonsratio:
V
12 f f m
(1 V f ) [10.6-1]
34
ECSSEHB3220Part2A
20March2011
10.7 Transverse modulus
10.7.1 General
Themethodsusedbyeachreferencedauthor,[See:Table10.42]aregivenalongwiththeirpredictions
fortransversemodulus:
Jonesmethod.
Frster/Knappemethod.
Schneidermethod.
Puckmethod.
Tsaimethod.
HSBmethod.
ThehomogenisationformulasofJones,Frster/Knappe,SchneiderandPuckaresuitableforisotropic
fibresonlyandcannotbeappliedtocarbonoraramidfibres,whicharenotisotropic.
10.7.2 Jones method
E m
E 2
(1 V f ) E mV
[10.71]
f E f
10.7.3 Frster/Knappe method
E m
E 2
(1 V f ) 1.45
V E E
0
[10.72]
f m f
0
E m
Where: E m
(1 m )
2

35
ECSSEHB3220Part2A
20March2011
10.7.4 Schneider method

0 3
E (1 V )
E2
m f
0
[10.73]
(1 V f ) 0.75 6V f E E m f
0
E m
Where: E m
(1 m )
2

10.7.5 Puck method

0 2
E (1 0.85V )
E (1
m f
2 0
[10.74]
V f
)1.25 V f E E m f
Where: E
0
E m

(1 m )
m 2
10.7.6 Tsai method

4k yGy
E [10.75]
2
k y mG y
4 x
2
Where: m 1
k y

E x
1 1 V f
V m
k y V f
V m k fy k
k m
k
V m
1 V f
1 1 V f
V m
G V y f
V m G fy G
G m
G
1
2 4 m G m
k 2(1 m) G fy

1
3 4 m G m
G 4(1 m) G fy

36
ECSSEHB3220Part2A
20March2011

G fy
k fy
1 2

f
G m
k m
1 2 m

10.7.7 HSB method

( V )
E 2 k y E m 1 2 ...
f
2(1 E m E fy
)

... 2
...
(1 ) 1 4V (1 E m E fy )
2
Em E fy f
1 2 (1 E m )
...arctan V f E fy [10.76]
1 2 (1 E E )
V f m fy

k y
=Correctionfactor;Ideallaminate k y
1
10.7.8 Graphs
The predicted relationship between fibre volume fraction and transverse modulus along with test
results,whereavailable,areshownfor:
Isotropicfibre:Glass/epoxyinFigure10.71.
Anisotropicfibres:
CarbonhighstrengthHT/epoxyinFigure 10.7-2.
CarbonhighmodulusHM/epoxyinFigure10.73.
Aramid/epoxyinFigure 10.7-4.
The data presented for anisotropic fibres, derived from the homogenisation formulas of Jones,
Frster/Knappe,SchneiderandPuckcanbeignoredinFigure10.71toFigure10.74,inclusive.
37
ECSSEHB3220Part2A
20March2011
Figure10.71Transversemodulus(E2)ofglass/epoxycomposites
Carbonfibresareanisotropic.DataderivedfromthehomogenisationformulasofJones,
Frster/Knappe,SchneiderandPuckcanbeignored.

Figure 10.7-2 - Transverse modulus (E2) of carbon (HT)/epoxy composites
38
ECSSEHB3220Part2A
20March2011
Carbon fibres are anisotropic. Data derived from the homogenisation formulas of Jones,
Frster/Knappe,SchneiderandPuckcanbeignored.
Figure10.73Transversemodulus(E2)ofcarbon(HM)/epoxycomposites
Aramid fibres are anisotropic. Data derived from the homogenisation formulas of Jones,
Frster/Knappe, Schneider and Puck can be ignored.
Figure 10.7-4 - Transverse modulus (E2) of aramid/epoxy composites
39
ECSSEHB3220Part2A
20March2011
10.8 Transverse Poisson's ratio

Eachofthereferencedauthors,[See:Table10.42]usesthesamemethodtoderivearelationshipfor
transversePoissonsratio:
E
E21 12
2
[10.8-1]
1
ThetransversePoissonsratioisthePoissonsratiooftheplywhenloadedinydirection.
10.9 Transverse shear modulus
10.9.1 General
Themethodsusedbyeachreferencedauthor,[See:Table10.4.2]aregivenalongwiththeirpredictions
fortransverseshearmodulus.
10.9.2 Jones method
G G m
[10.91]
12
(1 V ) V G G
f f m f
10.9.3 Frster/Knappe method

1 0 .4 V
G12 G m
f
[10.92]
(1 V f ) 1.45
V G G f m f
10.9.4 Schneider method

G (1 0.25 V )
G12
m f
[10.93]
(1 V ) f
1.25
1.25V G G f m f
40
ECSSEHB3220Part2A
20March2011
10.9.5 Puck method

1 0 .6 V
G12 G m
f
[10.94]
(1 V f ) 1.25
V G Gf m f
10.9.6 Tsai method

1 0 .6 V
G12 G m
f
[10.95]
(1 V f ) 1.25
V G Gf m f
V
Where: 1 1 m
12 2
G f
10.9.7 HSB method

( V )
G12 k 12 G m 1 2
...
f
2(1 G m G f )

... 2
...
(1 G G ) 1 4V (1 G m G f
) 2
m f f
1 2 (1 G m )
... arctan V f G f [10.9-6]
1 2 (1 G G )
V f m f

k 12
= Correction factor. For an ideal laminate k 12
1
41
ECSSEHB3220Part2A
20March2011
10.9.8 Graphs
The predicted relationship between fibre volume fraction and transverse shear modulus along with
testresults,whereavailable,areshownfor:
Glass/epoxyinFigure10.9.1.
CarbonhighstrengthHT/epoxyinFigure10.9.2.
CarbonhighmodulusHM/epoxyinFigure10.9.3.
Aramid/epoxyinFigure10.9.4.
Figure 10.9-1 - Shear modulus (G12) of glass/epoxy composites
42
ECSSEHB3220Part2A
20March2011
Figure10.92Shearmodulus(G12)ofcarbon(HT)/epoxycomposites
Figure 10.9-3 - Shear modulus (G12) of carbon (HM)/epoxy composites
43
ECSSEHB3220Part2A
20March2011
Figure 10.9-4 - Shear modulus (G12) of aramid/epoxy composites
10.10 In-plane stress calculation methods

Thestiffnessofaunidirectionalfibrereinforcedlayer,likeanyotherstructuralmaterial,isdefinedby
appropriatestressstrainrelationsandiscoupledtothestressstrainrelationofathinlaminateplate
accordingtothearbitraryorientationandthicknessofeachply.
[See:10.12forunidirectionalplies;10.13foronaxisrelationship;10.14forplyorientation]
10.11 Analytical notation for in-plane methods
Systemofreferenceaxes:
1,2,3 Plyreferenceaxes(onaxis).Principalaxesoforthotropyofaply.
x, y , z (x,y,z)Structuralaxes(offaxis).Platereferenceaxes.
Strains:
1

Columnmatrixofdirectandshearstrainsreferredto(1,2)coordinates
k
2
12 k
forkthlayer.

x Columnmatrixofdirectandshearstrainsreferredto(x,y)coordinates
k
y

xy
forkthlayer.
k
44
ECSSEHB3220Part2A
20March2011
Systemofreferenceaxes:

Stresses:
1

2
Columnmatrixofdirectandshearstressesreferredto(1,2)co
k
ordinatesforkthlayer.
12 k

x Columnmatrixofdirectandshearstressesreferredto(x,y)co
k
y
ordinatesforkthlayer.
xy k

Platecurvatures:
W Deflectionofplateinzdirection.
K
x
K K y Columnmatrixofcurvaturecomponentsofplate.

K xy k
Columnmatrixofstrainscomponentsofplate.

Plateloadingintensities:
N
x Columnmatrixofdirectandshearloadintensitiesreferredto(x,y)co
Ny
ordinates.
N xy
M Columnmatrixofbendingmomentintensitiesactinginxzandyz
x

M y planesandtwistingmomentintensityactingaboutxandyaxes,

M xy respectively.
Platestiffnesses:
A Inplanestiffnessmatrixforplate.
B Coupledinplaneandflexuralstiffnessmatrixforplate.
D Flexuralstiffnessmatrixforplate.

Qk Matrixrelatingstressandstrainin(1,2)coordinatesforkthlayer.

Qk Matrixrelatingstressandstrainin(x,y)coordinatesforkthlayer.

Plyproperties:
E1,E2 Moduliofelasticityofaplyin1and2directions,respectively.
G12 Shearstiffnessofaplyin(1,2)direction.
1 / 1 1 1 2
E foraply.
p
E1

Geometry:
45
ECSSEHB3220Part2A
20March2011
n Numberofpliesinplate.
hk Thicknessofkthlayer.
Angleofprincipalaxisoforthotropy,x,ofkthlayermeasuredfrom
k
platereferenceaxis,1.
mn Numberofgroupsoflayershavingidenticalorientations.
10.12 Stress-strain relations for unidirectional plies
10.12.1 Fibre-oriented co-ordinate system

Asthethicknessoflayersissmall,theyareassumedtobethinplatesinaplanestressstate.
Thecoordinatesystemisdefinedbytheprincipalmaterialdirections,asshowninFigure10.12.1.
Axes,12,coincidewiththelongitudinalandtransversedirections.Thismaterialsymmetryiscalled
orthotropicandonaxis.
Figure 10.12-1 - Definition of the co-ordinate system for the analysis of unidirectional
plies
10.13 On axis stress strain relations
10.13.1 General
Theelasticpropertiesofanorthotropicplyaredefined,usingmatrixnotation,byequationsgivingthe
complianceandmodulusmatrices.
[Seealso:10.11fornotation]
46
ECSSEHB3220Part2A
20March2011
10.13.2 Compliance matrix

1 S 0
S 11
S
12 1

k k k 2 S 12 S 0
22 2 [10.131]
12 k 0
0 S 66 k

12 k
[Compliancematrix]
Where:
12(shearstrain)isdeterminedby(u/y+v/x).
; S 12 12
1 1 1
S 11
; S 22 ; S 66
E 1 E 2 G 12 E 1
10.13.3 Modulus matrix

1 Q
1
Q 0
Q 11 12
k k 2 Q Q 0 2 [10.132]
12 22

12 k
k

12 k 0 0 Q
66
k
[Modulusmatrix]
Where:
Q 11
pE 1
; Q
22
pE 2
; Q
66 G12 ; Q12 p 12 E 2

10.13.4 Symmetry of compliance and modulus matrices

TherearefourindependentelasticconstantsusedtodescribethebehaviourofathinUDlayer.These
constants are determinedby experiment or from fibre and matrix propertiesusing micromechanical
formulae,[See:10.4].
10.14 Stress-strain relations for a ply of arbitrary

orientation
10.14.1 General
AUDunidirectionalplywithprincipalmaterialaxesXandYisdefinedwithrespecttoanarbitrary
systemofaxes1and2bytheangle1:X,asshowninFigure10.14.1.
47
ECSSEHB3220Part2A
20March2011
Figure10.141Definitionofaxes
From elementary mechanics of materials, the transformation equations for stresses and strains are
knowntobe:
k M k k
k
M
1 T
k k
Q k M k Q k M k
T
Where:
2
n 2mn
2
m2
M n m
2
2mn

mn m n
2 2
mn

mn
2 2
m n
M
1 T
n
2
m
2
mn

2mn m n
2 2
2mn
With: m cos ; n sin
48
ECSSEHB3220Part2A
20March2011
10.14.2 Off axis stiffness of a unidirectional ply

Theoffaxisstressstrainrelationis:
Q xx
x
Q xy
Q xS x
k Q k k y Q Q Q yS y
[10.141]

xy yy

xS
xy Q Q yS
Q
SS xy
Therelationshipbetweenthecomponentsoftheonaxisandoffaxisstiffnessmatrixis:
Q
Q xx 4 4
2 2 2 2
11
m4 n 2m n 4m n
Q yy
4 2 2 2 2
n m 2m n 4m n
2 2 Q22
4m n
2 2 4 4 2 2
Q xy m n mn m n

QSS
2 2
m3 n mn
2
2m n
2
m n 2 2 2 2 2

[10.142]
mn mn m n 2mn m n
3 3 3 3 3
m n Q
m n m n mn 2m3 n mn3
Q xS 3 3 3 3 12
mn Q
Q yS 66
Theserelationshipsarelimitedtotransformationfromtheonaxisorthotropicwhereshearcoupling
componentsarezero.
Theoffaxisstressstrainrelationintermsofcomplianceis:
x S 11 S 12 S 1S
x

y S 12 S 22 S 2S y [10.143]
S S S
s 1S 2S SS s
Engineering constants are determined for onaxis loaddisplacement relationships only and not for
offaxis.
Intermsofcompliance,applyingthesamestepsgives:
S
m n S
4 4 2 2 2 2

xx
m n 2m n 11

S yy n m n
4 4 2 2 2 2
m 2m n

m n S
2 2 2 2 4 4 2 2
S xy m n mn m n
m2 n2 2
22
[10.144]

8m n
2 2 2 2 2 2
S ss 4m n 4m n

2mn 2 mn m n mn m n S
3 3 3 3 3 3
S xs 2m n3
2 m n mn m n mn S
12

S ys 2mn 2m n
3 3 3 3 3

66
49
ECSSEHB3220Part2A
20March2011
AlloftheprocessesdescribeduptothispointaresummarisedasshowninFigure10.14.2.
Qxx... Transformation Q11...

Ex Inversion Inversion

S xx... Transformation S11... E1
ON-AXIS OFF-AXIS
Figure 10.14-2 - Transformation of stiffness and compliance matrix
10.15 Stiffness matrix for a laminate
10.15.1 General laminates

The laminate is considered to be a thin structure of constant thickness t, consisting of n plies of
thicknesshk(k=1,2,,n).
Assumingthattheelasticconstantsofaplyarenotinfluencedbyadjacentpliesinthelaminate,the
stressstrainrelationforthekthplyis:
k Q k k [10.15-1]
As usual in the theory of plates, forces and moments per unit length of the cross section of the
laminateareshowninFigure10.15.1anddefinedby:
N x t 2 x
n zk x

N y y dz y dz [10.152]
N t 2 k 1
z k 1
s s k s k
And:
M x t 2 x
n zk x

M y y zdz y zdz [10.153]
M t 2 k 1
z k 1
s s k s k
MxyMsandNxyNs(giveninEq[10.152]andEq[10.153])
50
ECSSEHB3220Part2A
20March2011
Figure10.151Inplaneforcesandmoments
Itisassumedthatthepliesareperfectlybonded;inconsequence,discontinuitiesintheinplanestrains
arenotpermittedattheinterfaceoftheplies.
Thestressstrainrelationshipsforalaminateare:
N x A11 A
12 A
13 B
11 B
12 B
13
1

N y A
22 A
23 B
12 B
22 B
23 2
N A B B B N A B
S 33 13 23 33
S or [10.154]
M x D
11 D
12 D
13 1 M B D
2

M y D 22 D
23
M S S
33
D
MatricesA,BandDaredefinedby:
A Q k zk zk 1
n
Matrix of extensional stiffness
k 1
B 1 2 Q k z 2k z 2k 1
n
Matrix of coupling stiffness
k 1
D 1 3 Q k z3k z3k 1
n
Matrix of flexural stiffness
k 1
Where:
zkisthehigherlevelofkthlayer,andzk1thelowerone,asshowninFigure10.15.2.
51
ECSSEHB3220Part2A
20March2011
Figure10.152Geometryofannlayeredlaminate
10.15.2 Symmetric laminates

Bending and stretching problems can only be treated separately in the case of laminates which are

symmetricingeometryandmaterialproperties;inthatcase B 0 .
Thestressesandstrainsinalaminateplycanbeobtainedby:
k z ; k M Tk k
k Q k z ; k M k 1 k
52
ECSSEHB3220Part2A
20March2011
10.15.3 Flow chart

Figure 10.15.3 shows a flow chart which summarises the steps taken in determining laminate
properties.
Figure10.153Calculationofalaminate
53
ECSSEHB3220Part2A
20March2011
10.16 Calculation methods with interlaminar stresses and

strains
10.16.1 Calculation with free-edge stresses

Composites failure produces a large amount of interlaminar separation. Delamination between
adjacent layers is induced by interlaminar stresses which can occur at free edges, such as borders,
holesorcracks.
Figure 10.16.1 shows that, even in the case of uniaxial loading, individual layers show biaxial
stressing. This arises from variations in stiffness, and transverse contraction of layers with different
fibre directions. At the free edges of the composite, stress normal to the edge vanishes (to maintain
equilibriumconditions).Asaresult,interlaminarstressesoccurwithintheinterfacelayers.
Figure10.16.2showsanexampleofinterlaminarstressesandtheirdecaybetweenthedifferentlayers
ofalaminate.However,thestressesatthefreeedgearenotrealisticandcannotbeappliedinafailure
criterion.
Analytical investigations have shown that the decrease of interlaminar stresses is significantly
influenced by the stacking sequence. Therefore it can be useful to have criteria that enable the
evaluationoftheinterlaminarstrengthofcompositeswithdifferentstackingsequences.
Figure 10.16-1 - Uniaxial loading of a composite
54
ECSSEHB3220Part2A
20March2011
Figure 10.16-2 - Typical decrease of interlaminar stresses with high peaks near the edges
55
ECSSEHB3220Part2A
20March2011
10.17 Qualitative evaluation of interlaminar strength for

design purposes
10.17.1 General
Currently,novalidandverifiedfailurecriterionforinterlaminarstressesexists.Thereforeitishelpful
to have data that enable the qualitative comparison of laminated plates with different stacking
sequences in order to minimise the risk of delamination. Such data for the uniaxial load case are
includedfromRef.[109].Figure10.17.1definestheloadandfibredirections.
Figure10.171Externalloadonacompositeplate
Characteristic data of the loaded laminate are combined as a parameter that is considered to be a
relevantvalueforevaluationofstrength:
Nf 1

E
[10.171]
1 z
Wheretheindividualfactorsaredefinedas:
f; a factor taking into account the interaction between the three relevant interlaminar stresses,
thus:
56
ECSSEHB3220Part2A
20March2011
2
y
2
f 1 yz

xy 3 xy
atthelocationxf,wherexfisdefinedbytherelation:
xf

0
xy
dx 0.5N
N;theforceflowisgivenby:
x

0
xy
dx N
(xy,yxandyarecalculatedbymeansofafiniteelementmethodorthetransfermatrixmodel).
;termedthecharacteristicdecaylength,with:

0
xy
dx 0.9N
E isthereferencestressinordertoformthenondimensionalparameter,where:
1 z
E1istheYoungsmodulusofa0layerinlongitudinaldirection.
ztheappliedstraininloaddirection.
Theterm E isroughlythestresswithinthe0layerofthelaminate.
1 z
The function enables one to choose from several considered composites, those with favourable
stacking sequences, i.e. those laminates which give low interlaminar stresses and reduce the risk of
delamination.
A high value of relates to a high interlaminar loading, a small value implies small risk of
delamination.
Thecalculationgivesthreegroupsoflaminatedplatesthatareinvestigatedfortheirparameter,[See:
Table10.17.1].UDmaterialdataaretypicalvaluesfora914C/T300laminate:
E 1
15
E 2
E 1
25
G 12
12
0.28
NOTE 1=longitudinal,2=transverse.

57
ECSSEHB3220Part2A
20March2011
10.17.2 Variation of fibre direction within a [, 0, ] laminate

Theresultsfor=f()aregiveninFigure10.17.2.Thecriticalinterfacesarethosebetweenthelayer
anda0layer.
Material:914C/T300
Figure10.172Determinationofthecriticalfibredirection
Theresultsindicatethatthemostcriticalcaseforinterlaminarseparationoccursfor=60ina[,
0,]laminate.
Interlaminareffectsdonotoccurif=0becausethemoduliarethesameineachlayer.Thedecayin
thefactorfor>60ismainlyaresultofdecreasinginducedforcesN.Themaximuminteraction
factorfis1.09.Thissuggeststhatthemainstressthatoccursasaresultofuniaxialloadingistheshear
stressxywithintheplanetransversetoloaddirection.
10.17.3 Variation of the thickness of the 0 layer within the [ 30,

0n, 30] laminate
Table 10.17.1 shows that this calculation gives three different laminates with the relating value
betweena30anda0layer:
58
ECSSEHB3220Part2A
20March2011
Table10.171Intralaminarstressesandstrains:valueforvariousstacking
sequences
Stacking Sequence

[ 30, 0 , 30] 0.0125
[ 30, 0 2 , 30] 0.0134

[ 30, 0 3 , 30] 0.0136
Owing to increasing N and with increasing number of 0 layers, the value of remains nearly
constant.Thisresultindicatesthattheriskofdelaminationdoesnotbecomesignificantlygreaterifthe
numberof0layerswithinthislaminatechanges.
10.17.4 Variation of the sequence of layers

Asymmetriclaminatewithfour30andtwo0layersenablesthreedifferentstackingsequences,as
showninTable10.17.2.
Table10.172Intralaminarstressesandstrains:
valueatinterfaceforvariousstackingsequences
Stacking Sequence at interface ""
[ 30, 0]s 0.013
[0, 30]s -0.013
[+ 30, 0, - 30]s -0.015
NOTE Theminussignofindicatesachangeinthedirectionofxyand
isnotrelevantforstrength.
Theresultsshowthe[+30,0,30]slaminatetobethemostcriticalcase.
10.18 References
10.18.1 General
[101] R.M.Jones
MechanicsofCompositeMaterials
McGrawHill,1975
[102] R.Frster&W.Knappe
ExperimentelleundtheoretischeUntersuchungenzur
RibildungsgrenzeanzweischichtigenWickelrohrenausGFKunter
Innendruck
Kunststoffe,61(8),1971
59
ECSSEHB3220Part2A
20March2011
[103] H.Schneider
ExperimentelleundtheoretischeBetrachtungenzurErmittlungvon
ElastizittskenngrenvonCFKLaminaten
DGLRVortragNr.81058,Aachen,Mai1981
[104] A.Puck
ZurBeansprungundVerformungvonGFKMehrschichtenverbund
Bauelementen
Kunststoffe,57(4),1967
[105] S.W.Tsai&H.T.Hahn
IntroductiontoCompositeMaterials
Technomic,1980
[106] LuftfahrttechnischesHandbuch
HandbuchStrukturberechnung(HSB)
Publisher:IABGOttobrunn,Germany
[107] J.E.Ashton,J.C.Halpin&P.H.Petit
PrimeronCompositeMaterials:Analysis
Technomic,1969
[108] B.D.Agarwal&L.J.Broutman
AnalysisandPerformanceofFibreComposites
JohnWileyandSons,1980
[109] MBB/ERNO
PreviouslyunpublishedworkbyS.Dieker
DepartmentRB514,1984
60
ECSSEHB3220Part2A
20March2011
11
Strength prediction and failure criteria
11.1 Introduction
11.1.1 Micro-mechanical strength models
11.1.1.1 Fibres
Similar fibres under load do not all have the same fracture strength neither do they fracture in the
sameplace.Astatisticalanalysisisnecessarytorationallydefinethestrengthofacomposite.
11.1.1.2 UD composites
Based on experimental observations of UD composite failure, statistical strength models have been
developedfortensionandcompressioninthefibredirectionaswellasfortensioninthetransverse
direction,[See:11.2;11.3;11.4].
Themethodsdescribedarenotyetadequatefordefinitivequantitativestrengthprediction,butthey
canbeusefulforanassessmentoftherelativemeritsofdifferentmaterials,[Seealso:WWFE].
11.1.1.3 Laminate
Thestrengthdesignverification(stabilityisanotherissue)foralaminatethisisperformedatdifferent
levels:
Stress:atalocalpointinamaterial,bystrength,
Stressconcentration:stresspeakatanotchorajoint,bynotchedstrength(Neuber),
Stressintensity:atadelaminationsite,byfracturetoughness.
11.1.2 Lamina failure models
11.1.2.1 Complex stressing

Designvaluesfortheultimatestrengthintheaxialandtransversedirections,togetherwithshear,are
determined experimentally. However, experimental information of this type is not sufficient for the
strengthcharacterisationofsingleUDlaminabecauseinmoststructuralapplicationsthematerialis
subjectedtocomplexstresses,i.e.stresscomponentsactingsimultaneouslyinmorethanonedirection.
Thestresscomponentsaredescribedbythreemutuallyperpendicularnormalstressesandthreeshear
stresses acting on an element. The problem for the designer is to predict failure under complex
stressesusingconventionaluniaxialstrengthdata.
61
ECSSEHB3220Part2A
20March2011
11.1.2.2 Failure theories
Failure theories are mathematical functions that connect such strength data with the actual state of
stress or strain to predict the failure of a structure under complex loading conditions. Numerous
theories have been proposed, [See also: 11.8]. Also, reference is made to an extensive assessment
initiatedbySodenandHinton,Ref.[1116].
Inastudylastingfor12years,WWFEworldwidefailureexerciseconsiders3Dfailuretheoriesthat
include: failure conditions to assess multiaxial states of stress, nonlinear stressstrain curves of UD
materialasinput,andnonlinearcodeinthestructuralanalysisofthelaminate.
Thenonlinearstressstraincurveshaveahardeningbranch(fromtestingofisolatedtestspecimens,
which break due to weakestlink behaviour) and a softening branch (due to the redundancy
incorporatinginsitueffectsofthelaminamaterialembeddedinthelaminate).
Althoughissuesofwhetheramodelisrightorwrong,trueoruntrueareconsideredtobeoutsidethe
WWFEremit,guidelinesfordesigners(fromWWFEI)illustratehowsomerobustmodelscouldbe
sufficient to describe all of the conditions that the 1st study considered, Ref. [1137]. However, the
resultsfromWWFEIareonlypartoftheoverallstory.Therobustnessofmodelsisundercontinual
assessmentandevolution.AnoverviewoftheentireWWFEstudyisdescribed,Ref.[1138].
11.1.3 Failure criteria studies
11.1.3.1 ESA failure criteria study

An ESAfunded study provided detailed guidelines for the proper use of failure criteria during the
differentstagesofdesignofspacecraftcompositestructures,forcompositesmadeoffibrereinforced
polymerorCMCceramicmaterials,Ref.[1118],[1122].
[See:64.5]
11.1.3.2 WWFE worldwide failure exercise

As part of WWFE, 19 theories were evaluated and a comparison made between accuracy and
performanceofpredictionsusingseveralexamplecases.Adviceisalsoprovidedontheuseofvarious
theories for engineering design, Ref. [1117], [1137]. The failure criteria employed in WWFE are
intended to represent design methodologies capable of describing the modes of failure, failure
locations,ultimatestrengthsanddeformationsuptoultimatefracture,Ref.[1138].
62
ECSSEHB3220Part2A
20March2011
11.1.4 Summary of World Wide Failure Exercise (WWFE)
11.1.4.1 WWFE I
Completedin2004,WWFEIconsidered,Ref.[1138]:
Useofmicromechanicsforpropertyprediction.
Predictionofbiaxialfailureofalaminainisolation.
Predictionof2Dmodesoffailure.
Predictionoffailureenvelopesforwholelaminates.
Matrixfailureintension,shearandcompression.
Materialnonlinearity.
Postfailuremodellingunder2Dstresses.
Predictionoffibrefailure.
SomeoftheconclusionsofWWFEI,whichinstigatedWWFEIIandIII,were,Ref.[1137],[1138]:
Improvementsintheoriesbyidentifyingweaknesses:
50%oftheoriesweremodified.
Theories,adoptedfor40years,modifiedforthefirsttime.
Guidelinesfordesignersontheaccuracyandboundsofapplicabilityforcurrentfailuretheories.
Manymodifiedtheoriesnowembeddedinnumericalsoftwarepackages.
Reviseddefinitionsoffailure,creatingnewinsights.
Bringingtogetherkeymembersoftheworldcommunityfordiscussionandargumenttoadvancethe
science.
Ideasfordesignersandacademicstofurtherimprovetheoreticalandpracticalareas.
Highlightingofgapsinexperimentalresults.
Highlightingofsomeareaswherealackoffundamentalknowledgeexists.
11.1.4.2 WWFE II
WWFEIIaimsto,Ref.[1138]:
Extendthephilosophyusedpreviouslytoinclude3Dfailurecriteria.
ResolvethegapsidentifiedinWWFEI.
Identifymainworkerson3Dfailurecriteria.
Validateandbenchmarktheirmodels.
Identifyanygapsinthecurrentmodels.
Identifyspecificshortfallsinexperimentaldataortestresults.
Stimulatethecompositescommunitytoprovidebettertools.
Providedesignerswithguidanceforpredictingaccuratestrengthsin3Dstressstates.
63
ECSSEHB3220Part2A
20March2011
Itconsiderstheparticularpoints,Ref.[1138]:
Applicabilityofcompositefailurecriteriaforisotropicmaterials.
EffectsofpressureonshearstrengthanddeformationofUDlamina.
Failureofunidirectionallaminaunderhydrostaticpressure.
EffectofthroughthicknessstressonbiaxialfailureofUDlamina.
Effectsof3Dstressesonthefailureofmultidirectionallaminates.
Deformationoflaminatesunderhydrostaticpressure.
3Delasticconstantsofmultidirectionallaminates.
Effectsoflayuponthroughthicknessstrengthoflaminates.
Failureoflaminatesundercombinedthroughthicknessandshear.
11.1.4.3 WWFE III

WWFEIIIisconcernedwithdamage,fractureandcontinuummechanicsofcompositesandaimsto,
Ref.[1138]:
Extendthephilosophyusedpreviouslytoincludedamage,fractureandcontinuummechanics
modelsforcomposites.
SolvethegapsidentifiedinthefirstWWFE.
Identifymainworkersondamageandfracturecriteria.
Validateandbenchmarktheirmodels.
Identifyanygapsinthecurrentmodels.
Identifyspecificshortfallsinexperimentaldataortestresults.
Stimulatethecompositescommunitytoprovidebettertools.
Provide designers with guidance for predicting damage, fracture and deformation in
composites.
Thespecificissuescoveredinclude,Ref.[1138]:
Damageinitiationandevolution.
Matrixcrackinitiationandcrackdensityevolution.
Delaminationinitiationandpropagation.
Effectsofplythicknessandconstraints.
Effectsofplystackingsequence.
Crackingunderthermalloading.
Monotonicloading,unloadingandreloading.
Failureatstressconcentrations(e.g.openhole).
Statisticalandprobabilisticnatureoffailure.
Leakageprediction.
64
ECSSEHB3220Part2A
20March2011
11.2 Tensile strength of UD composites in fibre direction
11.2.1 General
Thestatisticaltensilefailuremodeltakesintoaccountthe:
Lengthstrengthrelationship,
Statisticalvariationoffibrestrength,and
Differencebetweenthestrengthofthebundleandanaveragestrengthofthefibres.
Thusthesemodelspredictthestrengthofthecompositebytakingintoconsiderationtheinsitufibre
strength.
AWeibulldistributionprovidesagoodfittoexperimentaldata:
f La 1e La

[11.2-1]
Where:
andareparameterstofittheexperimentalfibretensilestrengthdata,andLis
thefibrelength.
Asummaryispresentedofsomerecentmodels,Ref.[1119].
11.2.2 Weakest-link failure model

Theweakestlinkfailuremodelassumesthatthebreakageofasinglefibreresultsinlocalisedstress
concentrations, which cause fracture of adjacent fibres. In turn the failure of these fibres can affect
further fibres and eventually leads to an overall failure. This model leads to a realstic failure
predictiononlyinthecaseofsmallsampleswithfibresofanarrowdispersionofstrength.
Forpracticalmaterialsinrealisticstructures,thestrengthpredictedbytheweakestlinkfailuremodel
istoolow.
11.2.3 Cumulative weakening failure model

If failure does not occur at the breakage of the weakest link, then the broken fibre causes a
redistributionofstressesnearthefracture,givingrisetothecumulativeweakeningfailuremodel.
Stress transfers from one end of the broken fibre, past the break, to the other. This stress transfer is
achieved by high shear stresses in the matrix near the fibre break; as shown in Figure 11.2.1,
Ref.[111].
65
ECSSEHB3220Part2A
20March2011
Figure 11.2-1 - Fibre reinforced composite: Tensile failure model
Foranelasticfibreandaperfectlyplasticmatrixanestimateoftheineffectivelengthisobtainedby:
df
f 4 y [11.2-2]
Where:
fisthefibretensilestrength,dfisthefibrediameter,andyisthematrixshearyieldstrength.
For fibres having a strength distribution as given in Eq [11.21], the cumulative weakening failure
modelestimatesthecompositetensilestrengthbasedonfibreareaas:
* e 1 / [11.23]
66
ECSSEHB3220Part2A
20March2011
11.2.4 Fibre break propagation model

Theloadconcentrationinfibresadjacenttoabrokenoneincreasestheprobabilitythatoneormoreof
themwillalsobreak.Thatleadstothe fibrebreakpropagationfailuremodel,whichstatesthatthe
propagationoffibrebreaksasthemechanismoffailure.
The occurrence of the first fracture of an overstressed fibre is used as a failure criterion. This can
providegoodcorrelationwithexperimentaldataforsmallspecimens,butitgivesratherlowstrength
valuesforlargevolumesofmaterials.
11.2.5 Cumulative group mode failure model

Thecumulativegroupmodefailuremodelincorporatesthreeeffects:
Variabilityoffibrestrengthresultsindistributedfractureswellbelowthecompositestrength.
Loadconcentrationsinfibresadjacenttobrokenfibresinfluencethegrowthofthecrackregion
toincludeadditionalfibres(fibrebreakpropagation).
Highshearstressesinthevicinityofbrokenfibrescausematrixshearfailure,whichhasacrack
arrestingeffect.Thiscanleadtogroupsofbrokenfibres.
This model can be viewed as a generalisation of the Cumulated Weakening Failure model with
chains of bundles instead of single fibres. For practical application the model is complicated by
bundles of various sizes, where both the number of broken fibres in a bundle and the ineffective
lengthofthatbundlevary.Consequently,statisticalparameterstodescribethemodelaredifficultto
obtain.
11.2.6 Status of models

Although the statistical models described are not yet adequate for definite quantitative strength
prediction,theycanbeusefultoassessrelativemeritsofdifferentmaterials.
Equation [11.23] can be applied in connection with the ineffective length of Equation [11.22] to
estimatetheactualstrengthandtoindicatetheroleofthepropertiesofeachconstituent.
11.3 Compressive strength of UD composites in fibre

direction
11.3.1 General
Whenfibrereinforcedcompositesaresubjectedtocompressiveloads,themodeoffailureappearsto
befibrebuckling.
In composites with very low fibre volume content, fibre buckling can occur even when the matrix
stressesarein the elasticrange. However, at practical volume fractions, (Vf> 0.40), fibre bucklingis
generally preceded by matrix yield and/or constituent debonding and matrix microcracking,
Ref.[1119].
67
ECSSEHB3220Part2A
20March2011
Fibre buckling can also result from shrinkage stresses developed during curing of the composite.
Thesearegeneratedbythedifferentialthermalcoefficientofexpansionofthematrixandfibres,[See:
Chapter12].
Twofibrebucklingmodesarepossible,asshowninFigure11.3.1.
Figure 11.3-1 - Buckling modes of fibres under compression
11.3.2 Extension mode buckling

Theadjacentfibresinacompositecanbuckleindependentlyandoutofphaserelativetoeachother.
Theresultingstrainsinthematrixarepredominantlyextensional.Thematrixalternatelydeformsin
extensionand compression, transverseto the fibres.This is called extension mode buckling, andis
only possible when the interfibre distance isquite large, i.e. when the fibre volumefraction is very
low.
11.3.3 Shear mode buckling

Thisbucklingmodeismorecommonandoccursatmostpracticalfibrevolumefractions.Inthiscase
the transverse deformation of adjacent fibres is in phase. The resulting strains in the matrix are
predominantlyshearstrains;sothereforecalledshearmodebuckling.
68
ECSSEHB3220Part2A
20March2011
11.3.4 Analysis of compression failure
11.3.4.1 General
For both buckling modes, the fibres are regarded as plates with thickness h separated by matrix 2c
wide.EachfibreissubjectedtoacompressiveloadPandisoflengthL.Thefibresarealsoregardedas
beingmuchstifferthanthematrix,i.e.Gf>>Gm;sothatfibreshearingdeformationsarenegligible.
The analysis to determine the fibre buckling load in each mode is based on the Rosen theory. This
equatesthechangeinstrainenergyforthefibreandmatrixtotheworkdonebythefibreforceduring
deformationtoabuckledstate.
11.3.4.2 Extension mode buckling

Forextensionmodebuckling,thecriticalstressinthecompositeis:
V E E
2V
31V
f m f
c max f
[11.31]
f
11.3.4.3 Shear mode buckling

Forshearmodebuckling,themaximumcompositestressis:
G m

1V
c max [11.32]
f
BothofthecompositestressexpressionsareshowninFigure11.3.2foraglass/epoxycomposite,Ref.
[113].
69
ECSSEHB3220Part2A
20March2011
Figure 11.3-2 - Comparison of predicted and experimental values of longitudinal

compressive strength
Eq[11.32]holdsforalinearelasticmatrixmaterial.Amoregeneralresultisobtainedbymodelling
thematrixasanelastic,perfectlyplasticmaterial,Ref.[1119].Thesecantmodulusvalueateachstrain
levelcanbeassumedtogoverntheinstability.Theseassumptionsleadto:

Vf Ef y
c max
3(1V f
[11.33]
where:yisthematrixyieldstress.
70
ECSSEHB3220Part2A
20March2011
Shear mode buckling gives the lowest strength for composites over a wide range of fibre volume
fractions(actualcomposites);transverseorextensionalmodegovernsthecompositestrengthforlow
fibrevolumefractionsonly.However,Eq[11.32]predictsvaluesmuchgreaterthanthosemeasured
inpractice.Initialfibremisalignementcanbeidentifiedasamajorreasonforthediscrepancy.Thisis
takencareofby:
G

m

1 0 y
c max [11.34]
where:oisthefibremisalignment;yisthecompositeshearyieldstrain.
In the literature the values given for 0 are usually 2 to 3 degrees, and the compressive strengths
predictedbymeansofEq[11.33]agreewellwithexperimentalresults,Ref.[1120].
11.3.4.4 Debonding
Theobservationthattransversesplittingordebondingisapossibleinitiatingeventcanbeformulated
asasimpletheoreticalexpression.Thisgivesthemaximumcompositestrengthas:
E V E V 1V 13
f f m m f mu
[11.35]
c max
V V f f m m
Where: muisthematrixultimatestrain.
Thepredictedstrengthisgreatlyinfluencedbythematrixultimatestrainandisalsoamaximumata
givenfibrevolumefraction.
11.4 Transverse tensile strength of UD composites
11.4.1 General
Thelongitudinalstrengthandstiffnessaregoodbecauseofthedominantroleplayedbythefibresin
this direction. Load is shared between the fibres and the matrix. However, because of their higher
strengthandstiffness,fibrescarryamajorportionoftheload.
Nosuchloadsharingtakesplacewhenthecompositeissubjectedtotransverseloading.Instead,the
matrixcarriestheloadwhilstthefibresactasstressconcentrators,andcompositefailureoccursata
much lower strain than that of the unrestrained matrix. Therefore, unlike longitudinal strength and
stiffness,andtransversemodulus,thetransversestrengthisreducedbythepresenceofthefibres.
71
ECSSEHB3220Part2A
20March2011
11.4.2 Prediction of transverse tensile strength

tu
The transverse tensile strength of the composite, R2 , is predicted on the assumption that it is
controlled by the matrixultimate strength m . It isfurther assumed that the composite strength is
u
lowerthanthematrixstrengthbyafactor,S,knownasthestrengthreductionfactor.
Thusthecompositetransversestrengthcanbewrittenas:
mu
R tu
2 ,intermsofstresses,
S
And:
2tu Sm1 ,intermsofstrains
Where: S1isthestrainmagnificationfactor.
WhenPoissonsratioeffectsareneglected,theequationsforSandS1takethesimplifiedforms:
1V 1 E E
1 E E
S
1 4V
f m f
12 [11.41]
f m f
1
S1
1 4V f
1 E E
12
m f
[11.42]
ThusonceSorS1isknown,thetransversestrengthintermsofstressesorstrainscanbecalculated.
11.4.3 Empirical analysis

Afurtherempiricalapproachforthepredictionoftransversetensilestrengthoffibrouscompositesis
describedbyNielsen,Ref.[114].Thecompositestraintofailurecanbeapproximatedas:
2tu mB 1V 13
f
[11.4-3]
Where:
2tu cB is thebreakingstrainofthecompositetransversetothefibres.
mB is the matrix breaking strain.
72
ECSSEHB3220Part2A
20March2011
If the matrix and composite have linear elastic stress strain curves, the Eq [11.43] can be stated, in
termsofstress,as:
R2tu mB E 1V
2 13
[11.4-4]
f
E m
Where:
E2=transversemodulusofthecomposite.
Em=matrixmodulus.
Both Eq [11.43] and Eq [11.44] assume perfect adhesion between phases, with failure occurring by
matrixfractureatorneartheinterface.
11.5 Static strength criteria for composites

Staticstrengthcriteriaforcompositesuseuniaxialstrengthdatatospecifythefailureenvelopeundera
complexstateofstress.
A large number of criteria have been proposed, which in certain cases predict very different failure
strengths.Onlylimitedexperimentaldataisavailabletojudgethesuitabilityofcriteria,especiallytrue
forthecompressioncompressionregion,wherethecriteriashowthelargestdifferences.Inthisregion
experimentsareextremelydifficulttoperformandavailabledatashouldbeusedwithcaution.
TheextensiveWWFEassessment,Ref.[1117],givessomeindicationofthestrengthsandweaknesses
ofanumberofcriteria.TheWWFEIguidelinesgiveusefulinformationtodesignerswhenselecting
anappropriatecriterioninagivensituation(onlywithintheconstraintsofWWFEI),Ref.[1137].It
alsomakesobviousthatnoneofthecriteriaisrightorwrong,butaremoreorlesssuitable.
[Seealso:11.1forWWFE]
11.6 Analytical notation for static strength criteria for

composites
11.6.1 Co-ordinate system

Figure11.6.1definesthenotationcoordinatesystem.
73
ECSSEHB3220Part2A
20March2011
Figure 11.6-1 - Analytical notation: Co-ordinate system
11.6.2 Formulae
E1, E2 modulusofelasticityin1and2directions,respectively
1 and 1
E1 E2
S 11 S 22
shearmodulus, 1
G12 G12
S 66
S11, S22, S12,
complianceoflayermaterial.NotethatS12=S21
S21, S66,
R12su shearstresstofailureoflayerin12plane
R1u directstresstofailureinfibredirection
R2u directstresstofailurenormaltofibre(inplane)
strain
12u directstraintofailurein12plane
1u directstraintofailureinfibredirection
2u directstraintofailurenormaltofibre(inplane)
12 longitudinalPoissonsratio,ratiooftransverse(inplane)
straintolongitudinal(fibredirection)strainduetoanapplied
longitudinaldirectstress: 12 S 12
S 11
stress
74
ECSSEHB3220Part2A
20March2011
1 directstressinfibredirection
2 directstressnormaltofibre(inplane)
12 shearstressin12plane
1, 2, 3 coordinatenotation,[See:Figure11.6.1].
Subscripts:
c compression
t tension
s shear
1, 2, 3 refertocoordinatesystem,[See:Figure11.6.1].
Directfailurestressesarealwayspositive,intensionaswellasincompression.
11.7 Different types of failure criteria
11.7.1 General
AgreatmanyfailurecriteriaforUDcompositescanbefoundintheopenliterature.Inthecontextof
thishandbook,[See:11.8],theyaregroupedas,Ref.[1135]:
Independentconditions:
Maximumstress.
Maximumstrain.
InteractiveconditionsPureinterpolativeconditions:
Tensorcriteria.
Considerationofmaximumstrengthinfibredirection.
InteractiveconditionsPhysicalconsiderations:
Hashinsfailurecriterion.
Pucksactionplanefailurecriterion.
Simplifiedparabolicmodel.
CuntzeFMCbasedUDfailurecriterion,Ref.[1136].
Zinovievsfailurecriterion(notdescribedhere).
Bogettisfailurecriterion(notdescribedhere).
Tsaisfailurecriterion(notdescribedhere).
Topic11.8presentsanuptodateperspectiveoffailurecriteriacompiledfrommorerecentwork,Ref.
[1117],[1121],[1135],[1136],[1137].
75
ECSSEHB3220Part2A
20March2011
11.7.2 Evaluation studies

An ESTECfunded study considered failure criteria for nonmetallic materials; both FRP and CMC,
Ref.[1122].
[Seealso:64.5]
InWWFEI,theintentionwastovalidatefailurecriteriaatthe2Dlevel,Ref[1117],[1137].
In the WWFEII a test data platform is provided to validate 3D UD criteria. Various test cases are
planned with carbon and glassfibre reinforced composites having different epoxy matrices and
stackingsequences.Theloadingconditionsincludeuniandbiaxialtensionandcompression,torsion,
shearandhydrostaticloading,andcombinationsthereof.
Noteshouldalsobetakenofthereducedpropertiesinthethicknessdirection,Ref.[1136].
[Seealso:11.1WWFE]
11.8 Overview - Failure criteria
11.8.1 Introduction
ThisoverviewwaspreparedbyDLRin2005,Ref.[1135],toreflectthestateoftheartinfailurecriteria
togetherwithpointsarisingfromtheongoingWWFEWorldWideFailureExercise,Ref.[1117],[11
37].Cuntzesfailurecriterionwasaddedin2009,Ref.[1136].
ThecriteriaarelimitedtoUDlayershavingidealstraightfibres.Misalignmentandondulationarenot
takenintoaccount.
Someofthefailurecriteriainopenliteraturearebasedonfracturemechanicsaspectswhereasothers
evaluatethelocalstateofstressorstrain;onlythelattertypeisconsideredhere.
Thesecanbegroupedfurtherinto:
Independentconditionswheresinglestress/straincomponentsarecheckedagainstfailure.
Interactive conditions where several stress/strain components contribute to the effort to reach
failure.
Withineachgroupareanumberofcriteria,i.e.
Independentconditions:
Maximumstress.
Maximumstrain.
InteractiveconditionsPureinterpolativeconditions:
Tensorcriteria.
Considerationofmaximumstrengthinfibredirection.
76
ECSSEHB3220Part2A
20March2011
InteractiveconditionsPhysicalconsiderations:
Hashinsfailurecriterion.
Pucksactionplanefailurecriterion.
Simplifiedparabolicmodel.
Cuntzesfailurecriterion.
Zinovievsfailurecriterion(notdescribedhere).
Bogettisfailurecriterion(notdescribedhere).
Tsaisfailurecriterion(notdescribedhere).
11.8.2 Independent conditions
11.8.2.1 Maximum stress

Thecriterionstatesthatfailureoccursassoonasastresscomponentintheprincipalmaterialdirection
reachestherespectivestrength.Inthecaseofplanestress,thisis:
1 X t or Xc
or 2 Yt or Yc [11.81]
or 6 Q
11.8.2.2 Maximum strain

The criterion is similar to the maximum stress criterion, except that strains are limited rather than
stresses,with:
X (Xc ) maximumtensile(compressive)directstraininfibredirection
t
Y (Yc ) maximumtensile(compressive)directstrainintransversedirection
t
Q maximumshearstraininthe12plane
Failureispredictedif:
1 Xt or Xc
or 2 Y or Y
t c [11.82]
or 6 Q
77
ECSSEHB3220Part2A
20March2011
Thisisnotidenticaltothemaximumstresscriterion.Forlinearelasticityandaplanestateofstressthe
criterioncanbereformulatedintermsofstresses,resultingin:
1 122 X t or Xc
or 2 121 Yt or Yc [11.83]
or 6 Q
11.8.3 Interactive conditions Pure interpolative conditions
11.8.3.1 General
Inmostcasesthestateofstressinastructureconsistsofmorethanjustonecomponent.Interpolative
conditions assume that there is an interaction between the stress components at the analysed point,
contributinginsomewaytotheefforttoreachfailure.
11.8.3.2 Tensor criteria

A general criterion for anisotropic material is a tensor polynomial written in stresses, usually
truncatedafterthequadraticterm:
6 6
B
i i
i 1

j1
A
ij i j 1
[11.84]

WhereAijandBiarematerialconstants.
For a plane state of stress aligned with the principal axes of orthotropy, the general expression is
reducedto:
A1112 2A1212 A 2222 A6662 B11 B22 1 [11.8-5]
Terms containing shear stresses linearly are omitted since shear stress reversal does not affect the
strength.WithspecialassumptionsandadditionalconditionsEquation[11.83]producesmanyofthe
establishedcriteria.Substitutingthedirectandshearfailurestressesproducesfivelinearequations:
78
ECSSEHB3220Part2A
20March2011
A11 1 B1 1 1
X t Xc X t Xc
A 22 1 B2 1 1 [11.86]
Yt Yc Yt Yc
A66 12
Q
AdditionalbiaxialtestsareneededfordeterminingtheremaininginteractiontermA12.Sincetheyare
difficulttoperform,variousvalueshavebeenproposedinstead.
Amongtheseare:
0.5
A12 (Hoffman)
Xt Xc
0.25
A12 (DeTeresa/Larsen,Ref.[1123])
Xt Xc
0.5
A12 (Tsai/Hahn)
X t X c Yt Yc
A12 0 (Tsai/Wu)
11.8.3.3 Consideration of maximum strength in fibre direction

With A12 0 the tensor criterion predicts strength values which can be much higher than Xt or Xc,
respectively. In order to avoid such unrealistic results Ref. [1124] limits the tensor criterion by the
conditionsforfibrefailure:
12 X 2t 12 Xc2 [11.8-7]
By differentiating between fibre failure and interfibre failure, a more physically based criterion is
derived.
79
ECSSEHB3220Part2A
20March2011
11.8.4 Interactive conditions - Physical considerations
11.8.4.1 Hashins failure criterion

ForUDmaterialthefailurecriterionshouldbeinvariantunderanyrotationaroundthefibredirection.
Based on this, Hashin, Ref. [1125], uses four stress invariants to determine failure limits that are
differentintensionandcompression,inthefibredirectionaswellasthetransversedirection.Under
planestressconditions,theseare:
Tensilefibremode(1>0)
12 62
1
X2t Q2
Compressivefibremode(1<0)
12
1
Xc2
Tensilematrixmode(2>0)
22 62
1
Yt2 Q2
Compressivematrixmode(2<0)
2 22 62
Yc 1 2 1
4Q223 Yc 4Q223 Q2

WhereQ23istheshearstresstofailureinthe23direction.
80
ECSSEHB3220Part2A
20March2011
11.8.4.2 Pucks action plane failure criterion
In recent publications in connection with the WWFE WorldWide Failure Exercise, Puck and
Schrmann,Ref.[1126],proposedconditionsforfibrefailure(FF)andforinterfibrefailure(IFF).
TheFFcriterionisbasedontheassumptionthatthefibrestressf1atfailureisthesameforauniaxial
aswellasforaplanestateofstress:
1 E1 m 1 for()>0
X t 1 12 2 f12 Ef1 f 2
1 E1 m 1 for()<0
Xc 1 12 2 f12 Ef1 f 2
Where:
Ef1=Youngsmodulusofthefibre;
f1=Poissonsratioofthefibre.
Thefactormfaccountsforslightlyhigherstresses2inthefibresthaninthematrix(glassfibresmf
1.3;carbonfibresmf1.1).
IFFiscausedby2and6alone.Thecriteriondistinguishesthreedifferentfailuremodes:
ModeA:
2
p()
1 62
22 p() 2 f w
Yt Q Q 2 Q

0
for20;fractureplaneangle fp
ModeB:
2
1 2 p() p() f
w
Q 6 2 2

2 R A
0
6 0
6c
for2<0and ;fractureplaneangle fp
ModeC:
81
ECSSEHB3220Part2A
20March2011
2
2
6 Yc
2 fw
Yc ( )
2 1 p( ) Q 2

6 6c fwRA
for2<0and 0 ;fractureplaneangle fp
2 R A

(2 )
Evaluatingtheseconditionsneedsthetwotangentstothe(26)curveat2=0:
d d
p() 6 for20and p( ) 6 for20.
d2 d 2
2 0 2 0
These inclination parameters are determined from fracture curves for different load combinations.
TypicalvaluesforthermosettingmatrixasprovidedbyPucketal,Ref.[1127],are:
p() 0.3 p() 0.25

ForGFRP: ;
ForCFRP: p() 0.35 ; p() 0.3
Y
R A Q() 1 2p() c 1
2p Q
) R A
p( p()
Q
)
6c Q 1 2p(
The weakening factor fw takes care of the reduced IFF strength due to premature breakage of single
fibres.Iftestresultsaremissing,theexpressionusedis:
n

f w 1 1
1d

where 1d 1.1Xt or 1.1Xc for tensile or compressive stresses respectively and n = 8 or n = 6 for
matriceswithhighorlowfracturestrain.
82
ECSSEHB3220Part2A
20March2011
11.8.4.3 Simplified parabolic model
Pucks model can be simplified by specifying only one inclination parameter, p, which has some
stabilityadvantages.AcorrespondingcriterionbasedontheworkofJeltschFricker,Ref.[1128],was
presentedbyCuntzeetal,Ref.[1129].
11.8.4.4 Cuntze FMC-based UD failure criterion

Aglobalfailurecondition,i.e.F({},{R})=1,hasasingleconditionrepresentingthefailuresurfacefor
allfailuremodes,hencetheinteractionbetweenseveralfailuremodesismathematicallyfixed.Mode
failure conditions, F ({},R mode) = 1, have a separate failure condition for eachmode (asused in the
CuntzeFMCFailureModeConcept).
a. ForUDmaterialthe3Dplystressvectoris:
( 1, 2 , 3 , 23 , 31, 21 )T
andthestrengthvector:
b. With average strength properties applied in the stressstrain analysis and strength design
allowables (minimum values according to MILHDBK 5, no bar over R), for strength analysis
(designverification):
R||t ( X t ), R||c ( X c ), Rt ( Y t ), Rc ( Y c ), R|| ( S )
ThisnotationisinaccordancewiththeGermanguidelineVDI2014,Ref.[1133].
Stressstrainanalysisgivestheaverage(typical)structuralresponsewhichisbestwhenutilising50%
values.Therefore,averagephysicalpropertiesareused,includinganaveragestressstraincurveanda
R value.
According to material symmetry requirements, transverselyisotropic UD material has 5 elasticity
properties,5strengths,and5failuremodes.Eachfailuremodeisdominatedbyonestrength.
For generally orthotropic fabric material this number is 9. If the warp reinforcement equals the fill
(weft)reinforcementthereducednumberoffailuremodesis6.
Therefore,theFMCemploys5failuremodesintheUDmaterialcaseandeither9or6failuremodesin
the orthotropic case. Whether the ply can be modelled as a transverselyisotropic material or as an
orthotropicmaterialshouldbeestablished.
ForUDmaterial,the5(brittlefracture)failuremodesconsistof2FF(FibreFailure)and3IFF(Inter
FibreFailure)modes.
83
ECSSEHB3220Part2A
20March2011
ThesefailureconditionsaredefinedinFigure11.8.1.Thesuperscriptsandindicateeithernormal
or shear stress dominated failure. Strength and elasticity properties are shown as symbols to avoid
confusion.
FF1 F||
FF2 F||
IFF1 F
IFF2 F
IFF3 F||
Figure 11.8-1 Cuntze failure criteria - Schematic diagram of failure modes in

transversely-isotropic UD material
Each failure mode is characterised by one strength and therefore, for a nonisotropic material, an
equivalentstressexistsforeachofthe5failuremodes:
mod e
eq ||
eq
||
, eq
, eq , eq , eq||
T
84
ECSSEHB3220Part2A
20March2011
In practice Material Stress Effort gives a better estimate of the remaining loadcarrying capacity of
theplymaterial.Thisislinkedtotheequivalentmodestress:
Eff mode = eqmode/ Rmode
Only Material Stress Effort is applicable in nonlinear strength analysis, not the material reserve
factor(thisispermittedonlyforlinearanalysiswherestressisproportionaltoload,[Seealso:Pucks
criterion]).
FF1(tension)andFF2(compression)aremoreaccuratelytermedfilamentmodesbecauseinthecase
ofbiaxiallylateralcompressionthefilamentcanbreakintensionduetothePoissonseffectwithout
anyappliedaxialstress.Replacing1inthe3Dcase,shownbelow,enablesFEAmacrostresslevelto
beusedinsteadofmicromechanicalfilamentstress.
The IFF modes, also termed matrix modes, IFF1 (lateral tension) and IFF3 (shear) are relatively
benign failure modes. However, the 2cdependent wedge failure IFF2 (lateral compression) of an
embedded lamina can also lead directly to final failure, Ref. [1139], [1140], [1141], because it can
causedelaminationofthestackedlaminate.
3DcaseUDmodefailureconditions
Where:Effmode=100%=1formodefailure
FF1 Eff || 1 / R||t eq

||
/ R||t with 1 1t E||
FF2 Eff || 1 / R||c eq|| / R||c with 1 1c E|| ,
IFF1 Eff [( 2 3 ) 2 2 2 2 3 32 4 232 ] / 2Rt eq / Rt ,
IFF2 2 2 2
Eff [(b 1)( 2 3 ) b 2 2 2 3 3 4 23 ] / Rc eq / R ,
c
IFF3 2
Eff || {[b|| I 235 ( b|| I 235 4 R|| ( 312 21
2 2 2 2 3
) ] /(2 R|| )}0.5
With: I 23 5 2 2 21
2
2 3 312 4 23 31 21 .
These stresses include the nonlinearly loaddependent stresses {}L and the equally nonlinear
residual stresses {}R. Curing of a thick wall generally results in higher residual stresses than thin
walls.Delaminationconditionsarealsotakenintoaccountintheseequations.
DeterminationofthetwoFMCbasedfrictionparametersfromtestresultsisdemonstratedinRef.[11
31]. Bounds for typical epoxybased GFRP, CFRP and AFRP are: 0.1 b|| 0.45 and
1.0 b 1.6 .
Forpredimensioning,avalueof bx 1 0 isadvisable.

2DcaseUDmodefailureconditions:
85
ECSSEHB3220Part2A
20March2011
1
Eff||t
R||t
1
Eff
R||c
2
Eff
Rt
2
Eff
Rc
21
Eff ||
R|| || 2
Thefrictioncoefficient || inthesimplified2DMohrCoulombformulacorrespondsto b|| ,Ref.[11

31].Itcanbeestimatedfromatypicalfracturepointaccordingto:
|| ( R|| 21 ) / 22
fracture fracture

Failuremodeinteraction
Where:Eff=100%=1forpseudoglobalfailure.
The FMC generates a phenomenological 3D lamina stressbased engineering approach for the
derivationoffailureconditions.Itaddressesthemechanics(physics)inthepurefailuremodedomains
and applies simple probabilistic methods in the transition (interaction) zones of the failure modes,
henceagainproducingarelationshipwhichappearsasaglobalfailuresurface.
The probabilistic, theorybased, roundingoff of the failure modes in the transition domains uses a
seriesfailuresystemmodel:
mode 1 m mode 2 m
Eff m
( Eff ) ( Eff ) ..... 1
with the (global) stress effort Eff and the mode material stress efforts Eff mode contributing each
participating failure mode, representing the actual portion of the loadcarrying capacity of the
material,andwiththe(Weibullrelated)interactioncoefficientm.
Inpracticeamaximumof3ofthe5failuremodesinteract.Theinteractionformulais:

( Eff ) m ( Eff ) m ( Eff ) m ( Eff ) m ( Eff ) m 1
m
Eff
showingagainasinglefailureequationwithaninteractionexponent:
2.5<m<3(thelowervaluebeingsafer)
86
ECSSEHB3220Part2A
20March2011
Remarksonapplicability
The application of Cuntze FMC to transverselyisotropic UD material was successful. The
validationofthefailureconditionsbytheWWFEItestdatawassatisfiedafterreevaluationof
theprovided(sometimescontradictory)statictestdatabase.
Ifthezthreaddensityisnottoohighandifthefibresremainstraight,theUDfailureconditions
abovecanalsobeappliedtononcrimpfabricsandaxiallyweftknittedcomposites.
AgoodmappingofCFRP,GFRPandUDtapebasedC/Cmaterials,Ref.[1134],wasachieved
Incontrasttoglobalcriteria,theutilisedinteractionproceduredoesnottakeintoaccountthata
changeinoneindependent(pure)failuremoderegimehasaneffectonanothermode.
The application of Cuntze FMC to orthotropic fabrics is possible, as shown Figure 11.8.2, for
predictedfailurecurvesofCMC(CC/SiC)laminae.
Figure 11.8-2 Cuntze failure criteria Notation and formulae for fabric failure
conditions
11.8.4.5 Other failure criterion

AccordingtotheWWFEWorldWideFailureExercise,Ref.[1117],[1138],criteriathatalsoprovide
goodresultsare:
Zinovievsfailurecriterion.
Bogettisfailurecriterion.
Tsaisfailurecriterion.
[Seealso:64.5forfailurecriteriaforCMCs]
87
ECSSEHB3220Part2A
20March2011
Figure 11.8-3 Comparison between predicted failure curves for C-C/SiC composites
with experimental data
11.9 Comparison between test data and various failure

criteria
11.9.1 Effects on failure mode

Theresultsoftubetestsusingglassfabricreinforcedpolyesterresinsareshownfor,Ref.[117]:
VariationsinthicknessinFigure11.9.1
StaticanddynamicloadinginFigure11.9.2
StaticstrengthinFigure11.9.3
Theseillustratethevariouseffectsonfailuremode,Ref.[117].
Itcanbeshownthatthefailuremodedependsonthe:
Biaxialstressratio,
Fatiguelife,and
Materialcombination.
88
ECSSEHB3220Part2A
20March2011

Thefitoffailuretheoriestotestdatatendstodeterioratewithincreasingfatiguelife,Ref.[117].
Example:Y449/T500glassfabric.
Figure 11.9-1 - Test results: Effect of thickness variation on static failure
89
ECSSEHB3220Part2A
20March2011
TestresultsforbiaxialstrengthY449/T500PfromRef.[117].
Figure 11.9-2 - Test results: Effect of static and dynamic loading
90
ECSSEHB3220Part2A
20March2011
ForY449/T500glassfabric,t=2.0mm,Ref.[117].
Figure 11.9-3 - Comparison: Failure theories applied to static strength result-
91
ECSSEHB3220Part2A
20March2011
11.10 Description of failure modes
11.10.1 Laminates
11.10.1.1 General
Failure modes of fibrereinforced laminates are described in which the individual layers consist of
continuousfibresthatarestrongandstiffalongtheirlength,withinamoreflexiblematrix.
The fibres and matrix are designed to perform complementary functions and therefore react to the
variousloadingconditionsindifferentways.
11.10.1.2 Role of fibres

In each layer of a laminate the principal function of the fibre is to transmit loads in its longitudinal
direction,althoughitalsotransmitstransverseandshearloads.
11.10.1.3 Role of matrix

The resin matrix connects the fibres, maintaining their relative position while distributing the load
amongthefibresandensuringcontinuityoftransverseandshearloadinginthelaminate.
11.10.1.4 Loading
The laminate can be subjected to any combination of loading, though the layers are most efficient
underinplane,longitudinaldirectload.
11.10.1.5 Fibre orientation

The orientation of the fibres normally varies from layer to layer through the laminate thickness.
Consequentlytherearemanymorepotentialfailuremodesfortheselaminatesthanforconventional
metallicmaterials.
11.10.2 Failure
Theprincipalwaysofconsideringthefailureoflaminatesarebasedonthe:
Behaviouroftheindividuallayersandtheinterfacesbetweenthem.
Laminateasawhole.
Table 11.10.1 gives failure modes for UD unidirectional laminates and Table 11.10.2 for MD
multidirectionallaminates.
92
ECSSEHB3220Part2A
20March2011
Table 11.10-1 - Failure modes of unidirectional laminates
Mode of Failure Nature of Loading Primary Factors (see: Key) Secondary Factors
1 Fibre failure transmitted laterally Longitudinal Tension - Fibre tensile strength. - Cure shrinkage stress.
(brittle failure). - Fibre volume fraction.
+ Matrix stiffness and strength.
+ Interface bond strength.
2 Fibre failure transmitted - Fibre tensile strength. + Fibre strength variability.
longitudinally and laterally - Fibre volume fraction.
(brushing). - Matrix stiffness and strength.
- Interface bond strength.
+ Cure shrinkage stresses.
3 Brittle fibre failure (inclined Longitudinal Compression - Fibre strength and volume fraction.
shear). Local and overall stability.
+ Local distortion/eccentricity.
-
4 Kink band failure (timber like). - Matrix shear stiffness. - Fibre shear stiffness.
+ Local distortion/eccentricity. - Fibre volume fraction.
- Fibre diameter. + Moisture and temperature
(matrix).
5 Fibre microbuckling. - Matrix transverse stiffness. + Moisture and temperature
- Fibre diameter and volume fraction. (matrix).
Fibre Young's Modulus.
-
6 Locally originated delamination. + Local distortion/eccentricity. - Matrix shear stiffness.
- Fibre diameter and modulus. + Moisture and temperature
- Matrix tensile strength. (matrix).
(continued...)
93
ECSSEHB3220Part2A
20March2011
Mode of Failure Nature of Loading Primary Factors (see: Key) Secondary Factors
7 Matrix/bondline tension fracture. Transverse Tension - Matrix/fibre average failure strain. + Disparity between fibre and
Fibre/matrix bond strength. matrix stiffnesses.
- Cure shrinkage stresses. Matrix strength.
+ - Fibre distribution
+ irregularity.
Fibre volume fraction.
+
As above. Short Transverse Tension (normal) As above. As above.
(delamination possible) + Curved laminate bonding.
8 Matrix/bondline fracture(inclined Transverse Compression As for Mode 7. As for Mode 7.
shear).
9 Transverse layer buckling. - Matrix/fibre transverse modulus. + Fibre volume fraction.
- Matrix/fibre shear modulus.
10 Shear in matrix, fibre/matrix Longitudinal Short Transverse Shear - Matrix shear strength. + Fibre volume fraction.
debonding, interlaminar shear in - Fibre matrix adhesion. + Cure shrinkage stresses.
laminate. + Moisture and temperature (resin).
11 Shear in matrix, interlaminar shear Longitudinal/Transverse Shear - Matrix shear strength. + Cure shrinkage stresses.
in laminate. Moisture and temperature
+ (matrix).
12 Shear in matrix (cross-fibre shear), Transverse/Short Transverse Shear As for Mode 10. As for Mode 10.
interlaminar shear in laminate.
Key: + The greater the quantity, the more likely the failure.
- The lower the quantity, the more likely the failure.
For exhaustive description, see Ref. [11-6].
94
ECSSEHB3220Part2A
20March2011
Table 11.10-2 - Failure modes of multidirectional laminates
Nature of Loading
Mode of Failure Primary Factors (See: Key) Secondary Factors
on Laminate
Layer transverse tension Tension + Transverse tensile strain components in layer, Some effective transverse and
cracking (regularly spaced Shear crack frequency increases with strain. shear stiffness retained initially
crack through independent Compression Cure shrinkage stresses. via uncracked zones.
layers). +
Layer longitudinal tensile See: Table 11.10.1 - unidirectional longitudinal Laminate lay-up and stacking
fracture. tensile. sequence.
+ Stress concentration from adjacent cracked + Cure shrinkage stresses.
layers.
Delamination from free - Inter-layer shear strength. In tension or shear, delamination
edges or notches. + Moisture and temperature (resin matrix). often local to stress raisers.
Varies with lay-up and stacking sequence. Usually catastrophic in
compression.
Delamination/layer + Local layer distortion. - Matrix shear stiffness.
buckling. + Layer thickness. + Moisture and temperature (resin
- Longitudinal layer modulus. matrix).
- Matrix tensile strength.
Interlaminar shear. Short Transverse See: Table 11.10.1 - Modes 10 and 12. See: Table 11.10.1 - Modes 10
Shear (normal). Will vary with stacking sequence. and 12.
Interlaminar tension. Normal Tension See: Table 11.10.1 - Mode 7. See: Table 11.10.1 - Mode 7.
Will vary with stacking sequence.
Key: + The greater the quantity, the more likely the failure.
- The lower the quantity, the more likely the failure.
For exhaustive description, see Ref. [11-6].
95
ECSSEHB3220Part2A
20March2011
Thelaminatestrengthcanbeinfluencedby:
Crackingofthematrix.
Voids, affecting the stresses in the laminate, can also be present as a result of imperfect
manufacturingprocesses.
Environmentalconditions,wherethetemperatureandmoisturecontentsignificantlyaffectthe
matrixbehaviour,andconsequentlythosefailuremodesthatareheavilymatrixdependent.
ExamplesofsomefailuremodesareshowninTable11.10.3.
Table11.103Examplesofsomefailuremodes
Fibre Microbuckling
Mode 3
Initially similar to the out-of-
phase buckling shown here.
Fibre Microbuckling
Mode 4 The in-phase fibre buckling
shown here often precedes
"Kink-band" failure.
Longitudinal/Short
Mode 10
Transverse Shear Failure
Longitudinal/Transverse
Mode 11
Shear Failure
Transverse/Short
Mode 12 Transverse Shear Failure
96
ECSSEHB3220Part2A
20March2011
11.11 Fatigue strength of composites
11.11.1 Background
11.11.1.1 General
InthefirstinvestigationsofthefatiguebehaviourofCFRPcomposites,emphasiswasplacedonthe
cyclic tension/tension loading of unidirectional laminates. In contrast to what is known about the
behaviourofmetals,nosinglethroughcrackwasfoundtonucleateandgrow.
Fromsuchobservations,ithasoftenbeenconcludedthatcompositesexhibitbetterfatigueproperties
thanmetalsandeventhattheyareinsensitivetofatigue.However,suchaconclusionisnotjustifiedin
thecaseofmultidirectionallaminates.
11.11.1.2 Factors influencing residual strength

Variousparametershaveastronginfluenceontheresidualstrengthofaspecimenwhenitisfatigue
loaded,[See:4.7].Theseincludethe:
Typeofloading(loadratioR).
Laminatelayup.
Environmentalconditions.
Discontinuities,e.g.openholes,delamination,flaws,andcracks.
11.11.1.3 Analysis
Althoughthereareanumberofanalyticaldescriptionsofthefatiguebehaviourofmetallicmaterials,
the literature contains very little information on the subject of composites. Some preliminary
information(estimates)onwaysofinvestigatingthebehaviourofcompositesunderfatigueloadingis
provided.
97
ECSSEHB3220Part2A
20March2011
11.11.2 Analytical notation

D11 flexuralstiffnessofdelaminatedplies
K stressintensityfactor
Kb bucklingcoefficient
N loadcycles
Ncr loadcyclestofailure
Ns loadcyclesrelatedtoaspecificresidualstrength
Nx loadperunitwidth
a lengthofdelamination
acr criticallengthofdelaminationatfailure
as lengthofdelaminationrelatedtospecificresidualstrength
lcr criticaldelaminationlength(buckling)
t thicknessofdelaminatedplies
stressinundisturbedcrosssection
cr staticultimatecompressionstrength
c compressionstrength
cs residualcompressionstrength
min minimumstresslimitoflastcompressionloadcycle
11.11.3 Approximation of fatigue life

Crack propagation (propagation of delaminations) in composite materials induced by cyclic loading
canbedescribedbyfracturemechanicsformulation,Ref.[119],wheretheincreaseincracklength,da,
withincreasingnumberofloadcycles,dN,isexpressedas:
C K c a
da n n m
[11.11-1]
dN
Where the stress intensity factor K a and c, n and m are constants that should be
determinedexperimentally.
IntegratingEquation[11.111],onegetsthenumberofcyclesrelatedtoaspecificdelaminationlength:
1 m 1 m
a0
N a s
1 m c n
s [11.112]
Failure of the specimen occurs when the length of the delamination is equal to the critical buckling
lengtha=lcrandtheremainingcrosssectionofthespecimenisnotabletowithstandtheloads.
Thefurtherassumptionsmadeare:
98
ECSSEHB3220Part2A
20March2011
at: N 0 l cr a0 ; c cr
at: N N cr
l cr acr ; c min
For given loading cycles, Ns, residual compression strength cr = cs is related to the critical
delaminationlengthlcr=as.
Ref.[119]presentsanapproximationforthecriticaldelaminationlengthbasedonthebucklingofthin
plates;seealsoRef.[112]:
Kb
2 D 11
N x 2 [11.113]
a
With: N xt
x b

FromEquation[11.113]:

2 D 11
[11.11-4]
a cr K b
N x
Introducing Equation [11.114] into Equation [11.112] gives a formula representing the relationship
betweentheloadcyclesandtheresidualstrengthofalaminate:
1 m 1 m
K b D11 K b D11
2
2 2
2
t t [11.115]
cs cr
Ns 1 m c n
Thenumberofcyclestofailureisgivenbyasimilarequation:
1 m 1 m
K b D11 K b D11
2
2 2
2
t t [11.116]
min cr
N cr 1 m c n

99
ECSSEHB3220Part2A
20March2011
11.12 References
11.12.1 General
[111] N.F.Dow&WalterB.Rosen
EvaluationofFilamentReinforcedCompositesforAerospace
StructuralApplications
NASACR207,April1965
[112] R.M.Jones
MechanicsofCompositeMaterial
McGrawHill,1975
[113] B.D.Agarwal&L.Y.Broutman
AnalysisandPerformanceofFibreComposite
JohnWileyandSons,1980
[114] L.E.Nielsen
MechanicalPropertiesofPolymersandComposites
Vol.2,MarcelDekker,NewYork,1974
[115] ESDU83014
FailureCriteriaforanIndividualLayerofaFibreReinforced
CompositeLaminateunderInPlaneLoading
[116] ESDU82025
FailureModesofFibreReinforcedLaminates
[117] ASTMSTP787
BiaxialStrengthBehaviourofGlassReinforcedPolyesterResins
TechnomicPublishingCo.,Inc.1980
[119] DFVLRMitteilung8708
SchadensmechanikkohlenstoffaseverstrkterKunstoffebei
Schwingbelastung
[1110] Notstated
[1111] Notstated
[1112] Notstated
[1113] Notstated
[1114] Notstated
[1115] Notstated
100
ECSSEHB3220Part2A
20March2011
[1116] SodenandHinton
CompositeScienceandTechnologyVol.58(1998)
CompositeScienceandTechnologyVol.62(2002).
[1117] M.Hintonetal.:DERA(UK)/UMIST(UK)/QinetiQLtd,(UK)
FailureCriteriainFibrereinforcedPolymerComposites
Elsevier,ISBN:008044475X(2004)
[1118] C.Kaiseretal:HPS/KayserThrede/DLR,(D)/ESTEC(NL)
EuropeanConferenceonSpacecraftStructures,Materialsand
MechanicalTesting,ESTEC(NL).1012May2005.
ESASP581(August2005)CDROM
[1119] B.W.RosenandZ.Hashin.
AnalysisofMaterialProperties
Reinhardt,T.D.etal.(Eds.):EngineeredMaterialsHandbook,Vol.1,
Composites.MetalsPark,Ohio;ASMInternational,1989,pages185
205.
[1120] S.Goutianosetal
Mechanismsofstresstransferandinterfaceintegrityincarbon/epoxy
compositesundercompressionloading,PartI:Experimental
investigation
InternationalJournalofSolidsandStructures39(2002)p.32173231
[1121] K.Rohwer:DLR(D)
DraftfordiscussionFailurecriteria
StructuralMaterialsHandbookWorkingGroup(April2004)
[1122] C.Kaiseretal:HPS/DLR/KayserThrede/ESTEC
FailurecriteriaforFRPandCMC:Theory,Experimentsand
Guidelines
ProceedingsoftheEuropeanConferenceonSpacecraftStructures,
MaterialsandMechanicalTesting,ESTEC,Noordwijk,NL1012May
2005
ESASP581(August2005)CDROM
[1123] S.J.DeTeresa&G.J.Larsen
ReductionintheNumberofIndependentParametersfortheTsaiWu
TensorPolynomialTheoryofStrengthforCompositeMaterials.
JournalofCompositeMaterials,Vol.37,No.19(2003),17691785.
[1124] Krber
BruchhypotheseundReservefaktorenfrunidirektionale
EinzelschichteninFaserverbunden.Handbuchfr
Strukturberechnung,5130101,1982.
[1125] Z.Hashin
FailureCriteriaforUnidirectionalFiberComposites.Journalof
AppliedMechanics,TransactionsoftheASME,Vol.47(1980),329334.
101
ECSSEHB3220Part2A
20March2011
[1126] A.Puck&H.Schrmann:
FailureanalysisofFRPlaminatesbymeansofphysicallybased
phonologicalmodels.CompositesScienceandTechnology58(1998)
10451067andCompositesScienceandTechnology62(2002)1633
1662.
[1127] A.Puck,J.Kopp&M.Knops
GuidelinesforthedeterminationoftheparametersinPucksaction
planestrengthcriterion.CompositesScienceandTechnology62(2002)
371378.
[1128] R.JeltschFricker
BruchbedingungenvomMohrschenTypfrtransversalisotrope
WerkstoffeamBeispielderFaserKunststoffVerbunde.ZAMMVol.
76(1996),No.9,505520.
[1129] R.G.Cuntzeetal.:
NeueBruchkriterienundFestigkeitsnachweisefrunidirektionalen
FaserkunststoffverbunduntermehrachsigerBeanspruchung
ModellbildungundExperimente.FortschrittberichteVDI,Reihe5:
GrundundWerkstoffe,No.506,Dsseldorf:VDIVerlag,1997,p.198
[1130] M.Hintonetal.:DERA(UK)/UMIST(UK)/QinetiQLtd,(UK)
Elsevier,ISBN:008044475X(2004)
[1131] R.G.Cuntze
ThePredictiveCapabilityofFailureModeConceptbasedStrength
ConditionsforLaminatescomposedofUDLaminaeunderstatictri
axialstressstates
WWFEII,PartA,issueinCompositesScienceandTechnology2009
(planned)
[1132] R.G.Cuntze
ThePredictiveCapabilityofFailureModeConceptbasedStrength
CriteriaforMultidirectionalLaminates
WWFEI,PartB,CompositesScienceandTechnology64(2004),487
516
[1133] VDI2014:GermanGuideline,Sheet3DevelopmentofFibre
ReinforcedPlasticComponents,Analysis.BeuthVerlag,2006.(in
GermanandEnglish)
[1134] R.G.Cuntze
StrengthPredictionforMultiaxiallyLoadedCMCMaterials.3rd
EuropeanWorkshoponthermalProtectionSystems
ESAESTEC:Noordwijk,March1998,WPP141
[1135] KlausRohwer:DLR(D)
OverviewofFailureCriteria
ECSSWorkingGroup(2005)
102
ECSSEHB3220Part2A
20March2011
[1136] RalfCuntze:formerlyMANTechnologieAG,Augsburg(D)
CuntzesFMCbasedFailureConditions
ECSSWorkingGroup(2009)
[1137] M.J.Hinton,A.S.Kaddour,P.D.Soden
Afurtherassessmentofthepredictivecapabilitiesofcurrentfailure
theoriesforcompositelaminates:comparisonwithexperimental
evidence.
CompositesScience&Technology64(2004)p549588.
[1138] Dr.ASKaddour,QinetiQ(UK)
Noneofthecriteriaisrightorwrong,butmoreorlesssuitable
DiscussionsbyemailbetweenDr.IngK.Rohwer,DrIngCuntze,
QinetiQ,(August2009)
[1139] A.Puck
FestigkeitsanalysevonFaserMatrixLaminatenModellefrdie
Praxis,Hanser,1996
[1140] R.G.Cuntzeetal
NeueBruchkriterienundFestigkeitsnachweisefrunidirektionalen
FaserkunststoffverbunduntermehrachsigerBeanspruchung
ModellbildungundExperimente
VDIFortschrittbericht,Reihe5,Nr.506(1997),250pages(inGerman)
Newfracturecriteria(HashinPuckactionplanecriteria)andStrength
DesignVerificationsforUnidirectionalFRPssubjectedtoMultiaxial
StatesofStressModeldevelopmentandexperiments
[1141] R.G.Cuntzeetal
103
ECSSEHB3220Part2A
20March2011
12
Calculation of thermal stress and
displacement
12.1 Introduction
12.1.1 General
Aswithmostotherpropertiesofunidirectionalcomposites,thecoefficientofthermalexpansionCTE
inthelongitudinaldirectionisdifferentfromthatinthetransversedirection.
12.1.2 Longitudinal CTE

Thelongitudinalcoefficient1isgenerallysmallbecausethefibresthemselvesusuallyhaveasmaller
coefficientthanthematrixresinandimposeamechanicalrestraintonthematrixmaterial.
12.1.3 Transverse CTE

Thetransversecoefficient2islarger.Atlowfibrevolumefractionitcanbeevengreaterthanthatof
the neat resin, because the matrix is prevented from expanding in the longitudinal direction, so is
forcedtoexpandmoreinthetransversedirection.
104
ECSSEHB3220Part2A
20March2011
12.2 Analytical notation for thermal stress calculations

Formulae:
A In-plane stiffness
B Coupling stiffness
D Bending stiffness
E E-Modulus
G G-Modulus
N In-plane forces
M Bending moments
S Stress - strain matrix
T Temperature
X, Y, Z Laminate co-ordinate system
h Laminate thickness
Coefficient of thermal expansion, CTE
Strain
Curvature
Poisson's ratio
Stress
Vf Fibre volume ratio
Subscripts:
m Matrix
f Fibre
1 Parallel to fibre direction
2 Perpendicular to fibre direction
Superscripts:
M Mechanical
T Thermal
105
ECSSEHB3220Part2A
20March2011
12.3 Calculation of CTE from constituents
12.3.1 CTE in fibre direction

The coefficient of thermal expansion for a UD layer parallel to the fibre direction is given by the
simplemixingrule:
E V
E m 1V
1
f1 f f1 f m
[12.3-1]
E 1
whereE1isgivenby:

E1 E f 1 V f E m (1 V f )
12.3.2 CTE perpendicular to fibre direction
12.3.2.1 Calculation Methods

SomemethodsforthecalculationofthetransverseCTEaregiveninreferences121and122.
Figure12.3.2providesanexampleofCTEforvariousunidirectionalcompositesandtheirdependence
onfibrevolumecontent.
Figure 12.3-1 - CTE of unidirectional composites
TheCTEforT300/9140/45/90laminatesisshowninFigure12.3.3.
106
ECSSEHB3220Part2A
20March2011
Figure 12.3-2 - CTE for T300/914 0/45/90 laminates
12.4 CTE for a laminate

Theconstitutiveequationrelatesthestressresultants(inplaneforcesandbendingmoments)tothein
planestrainsandcurvatures:
N A B
0
M B D
[12.4-1]
0
Theoverallstressresultants{N}and{M}aregivenby:
N N N T
M M M T
[12.4-2]
Wheretheequivalentthermalforcesandmomentscanbewrittenas:
N S dz
h 2
T T
[12.43]
h 2
107
ECSSEHB3220Part2A
20March2011
M S zdz
h 2
T T
[12.44]
h 2
InthisanalysisitisassumedthatHookesLawfororthotropicmaterialappliestoeachlayerandthat
thelayersareinastateofplanestress,whichistrueonlyawayfromtheedgeofthelaminate.
Thestressstrainmatrixisgivenby:

E1 12 E 1
0

1 12 21 1 12 21

S i SYMM. E 2
0
1
[12.45]
12 21

G12

andfromthedefinitionofstressresultantsoneobtainstheinplane,bendingandcouplingstiffnessof
thelaminatedcomposite:
A S dz
h 2
i
[12.46]
h 2
D S z dz
h 2
i
2
[12.47]
h 2
B S zdz
h 2
i
[12.48]
h 2
Where: hdenotesthethicknessofthelaminatedplateand[S]iisthetransformedstressstrainmatrix
inlaminatecoordinatesxy.Thestressesintheithlayerarethengivenby:
i S i 0 Z T
i
[12.49]

108
ECSSEHB3220Part2A
20March2011
Wherethethermalstrainscanbewrittenas:
T
T i i i
[12.410]
Dividing the thermal strains and curvatures by the temperature differences T, one obtains the
coefficientsofthermalexpansionforthelaminate.
1T
c
[12.411]
Where{c}isa6x1vector.
WiththeconstitutiveexpressionsEq[12.41],Eq[12.43],Eq[12.44]andEq[12.410],andassuming
thattheoverallstressresultantsareequaltozero,thevector{c}canbewrittenas:
A B 1 n A B
c

D [12.4-12]
B i 1 B D i 0 i
WhereiaretheCTEsoftheithlayer.
In the case of a symmetrical laminate, the coupling stiffness [B] vanishes and only the first three
coefficientsof{c}arenonzero.Foranunsymmetricallaminate,thermallyinducedcurvaturesoccur
andthevector{c}isfullypopulated.
12.5 Thermal stresses within laminate layers
12.5.1 General
The behaviour of the resin matrix, determines the extent to which the mechanical and thermal
properties of CFRP are temperature dependent. Consequently, the Youngs moduli parallel and
perpendiculartothefibredirection,theshearmodulus,Poissonsratioandthecoefficientsofthermal
expansionofanunidirectionallaminatearefunctionsoftemperatureinanunrestrictedmanner.Asa
resultthetotalstiffnessmatrix,thethermalstressresultantsandthevectorofcoefficientsofthermal
expansionforamultidirectionallaminatearelikewisetemperaturedependent.
12.5.2 Residual curing stresses

The anisotropic thermal expansion characteristic of the fibres, together with that of the resin matrix
whichisisotropic,leadtoresidualstressesinthelaminateaftercuring.
109
ECSSEHB3220Part2A
20March2011
12.6 Stress strain temperature relation
12.6.1 General
Theresultingstrainsinanunidirectionallayerofalaminate,whichissubjectedtoatemperatureload
andstresses,areobtainedfromthesumofmechanicalandthermalstrains.
12.6.2 Mechanical strains

The mechanical strains depend on the temperature, because the compliance is a function of
temperature,whilethethermalstrainsareindependentofthestresses.Hence,evenatthesamestress
levelthemechanicalstrainsareinfluencedbythetemperature.
, T T
M T
i i i i
[12.61]
Whenthemechanicalstrainsarelinearlydependentonthestressesi,asimplifiedformofEq[12.61]
isobtained:
S ij T i
M
i
[12.62]
Where:Sij(T)denotesthetemperaturedependentcompliance.
FromEq[12.61]onecannowcalculatethechangeofstrainduetoaninfinitesimalchangeofstress
andtemperature:

M T
di S ij T d
i

T
i

T
i dT

[12.63]

Where:iM/Tisthecouplingcoefficient.
12.6.3 Incremental strain theory

Inincrementalstraintheory,wherethethermalstrainandstressstrainrelationsareonlyofinterestat
thefinaltemperature,thecouplingterm,iM/T,isomitted.Thisiscorrectonlywhenthecompliance
istemperatureindependent.Insomecasesthisomissioncanresultinasmuchas35%differencein
thecuringstressprediction,Ref.[121].
Thestepwiseevaluationofthermalstrainsandstressesisperformedassumingthat:
Theoverallstressresultantsareequaltozero.
Temperaturedependent material properties are given in a table with constant temperature
differencestepsT.
Theuppertemperatureboundisthecuringtemperature,wherethelaminateisinastressfree
state.
110
ECSSEHB3220Part2A
20March2011
ThethermalstrainsinalayerjatpresenttemperatureTNaregivenby:
T 12 T T T
T
N j i j i 1 j
[12.6-4]
AndthetotalthermalstrainsinthelaminateatTNaregivenby:

0 AT N
TN
BT N T 1 T
DT M T
N
T N B T N
T
N
[12.6-5]
N N
Withthethermalstressresultants:
N T S T T dz
h 2
T T
N N N [12.6-6]
h 2
And:
M T S T T zdz
h 2
T T
N N N [12.67]
h 2
Where: hdenotesthelaminatethickness.
From the total thermal strains, the coefficients of thermal expansion for the laminate at (TN + TN1)/2
are:
0 0
T N T N 1

T N T N 1 [12.68]
c
T
andfinallythesinglelayerthermalstressesaregiveninlayerjby:
j S T N j 0 T N z j T N T T N j [12.69]
111
ECSSEHB3220Part2A
20March2011
12.7 Microstress analysis
12.7.1 General
Microstressesariseduetothedifferencebetween:
curingtemperatureandthetemperatureatuse,and
thecoefficientsofthermalexpansionoffibresandmatrix.
Therearestressesbothalongandperpendiculartothefibredirection.
12.7.2 Microstresses on fibre axis

Alongthefibreaxisthehighercoefficientofthermalexpansionoftheresinproducesatensilestressin
theresinwhenthecompositeiscooledfromcuringtoworkingtemperature.
Thetensilestressintheresincanbeevaluatedby:
( f )TE

m x
m
m
1 E
[12.71]
m
Ef x
wheretheassumptionisthatthereisaperfectbondattheresintofibreinterface,Ref.[124].
12.7.3 Microstresses normal to fibre axis

The microstresses perpendicular to the fibre direction are of more interest, because this is the more
criticaldirectionofasinglelayerwithrespecttothestrengthandstiffnessproperties.
To evaluate the internal stresses in fibre and matrix perpendicular to the fibres a finite element
programshouldbeused,Ref.[121].
12.8 References
12.8.1 General
[121] ESTECContractNo.2930/76/NL/PP(SC)
Finalreportforastudyonevaluationofinbuiltstressesduring
manufactureofelementsmadeofcarbonfibrereinforcedplastics
112
ECSSEHB3220Part2A
20March2011
[122] R.A.Schapery
Thermalexpansioncoefficientsofcompositematerialsbasedon
energyprinciples
J.CompositeMaterial2(3),1968
Technomic,1980
[124] R.C.Novak&M.A.deCrescente
Fabricationstressesingraphite/resincomposites
ASMEPaperNo.70GT84,1970
113
ECSSEHB3220Part2A
20March2011
13
Moisture effects on composite properties
13.1 Introduction
13.1.1 General
The properties of composite materials are susceptible to changes induced by the absorption of
moisture.Sucheffectsshouldbeconsideredbecausemoistureislikelytobepresentintheatmosphere
during the production of components and launch of a spacecraft. The subsequent desorption
occurringintheoperationalspacevacuumalsoneedconsideration.
13.1.2 Moisture penetration

Moisture penetrates the organic resin material (and also aramid fibres) by a diffusioncontrolled
process. Additional heat can increase the moistureretention capacity of the material. Such moisture
acts like a plasticiser, which softens the matrix and lowers the glass transition temperature. Such
processesare often reversible. The process can, however,lead tomodificationof residualstrain and
microcracks within the resin system. Thermal cycling can cause some permanent changes in the
coefficient of thermal expansion and possibly also in mechanical properties. Such effects should be
avoidedinstructuresthathavetobeverystable,eitherdimensionallyoroptically,[See:Chapter28].
13.1.3 Moisture effects
13.1.3.1 Swelling
Animportanteffectoftheabsorptionofmoisturebycompositesisswellingoftheresinmatrix.This
can lead to a modification of the preexisting residualstress conditions, which can result in the
production of microcracks both in the matrix and at the interface with the fibres. In certain cases,
swellingduetomoistureabsorptioncancauseareductioninthethermallyinducedstresses.
114
ECSSEHB3220Part2A
20March2011
13.1.3.2 Mechanical properties
The mechanical properties of the composites that are altered by moisture absorption are those
governedbythebehaviourofthematrix,[See:13.3].
Thepropertiesaffectedincludethe:
Interlaminarshearstrength(ILSS),
Inplaneshearstrength,
Compressionstrength,and
Coefficientofthermalexpansion(CTE).
These alterations become more evident at high temperatures, owing to the depression of the glass
transitionpoint,Tg.
13.1.3.3 Prediction of moisture effects

Analyticalmethodshavebeendevelopedtopredictthechangesinmaterialpropertiesboughtabout
bythepresenceofmoisture.Methodsarepresentedfordeterminationofthe:
Moisturedistribution,[See:13.5].
Finalmoisturecontent,[See:13.5].
Swellingcoefficientofunidirectionalplies,[See:13.6].
Swellingcoefficientforthelaminate,[See:13.6].
13.2 Analytical notation for moisture effects
a, b constants
d thickness of specimen
D state of diffusion
D1 diffusion coefficient parallel to the fibres
D2 diffusion coefficient transverse to the fibres
Dm diffusion coefficient of the matrix
Dz diffusion coefficient normal to the surface
E Young's modulus
E2 transverse modulus of elasticity (Young's modulus transverse to
unidirectional plies, here in tensile)
F force
G time dependent parameter, [See: Equation [13.5-4].
H thickness
L swelling strain transverse to unidirectional plies
M moisture content (percentage weight gain)
Mi initial moisture content
M maximum moisture content
N number of cycles
R 22
tu
2
transverse tensile strength of unidirectional plies
S equivalent thickness
115
ECSSEHB3220Part2A
20March2011
T temperature
T time
Vf volume fraction of fibres
Vr volume fraction of resin
angle between surface plane and the normal x-axis ( = 90 fibres are
in a plane parallel to the surface)
swelling coefficient
tensile strain
m swelling strain
relative humidity
Poisson's ratio
fatigue strength of unidirectional plies
t tensile strength of neat resin
r weight fraction of resin
Subscripts:
c composite
f fibre
m matrix
1 fibre direction
2 transverse to fibre direction
i i-th layer of a laminate
13.3 Typical effects of moisture
13.3.1 General
Ingeneral,moistureabsorptionbythematrixofacomposite:
changesthechemicalstructure(softensmatrix)
causesswellingofthematrix
13.3.2 Sample data: Effects of moisture
13.3.2.1 General
Therateofdiffusionofmoisturedeterminestheextenttowhichmoistureinfluencesthesofteningor
swellingeffects.
Sampledataarepresentedfor:
MatrixplasticisinginFigure13.3.1,Ref.[134].
MatrixswellinginFigure13.3.2
TemperaturetimedependenceofmoistureabsorptioninFigure13.3.3andFigure13.3.4.
116
ECSSEHB3220Part2A
20March2011
Theeffectsofmoistureonmechanicalpropertiesareshownfor:
ChangesinstrengthinFigure13.3.5,Ref.[134].
ChangesinmoduliinFigure13.3.6,Ref.[134].
ChangesinfatiguebehaviourinFigure13.3.7,Ref.[134].
[Seealso:4.5forfurtherdataonhygrothermaleffectsonmechanicalandphysicalproperties]
13.3.2.2 Matrix plasticising
Figure13.31Stress/straincurvesofFibredux914(neatresin)fordifferent
moisturecontentstestedatRT
117
ECSSEHB3220Part2A
20March2011
13.3.2.3 Matrix swelling
Figure 13.3-2 - Swelling strain, transverse (UD laminate)
13.3.2.4 Temperature-time dependence of moisture absorption
Figure 13.3-3 - Influence of temperature and time on weight gain
118
ECSSEHB3220Part2A
20March2011
Figure 13.3-4 - Test results: Moisture absorption versus time
13.3.2.5 Changes in strength
Figure 13.3-5 - Transverse tensile strength
119
ECSSEHB3220Part2A
20March2011
13.3.2.6 Changes in moduli
Figure 13.3-6 - Transverse modulus of elasticity
13.3.2.7 Changes in fatigue behaviour
Figure 13.3-7 - Influence of temperature and moisture on fatigue strength of

unidirectional 0 laminates
120
ECSSEHB3220Part2A
20March2011
13.4 Approximate method for calculation of strength and

modulus retention of [0/90] laminates
13.4.1 General
Theabsorbedwaterinepoxycompositesnotonlyproduceschangesofchemicalandphysicalnature
butalsocausesadegradationofthemechanicalproperties.
The reduction of strength and stiffness is the result of softening of the fibre, the matrix and their
interface.
The level of degradation depends on the moisture content in the composite; it varies with the
structuralcomponentsandthequalityofcompositeandcannotbeignored.
[See:13.2foranalyticalnotation]
13.4.2 Modulus retention

Becausethedegradationinresinpropertiesresultsinaparalleldegradationincompositeproperties,
thesimpleruleofmixturescanbeusedasabasisforcalculatingthemoduliofUDcomposites,[See:
10.4].
E V E V E
c f f m m [13.4-1]
Where the subscripts c, f and m refer to the composite, fibreandresin respectively; Vis the volume
fraction,andEistheelasticmodulus.
Ec decreases when Em is reduced by moisture, which affects the interfacial bonding strength of the
compositeandcausesalossofefficiencyofstresstransferbetweenfibres.
AccordingtoRef.[133],Equation[13.41]ismodifiedto:
E V E V E E V W
c f f m m f f c [13.4-2]
Where:
isanempiricalconstantforafixedtypeofreinforcementandcouplingagent.
Ec,Em are the moduli of composite and resin matrix after moisture
absorption,Emcanbeobtainedfromtestresults.
Wc isthemoisturecontentbyweightofthecomposite.
Fora[0/90]laminate,Eq[13.42]canbeapproximatedas:
121
ECSSEHB3220Part2A
20March2011
V
E E V E E V W [13.4-3]
f
c
2 f m m f f c
This equation correlates very closely with the experimental results for both glass and carbon
composites;when=2forglassand=3forcarbon.
Figure13.4.1showsacomparisonofcalculatedandmeasuredresultsforGFRPandFigure13.4.2for
CFRP,Ref.[133].
Material:Araldite507/HY956,glassASAH,MS252fabricandwovenrovings
Figure 13.4-1 - Moisture: Comparison of calculated and experimental results of strength

and stiffness of [0/90] GFRP laminates
122
ECSSEHB3220Part2A
20March2011
Material:Araldite507/HY956,carbonCarborundumGSGC2
Figure 13.4-2 - Moisture: Comparison of calculated and experimental results of strength

and stiffness [0/90] CFRP laminates
13.4.3 Strength retention

According to the general engineering equation for strength of UD composites, [See: 11.2], including
theeffectsofmoistureoncompositematerialanditscomponents,anapproximatecalculationcanbe
usedtodeterminethestrengthof[0/90]compositelaminateafterabsorptionofmoisture:

u u V f
Rc R f 2 V m E m

R
u
f VW f c
[13.4-4]
E f
Where:
isanempiricalconstantforafixedtypeofreinforcementandcouplingagent.
Rcu strengthofcomposite.
Ruf strengthoffibre.
[Seealso:Figure13.4.1forGFRP;Figure13.4.2forCFRP,Ref.[133]]
123
ECSSEHB3220Part2A
20March2011
13.5 Moisture content
13.5.1 Fick's law

ThebasisforthecalculationofthediffusionofmoistureintocompositematerialsisFicksLaw:
c c
Dx [13.5-1]
t x x
Thisisvalidforaninfiniteplatewherethemoisturecontentvariesonlyinthexdirection,asdefined
inFigure13.5.1.
Figure13.51Notation:Definitionofaxes
AdditionalassumptionsforthedefinitionofaFickianprocessare,Ref.[135]:
Theenvironmentalconditions(temperature,moisture)onbothsidesoftheplateareconstant.
The temperature inside the material approaches equilibrium much faster than the moisture
concentration,i.e.thetemperatureinsideisequaltotheenvironmentalconditions.
Thetemperaturedistributioninsidethematerialishomogeneous.
The diffusion coefficient depends only on temperature and is independent of moisture
concentrationorofthestresslevelsinsidethematerial.
124
ECSSEHB3220Part2A
20March2011
13.5.2 Determination of moisture content

Dependingontheenvironmentalconditionsandtheconditionofthematerialitself,moistureiseither
absorbedordriedout.Thisweightgainorweightlossisdescribedasthepercentagemoisturecontent
Mofthematerialasafunctionoftimet,Ref.[135],definedas:
weight wet material weight dry material

M M (t ) 100 [13.5-2]
weight dry material
If the process is assumed to be Fickian, [See: Ficks Law], the percentage moisture content during
absorptionordesorptionis:
M GM m
M i M i [13.5-3]
ThetimedependentparameterGisgiveninRef.[135]as:
G 1
8

exp 2 j 1
2
Dt
2
x
s2 [13.54]

2
j 0 2 j 1 2
AccordingtoRef.[135],Eq[13.54]canbeapproximatedbytheexpression,[See:Figure13.5.2]:
G 1 exp 7.3 D x t s 2 0.75

[13.5-5]
NOTE Eq[13.53]isonlyvalidfortimeperiodsgreaterthanoneday.
If the material is exposed only on one side, i.e. the material is insulated on one side; the equivalent
thicknesssisequaltotwicethethicknessh(s=2h),otherwises=h.
ThediffusioncoefficientDx(normaltothesurface)canbeexpressedforaunidirectionallaminateas:
D D cos
x 1
2
D2 sin 2 [13.56]
Generally, the diffusion of moisture in the fibres is small compared with that in the matrix (Dr).
Consequently, the diffusion parallel to the fibres, D1, and that transverse to the fibres, D2, can be
writtenas:
D 1 V D
1 f m [13.5-7]
D 2

1 2 V f
D m [13.5-8]
125
ECSSEHB3220Part2A
20March2011
SubstitutingEq[13.57]andEq[13.58]inEq[13.56]gives:
D D
x m
1 V cos 1 2 V sin
f
2
f
2
[13.59]
13.5.3 Maximum moisture content

Ithasbeenobservedfromexperimentthatthemaximummoisturecontentwithinthelaminateisnot
sensitive to temperature,but doesdepend on the moisture content of the environment towhich the
materialhasbeenexposed.
AccordingtoRef.[135],themaximummoisturecontentcanberelatedtotherelativehumidity by
theexpression:
M a
b
[13.510]
Where: aandbarematerialconstants.
Table13.5.1givesaandbvaluesforvariousmaterials,Ref.[135].
Table13.51Summaryofconstantsforthedeterminationofthemaximum
moisturecontent, M
M a %
b
Material a b
Narmco T300/5208(70vol.%) 1.5 1
Fiberite T300/1034(65vol.%) 1.7 1
Hercules AS/35015(65vol.%) 1.8 1

ResinNarmco 5208 6.7 1.32
ResinFiberite 934 6.3 1.77
ResinHercules 35011 6.3 1.46
Therearevariationsinthereportedvaluesofthesematerialconstants,probablyowingtodifferences
inthecuringprocesses.Thevalueofbcanbeapproximatedbyunity.
Themaximummoisturecontentofthecompositecanalsobeestimatedfromthemaximummoisture
contentoftheneatresin, M.Thebasisforthisistheassumptionthatthefibresdonotabsorbany
m
moisture.Theassumptionisonlyvalidforcarbonandglassfibres,aramidfibresdoabsorbmoisture.
126
ECSSEHB3220Part2A
20March2011
M M m [13.5-11]
m
Where:mistheweightfraction(percent)ofresininthecomposite.
13.5.4 Experimental determination of the diffusion coefficient

From Ref. [135], the material constants a and b in Eq [13.510] and the diffusion coefficient of a
compositeshouldbedeterminedexperimentally.Ifthediffusioncoefficientoftheresinisknown,Dx
canalsobecalculatedusingEq[13.59]providedthat:
Thespecimeniscompletelydriedinadesiccatoranditsdryweightismeasured.
The specimen is placed in a constant temperature, constant moisture environment and its
weightisrecordedasafunctionoftime.
Themoisturecontent(percentweightgain)isplottedagainstthesquarerootoft,asshownin
Figure 13.5.2, to determine the diffusion coefficient and against the relative humidity , as
showninFigure13.5.3,todeterminethematerialconstantsaandb.
Figure 13.5-2 - Illustration of change of moisture content with square root of time
127
ECSSEHB3220Part2A
20March2011
Figure 13.5-3 - Variation of maximum moisture content with relative humidity for a
T300/1034 composite
Thediffusioncoefficient,D,isobtainedfromFigure13.5.2fromtheinitialslopet<tLofthe M versus
tcurve.
2 2
h M 2 M 1
D
4M
[13.512]
t 2 t1
If all the fibres are parallel to the surface through which the moisture diffuses, the edges can be
neglected:
D D
x [13.513]
Ifedgeeffectscannotbeneglected,themethodoutlinedinRef.[135]shouldbeused.
128
ECSSEHB3220Part2A
20March2011
13.6 Calculation of swelling coefficient from constituents
13.6.1 General
Likethecoefficientofthermalexpansion,theswellingcoefficient,,foraunidirectionalcompositeis
notthesameinthelongitudinalandtransversedirections.Thelongitudinalcoefficient1isgenerally
small because fibres themselves usually do not absorb moisture. Aramid fibres are the exception
becausetheydoabsorbmoisture.
Asswellingispredominantlycausedbythematrix,thetransversecoefficientofswelling2isofmore
interest.
Thetwocoefficientscanbecalculatedinthesamewayasthecoefficientofthermalexpansion,ascan
thecoefficientsforthelaminateasawhole.
Amethodforcalculatingtheunidirectionalswellingcoefficientsfromtheconstituentsissummarised
fromthatprovidedinRef.[136].
13.6.2 Swelling coefficient 1 in fibre direction
f 1 1V E E
f
m f1
V m

1

1V
f
[13.61]
1
f
V Em E f 1
f
Theswellingcoefficient1canbeapproximatedby:
1 f1
[13.62]
butisnormallyneglected.
13.6.3 Swelling coefficient 2 transverse to fibres
V 1V 1V E E
f f 12 m f1 m
1V f
2 f1 f m f m f1
[13.63]
1
V Em E f1
f
Eq[13.63]canbeapproximatedby:
V
2 f1 f
m
1 1V
m f1 m f [13.64]
129
ECSSEHB3220Part2A
20March2011
ThevariationoftheswellingcoefficientswithfibrevolumefractionisshowninFigure13.6.1.
Figure 13.6-1 - Moisture: Coefficient of swelling for unidirectional composites
13.6.4 Swelling coefficient for a laminate

A change in the moisture content of a composite produces a change in its dimensions which is
proportional to the change in moisture content and its initial dimensions. Thus, swelling strains
developinthecompositesasaresultofchangesinmoisturecontent.
Theswellingstrainmisgivenby:
M
m i i i
[13.6-5]
DividingtheswellingstrainsbythechangesinmoisturecontentM,givesthecoefficientofswelling,
[See:Figure13.6.2]:
1
[13.66]
i
M i
M i
Where:idenotesthecoefficientsoftheithlayerofacomposite.
130
ECSSEHB3220Part2A
20March2011
Figure 13.6-2 - Moisture: Definition of swelling coefficient 2 for orthotropic material

(unidirectional laminate)
Calculationoftheswellingcoefficientofalaminateisanalogoustothatofthecoefficientofthermal
expansion,[See:12.4]andmakesuseoftheexpression:
B 1 n A B
A

B D
i 1 B D i 0
[13.67]
i
A c dz
h 2
With: i
h s
B c zdz
h 2
i
h s
D c z 2dz
h 2
i
h s
Where:
histhethicknessofthecomposite,
[c]denotesthetransformedstiffnessmatrixinthelaminatecoordinatesxy.
Figure 13.6.3 is an example of the calculated swelling coefficient of [0/45/90]s laminates for
T300/914composite.
131
ECSSEHB3220Part2A
20March2011
Figure 13.6-3 - Moisture: Swelling coefficients for T300/914 [0/ 45/90]s laminates
13.7 Coefficient of moisture expansion (CME)
13.7.1 Resin behaviour
13.7.1.1 General
Theexpansioncharacteristicsofacompositeareofinterestfordimensionallystablestructures,[See:
Chapter28].
[Seealso:2.4formatrixsystems;4.5forhygrothermaleffects]
13.7.1.2 In-plane
Thecoefficientofmoistureexpansionisprincipallyafibredominatedcharacteristic,[Seealso:5.3].
13.7.1.3 Through-thickness (transverse)

Expansion is both positive and large, reflecting the contribution of both resin and fibre. In this
directionanyswellingoftheresinduetomoistureabsorptionisdirectlyconferredonthecomposite.
Thecoefficientofmoistureexpansion(CME)oftheresinisthereforeanimportantparameter.
132
ECSSEHB3220Part2A
20March2011
13.7.1.4 Measurement of CME
Forthebaseresin,CMEismeasuredasanincreaseinspecimendimension(strain)accompanyinga
weightgainbymoistureabsorption,oftenby:
s C [13.7-1]
Where:
s=strainbyswelling
=CME
C=moisturecontent

Thisvaluedescribestheresinbehaviourinthesaturatedstateforaparticularrelativehumidityand
temperature.Ideally bothmoistureuptakeand change instrain are small. The relationshipbetween
the two parameters need not be linear with respect to changes in relative humidity (RH) or
temperature.
Themoistureabsorbencyofresinsdifferenormouslyandcanspanarangeupto6%byweightinthe
fullysaturatedconditionatelevatedtemperaturesandhighRH.
Movingfromthefirstgenerationepoxiestotoughenedepoxies,cyanateestersandthermoplastics,the
level of moisture absorption and hence expansion has reduced significantly. This is shown in Table
13.7.1forUHMcompositesmadewithFiberite9543cyanateesterresin.Theinformationisadvisory
onlyasthematerialswerenotnecessarilytestedunderthesameconditions.Thelevelofmoisturetake
upisdependentonthefibretype,layup,curetemperature,temperatureandRH.
[Seealso:Table6.35.1fordataonRS3cyanateesterresinalone]
Table13.71Moistureabsorption:CyanateesterFiberite954composites
Fiberite 954
Cyanate Resin Systems
954-3 954-2 934
Hygrostrain (ppm) 18.9 29.2 108.5
Water Absorption (%) 0.18 0.17 0.70
CME (ppm/%) 105 172 155
Lay-up (0,+60,-60)xs (0,30,60,90,120,150)s
Key: : 55% RH/EQ
: 50% RH/EQ
: CME: Coefficient of moisture expansion = hygrostrain/%M
Pseudo-isotropic P75 laminates; 30% RC
133
ECSSEHB3220Part2A
20March2011
13.7.2 Composite behaviour
13.7.2.1 General
Thevolumefractionofresinpresentincompositesistypically30%to35%withtheremainderasnon
moisture absorbing fibre. The exception is aramid fibres which can absorb large quantities of
moisture.
TheamountofmoistureabsorbedbyUHMCFRPcompositescanbeaslowas0.2%withthenewer
resins.
13.7.2.2 CFRP composites

ThemagnitudeofmoistureexpansioninCFRPcausedbymoistureuptakebytheresinisdependent
onthe:
Totalamountofmoistureabsorbed(underspecificsaturationconditions)
Resinexpansionduetomoisture(CME),
Carbonfibremodulus,
Fibrevolumefraction.
Thisindicatesthatcomparisonsbetweendifferentcompositescanbedifficultifneithertheresinnor
thefibreiscommon.
Toprovideameansofcomparisonatheoreticalcalculationcanbeundertakenwhichnormalisesona
specific moisture weight gain and enables the difference in composite modulus to be taken into
account. The value of such calculations is debatable unless the resins are radically different in their
moistureabsorptionbehaviours.
13.8 References
13.8.1 General
[131] C.H.Shen&G.S.Springer
Moistureabsorptionanddesorptionofcompositematerials
JournalofCompositeMaterials,10(1),1976
[132] P.E.McMahon&L.Ying
Effectsoffiber/matrixinteractionsonthepropertiesof
graphite/epoxycomposites
NASACR3607,1982
[133] D.P.Bashford:FulmerResearchLtd.,UK
Reviewofmoistureabsorption/desorptionanditseffectsonresidual
stressesincarbonfibrereinforcedplasticsforspaceapplications
[134] DFVLR
Developmentoffracturemechanicsmapsforcompositematerials
ESACR(P)2017,January1985
134
ECSSEHB3220Part2A
20March2011
[135] G.S.Springer
EnvironmentalEffectsonCompositeMaterials
TechnomicPublishingCompanyInc.1981
[136] G.Niederstadtetal
LeichtbaumitkohlenstoffaserverstrktenKunststoffen
Kontakt&Studium,Bd.167
ExpertVerlag,1985
135
ECSSEHB3220Part2A
20March2011
14
Stress concentrations and fracture
14.1 Introduction
14.1.1 General
Understeadilyincreasingtestloads,failureofnotchedCFRPspecimensisinitiatedinindividualplies,
longbeforethemaximumloadcapabilityisreached.Thisoccursalongortransversetothefibresand
alsobetweencertainplies.Thepartialdamageinorbetweentheplies,leadstoaredistributionofthe
localstressesresultinginacomplexfailure.
14.1.2 Fracture mechanics models

Basedupontheexperiencegainedwithbrittlemetallicmaterials,thelinearelasticfracturemechanics,
LEFM,approachwasappliedtodescribethebrittlefracturebehaviourofCFRPmaterials.
ResearchdirectedtowardstheapplicationofLEFMtocompositesindicatesthatitcanonlybeapplied
inalimitednumberofcasesbecausetheapproachbasedonmetallicmaterialsisphysicallywrongfor
composites.
Consequently,severalmodifiedfracturemodelshavebeenproposedtoovercometheshortcomingsof
LEFMforcompositematerials,[See:14.3].Primarily,thesecomparethecriticalstressintensityfactor
withthatofmetals.
Suchfracturemodelsalsoseektopredictthenotchstrengthofcomposites,butithastobeemphasised
thatthesemiempiricalfracturemodelspresentedheredonotdirectlyaddressthemicroandmacro
failureassociatedwiththecrackextensionprocessatthecracktipdamagezone.
The failure process, type of damage and its progression (parameters associated with the different
fracturemodels)stronglydependuponavarietyofintrinsicandextrinsicvariables,assummarisedin
Table14.1.1,Ref.[141].
Table14.11Intrinsicandextrinsicvariablesrelevantforstressintensityfactors
IntrinsicVariables ExtrinsicVariables
LaminateConfiguration LoadingFunction
StackingSequence LoadingRate
PropertiesofConstituents SpecimenGeometry
FibreVolumeFraction ShapeofDiscontinuity
FibreMatrixInterface TestTemperature
FabricationProcedure MoistureControl
136
ECSSEHB3220Part2A
20March2011
14.2 Analytical notation for stress concentrations
Formulae:
Ex,Ey Youngsmodulusofthelaminate
Gxy shearmodulusofthelaminate
Kt concentrationfactor(tensile)
Ks concentrationfactor(shear)
xy Poissonsratio
tensilestress
shearstress
Subscripts:
pointonthecircumferenceoftheholeforwhichthetangentialstressesare
A
ofinterest
x,y orthotropicaxesofthelaminate
angleinrelationtotheorthotropicaxis
circumferentialangleatthehole
undisturbedarea(infinity)
14.3 Summary of fracture models

Table14.3.1summarisessomeofthefracturemodelsappliedtocomposites,withcommentsontheir
scopeandauthors,Ref.[141].
Table14.31Somefracturemodelsandtheirauthors
Model Application
FractureCriterion Authors [Reference]
Abbr. Hole Crack
LinearElasticFracture Waddoups,M.E.
WEK Mechanics(LEFM) Eisenmann,J.R. Ref.[143]
[See:14.5] Kaminski,B.E.
PointStress
Whitney,J.M.Ref.[144]
WN AverageStress
Nuismer,R.Y. ,[145]
[See:14.6]
K PointStress Karlak,R.F. Ref.[146]
Pipes,R.B.Ref.[147],
PWG PointStress Wetherhold,R.C. [148],[149]
Gillespie,J.W.
Mar,J.W.
ML N0.5
Lin,K.Y. Ref.[1410],[1411]
Poe,C.C.
PS Strain
Sova,J.A. Ref.[1412]
137
ECSSEHB3220Part2A
20March2011
14.4 Evaluation of fracture models

Based on the research of Awerbuch, Ref. [141], some general comments on fracture models for
compositesincludes:
Goodagreementbetweenallfracturemodelsandallexperimentalresults,[See:Table14.3.1].
All fracture models are semi empirical, i.e. they can be applied providing that at least two or
three(dependingonthefracturemodel)notchedstrengthdataareknown.
The fracture models include certain parameters assumed to be material constants. i.e.
independentofspecimengeometry:
Theseparametersstronglydependonlaminateconfigurationandmaterialsystemaswell
asonthevarietyofintrinsicandextrinsicvariables,[See:Table14.1.1].
These parameters should be determined experimentally for each laminate layup and
materialsystemindependently
Thenumberofteststobeconductedindeterminingthefracturemodelsparametersdepends
onthemodelitselfandthelevelofaccuracyrequired.
Inallfracturemodelstheactualpatternanddetailsofthenotchtipdamagearebypassedby
assuminganeffectivenotchtipdamagezonewhichgrowsinaselfsimilarmanner.
Most CFRP laminates are highly notch sensitive. For many cases, the notched strength is
reduced by as much as 50% for notch length to width ratios of 0.2 to 0.3, as shown in Figure
14.4.1.
138
ECSSEHB3220Part2A
20March2011
Figure 14.4-1 - Strength ratio versus hole diameter: Crack length for WN fracture
model - 'Average stress criterion'
14.5 WEK fracture model
14.5.1 General
The basic assumptions for the Waddoups, Eisenmann and Kaminski (WEK) model, Ref. [143], are
based upon the LEFM linear elastic fracture mechanics approach, which can be directly applied to
anisotropicmaterialsgiventhattheconditionsare,Ref.[142]:
Theorientationoftheflaw(notch)withrespecttothemainaxisofsymmetryisfixed.
The stressintensityfactordefined for anisotropic cases is consistent with the isotropic case in
stressdistributionandincrackdisplacementmodels.
Thecriticalorientationcoincideswithoneofthemaindirectionsofelasticsymmetry.
Dependent upon material parameters, an equivalent damage tip zone is defined. The total defect is
consideredthesuperpositionofthedefectinitiallypresent(beforeloading)andthedamagetipzone.
Thestressintensityfactorandtheequivalentdamagetipzoneareassumedtobematerialconstants.
14.5.2 Circular holes

Bowie,Ref.[1413],providesastressintensityfactorforsymmetricalcracksemanatingfromacircular
holeinaninfiniteisotropicplate,asdefinedinFigure14.5.1,withtheresult:
139
ECSSEHB3220Part2A
20March2011
K a f a R

I N
[14.51]
Figure 14.5-1 - Notation: Infinite isotropic plate with circular hole

Failureoccurswhentheappliedstressreaches
N :

K IC
[14.52]
N
a f a R
Inthecaseofaninfiniteplatewithoutahole:
R0

a R
And,fromRef.[141]:
f a R 1
140
ECSSEHB3220Part2A
20March2011
Theunnotchedstrength0isdefinedas:
K IC
[14.5-3]
0
a
Thusthenotchsensitivitybecomes:
1

[14.5-4]
N 0
f a R
Itisassumedthataisindependentoftheholeradius.Thus,Eq[14.54]predictsthenotchstrengthto
beafunctionofabsoluteholesize.
Anapproximationforthefunctionf(a/R)isgivenby:
f a R C 1
C3 [14.5-5]
C2 a R
With: C=0.6866
C=0.2772
C=0.9439
The equivalent damage zone size, a, can now be adapted such that the equations for notched and
unnotchedconditionsarefulfilledbest.UsingEquation[14.55]avalueacisgivenas:
C1
a R
C 2 [14.5-6]
0 N C 3
c

The notch sensitivity
N 0 represented by this single parameter is shown for a [45/0]s CFRP
laminateinFigure14.5.2,Ref.[141].
141
ECSSEHB3220Part2A
20March2011
Figure 14.5-2 - Comparison between experiment and prediction by WEK model
14.5.3 Straight crack

Figure14.5.3showsthenotationforastraightcrack.
142
ECSSEHB3220Part2A
20March2011
Figure 14.5-3 - Notation: Infinite plate with straight crack
Inasimilarwaytotheplasticzonecorrection,Ref.[1414],thestressintensityisdefinedas:
N c ac

K IC
[14.5-7]
Intheunnotchedcondition:
c 0
0

N
Therefore:
K IC
0 a c
[14.5-8]
Hencethenotchsensitivitybecomes:
143
ECSSEHB3220Part2A
20March2011

N
a c
[14.5-9]
0 ca c
Anexampleofthestrengthratioversushalfcracklengthfora[0/90/45]sCFRPlaminateisgivenin
Figure14.5.4.
Figure 14.5-4 - Comparison between experiment and prediction by WEK model
If the equivalent damage zone ac is assumed to be independent of the crack length c, ac can be
calculatedwith0and derivedfromtestsanddifferentcracklengthc,usingEquation[14.510].
N
c

a c
1
0
2
N
[14.5-10]
BestfitcurvesfromtestsareshowninFigure14.5.5,Ref.[141].
144
ECSSEHB3220Part2A
20March2011
Figure 14.5-5 - Determination of ac for various laminates
14.6 WN fracture model
14.6.1 General
TheWNmodelpredictsthatfinalfractureofanotchedmultianglelaminateoccurswhenthestressin
acertainzoneexceedstheultimatestrengthoftheunnotchedmaterial.
Thedifferentfailurecriteriaformulatedare:
Pointstresscriterion.
Averagestresscriterion.
14.6.2 Failure criteria
14.6.2.1 Point stress criterion

Thepointstresscriterion,notationshowninFigure14.6.1,assumesthatfailureoccurswhenthestress,
x,oversomedistance,d0,awayfromthediscontinuityisequaltoorgreaterthanthestrengthofthe
unnotchedlaminate,Ref.[144]:
0, y
x 0
[14.6-1]
at: y Rd 0

145
ECSSEHB3220Part2A
20March2011
Figure 14.6-1 - Notation: Point stress criterion diagram
14.6.2.2 Average stress criterion

The average stress criterion, notation given in Figure 14.6.2, assumes that failure occurs when the
averagestress,x,oversomedistance,a0,equalstheunnotchedlaminatestrength,Ref.[144]:
R a0
1
0 0, y dy
x
[14.62]
a 0 R
146
ECSSEHB3220Part2A
20March2011
Figure 14.6-2 - Notation: Average stress criterion diagram
14.6.3 Characteristics of WN fracture model

FeaturesoftheWNfracturemodelare:
Bothstressfailurecriteriaaretwoparameterfracturemodels:xandd0ora0
For any material system they can be determined experimentally from testing unnotched and
notchedspecimens.
Itisrecommendedatleasttwoteststobeperformed,normallymoretestsareperformedtotake
intoaccountascatterintestresults.
The characteristic distances d0 and a0 depend on the type of laminate and are assumed to be
materialconstants.
The characteristic distances are determined experimentally for each material system and
laminatelayupindependently.
Whend0 anda0aredetermined,bothfailurecriteriacanbeappliedtopredictthetrendforthe
notchstrengthandthecriticalstressintensityfactor.
Excellent agreement with experiments when d0 and a0 are properly determined; illustrated in
Figure14.6.8andFigure14.6.9.
ResultsobtainedbyAwerbuch,Ref.[141],indicatethatd0anda0areindependentofnotchsize;
showninFigure14.6.11.
IncontrasttotheWEKmodel,[See:14.5],KIC=KQisnotamaterialconstant.
147
ECSSEHB3220Part2A
20March2011
Figure14.6.3showsadiagrammaticrepresentationofaninfiniteorthotropicplatewithcircularhole.
Figure 14.6-3 - Notation: Infinite orthotropic plate with circular hole

For an infinite orthotropic plate subject to a uniform stress,
x , applied parallel to the xaxis at
infinity,thenormalstress,xalongtheyaxisaheadoftheholecanbeexpressedas:
0, y 2

R 2 R 4 R 6 R 8
x
2 3 K
T
3 5 7 [14.6-3]
y y
x
y y
withtheorthotropicstressconcentrationfactorforinfiniteplates:
K

T
1 2 E E
1 2
12
12
0 .5 E 1 G 12
12
[14.6-4]
E1,E2,12yandG12yaretheeffectiveelasticmoduliofanorthotropiclaminate,[See:Figure14.6.3for
notation].
UsingEq[14.63]inconjunctionwiththepointstresscriterion:

2

N

0 2 3 K N 3 5 7
1
2
1
4
1
8
1
[14.65]
with:
148
ECSSEHB3220Part2A
20March2011
R
[14.6-6]
1
Rd 0

WhenR(largeholes,1=1): N 0
1 K T

WhenR0(smallholes,1=1): N 0
1.00
ForquasiisotropicmaterialsKT=3.0:
2
2 3

N 0
2 4
[14.67]
1 1
WhenR(largeholes, =1):

0
1 3
1 N

0
1.00
WhenR0(smallholes,1=0): N
ApplyingtheaveragestresscriteriontoEq[14.63]:

2 1
0
2
2 K 3
[14.68]
N 2 4 6 8

2 2 T 2 2
With:
R
2 R [14.6-9]
a0

1 K T

1.00
WhenKT=3.0(quasiisotropicmaterials):
149
ECSSEHB3220Part2A
20March2011

2 1

2
[14.6-10]

0 2 4
2
N
2 2

1 3

1.00
A comparison of analytical results for point stress criterion and average stress criterion with
experimentisgiveninFigure14.6.4,Ref.[141].
Figure 14.6-4 - Comparison between experiments and predictions of WN model
TheeffectsofthestackingsequenceofalaminateonthenotchsensitivityareshowninFigure14.6.5,
Ref.[141].
150
ECSSEHB3220Part2A
20March2011
Figure14.65Effectsofstackingsequenceonnotchsensitivity
TheeffectsoffibreanglevariationofalaminateonthenotchsensitivityareshowninFigure14.6.6,
Ref.[141].
Figure 14.6-6 - Effects of fibre angle on notch sensitivity
151
ECSSEHB3220Part2A
20March2011
14.6.5 Straight cracks

Figure 14.06.7 shows a diagrammatic representation of an infinite orthotropic plate with straight
crack.
Figure 14.6-7 - Infinite orthotropic plate with straight crack
Thestressinthevicinityofacrackinanorthotropiclaminateofinfinitewidthunderuniaxialloading,
x ,yieldsasingularityatthecracktip.
Theexactexpressionforthestressaheadofthecrackcanbewritten,afterRef.[1415],as:

y
0, y
x 2
x
2
yc [14.6-11]
y c
Thisisindependentofmaterialproperties.
152
ECSSEHB3220Part2A
20March2011
Becauseofthesingularityofthecracktiptheconceptofastressconcentrationfactorisreplacedbya
stressintensityfactor.
Foruniaxialtension(ModeI)thisfactorisdefinedas:
K

I x
c [14.612]
Thestressdistribution,Eq[14.611],intermsofthestressintensityfactorisgivenby:
0, y K y I
yc
c y c [14.6-13]
x 2 2
Forlargecracks,theapproximateexpressionis:
0, y 2K I c
x
y c
x
2 y c [14.6-14]
ThisgivespredictionsclosetotheexactsolutionsofEq[14.6.11]andEq[14.613].Forsmallercracks
Eq[14.611]or[14.613]canbeapplied.
14.6.6 Point stress criteria

Applying the point stress criterion in conjunction with Eq [14.611] or Eq [14.613], the notch
sensitivityofaninfinitelaminateplatewithacentrecrackbecomes:
1
2
N 0
[14.6-15]
3
With:
c
c
3
[14.6-16]
d 0
Thecriticalstressintensityfactor K Q becomes:
K Q
0 c1 2
3

[14.6-17]
153
ECSSEHB3220Part2A
20March2011
UsingEq14.06.14gives:
K Q
0 2 d 0 const [14.6-18]
14.6.7 Average stress criterion

Applying the average stress criterion in conjunction with Eq [14.611] or Eq [14.613], the notch
sensitivityratiogives:
1

4
[14.6-19]
N 0
1 4
With:
c
4 [14.6-20]
c a0
And:
2c
a [14.6-21]
N 1
0
0
Thecriticalstressintensityfactorbecomes:
K Q
0 c1 1 4 4
[14.6-22]
UsingEq[14.614]:
K Q
0 a0 2 const [14.6-23]
FromEq[14.619]andEq[14.620]:

N 0
a 2 c a 2
0 0
[14.6-24]
154
ECSSEHB3220Part2A
20March2011
ThisissimilartoWEKfracturemodel,[See:14.5(Eq[14.59])]witha0=2ac.
Figure 14.06.8 shows a good correlation between the analytical results for both criteria with tests,
whileFigure14.06.9showsthecorrelationofcalculatedstressintensityfactorswithtests,Ref.[141].
Figure14.68ComparisonbetweenexperimentsandpredictionsofWNmodel
Figure 14.6-9 - Critical stress intensity factor KQ
155
ECSSEHB3220Part2A
20March2011
Best fit curves for the determination of a0 from test are shown in Figure 14.06.10 for different
laminates,Ref.[141].
Figure 14.6-10 - Determination of a0 for various laminates, applying 'Average stress

criterion'
TheagreementwithresultsfromtheWEKmodelisfairlygood,[See:Figure14.05.4].
The WN model assumes that the characteristic lengths d0 and a0 are material constants and are
independentofcracklength.
FromFigure14.06.11,Ref.[141],itcanbededucedthatthiscanbetrueford0(pointstresscriterion)to
agreaterextentthanfora0(averagestresscriterion).
156
ECSSEHB3220Part2A
20March2011
Figure 14.6-11 - Characteristic dimensions as a function of half crack length
TheeffectofthelaminatelayuponnotchsensitivityisshowninFigure14.06.12,Ref.[141].
Figure14.612Effectofstackingsequence
157
ECSSEHB3220Part2A
20March2011
14.7 Finite plate models

TheWEK,[See:14.5]andWN,[See:14.6],fracturemodelswereformulatedassumingplatesofinfinite
width.
For a real application consisting of finite plates, it is necessary to add a correction factor. There are
severalanalyticalmethods,includingfiniteelementmethods,todeterminethefinitewidthcorrection,
FWC,factor,Ref.[141].
Fororthotropicmaterials,noclosedformsolutionsareavailable;thereforetheclosedformexpressions
forisotropicmaterialsareoftenused.
TheapplicabilityoftheisotropicFWCfactortoorthotropicmaterialsisjustifiedforthecaseshaving,
Ref.[141]:
Largelengthtowidthratios,whereL/W>3.0.
Largefreeedgedistances,where2c/word/w<0.5.
[See: 14.8 for numerical results for finite width correction; 14.9 for results from calculated stress
concentrationfactorsatcircularholes]
14.8 Finite width correction (FWC)
14.8.1 General
TouseisotropicFWConorthotropicmaterials,ananisotropyfactor,H,isaddedwhichdependson
thespecimengeometry,layupandmaterialproperties:
H Y Y iso
[14.81]
Where:Happroachesunity.
Values of H for various CFRP laminate layups are given in Figure 14.08.1 and Figure 14.08.2,
Ref.[141].
158
ECSSEHB3220Part2A
20March2011
Figure14.81Anisotropyfactorversusnetsectionreductionforcentrecracked
angleplyCFRPspecimens
159
ECSSEHB3220Part2A
20March2011
Figure14.82Anisotropyfactorversusnetsectionreductionforcentrecracked
CFRPspecimens
DefinitionoftheFWCfactor,Yiso,asgiveninEq[14.83]orEq[14.84],Ref.[1416],[1417],gives:
Y N

N
[14.82]
and: K Q Y N
enablingacomparisonbetweenexperimentalresultsandpredictions.

Forplatescontainingholes,Ref.[1416]recommendstheFWC:
2 1 2 R W
3
Y iso 31 2 R W
[14.83]
160
ECSSEHB3220Part2A
20March2011
14.8.3 Centre crack

Forplatescontainingcracks,Ref.[1417]recommendstheFWC:
2 3
2c 2c 2c
Y iso 1 0.1282 W 0.2881 W 1.5254 W [14.84]
14.9 Calculated stress concentration factor at circular

holes
14.9.1 NASA results

In Ref. [1418], stresses are calculated for finitewidth orthotropic laminates under uniaxial tension
load.Atwodimensionalfiniteelement,FE,analysiswasused.
Table 14.09.1 summarises the HT carbon/epoxy laminate configurations and elastic constants,
Ref.[1418]
Table14.91HTcarbon/epoxy
ElasticConstants
Laminate Ex Ey Gxy
xy
(MPa) (MPa) (MPa)
Quasi-isotropic [0/45/90]s 57,890 57,890 22,090 0.310
0 146,900 10,890 6,412 0.380
90 10,890 146,900 6,412 0.028
[0/90]s 79,500 79,500 6,412 0.052
[45]s 22,250 22,250 37,770 0.735
161
ECSSEHB3220Part2A
20March2011
Figure 14.9-1 - Stress concentration factor KTn for orthotropic CFRP laminates with a
circular hole, 2 L W 10
Figure14.9.1showsthatanisotropyhasastronginfluenceonthestressconcentrationfactor(KTn)for
longstripswithlargew/dvalues.
KTnisdefinedastheratioofmaxattheedgetothetensilenetsectionstressnwhere:
NW
n
W d
Where:Nisthegrosssectiontensilestressappliedacrossthelaminate.
Forw/d,Table14.9.2givesvaluesof,[Seealso:Table14.9.1forlaminateproperties].

Table14.92 K T forvariouslaminates,w/d
Laminate K

T
[0] 6.43 90
[90] 2.48 90
[45]s 2.89 60
[0/90]s 4.78 90
[0/45/90]s 3.0 90
Valuesof K T inTable14.9.2canbeeasilycalculatedbyusingEq[14.62],[See:14.6].
162
ECSSEHB3220Part2A
20March2011
14.9.2 Finite width correction (FWC)

Ref.[1418]notesthatforsufficientlylargeL/W(L/W=10)andw/d>3theFWCcanbeapproximated
byusingthequasiisotropiccase,[See:14.8].
14.9.3 MBB/ERNO study

The stress distribution at holes in an infinite plate can be calculated according to Ref. [146]. The
resultsofthesecalculationsforsomeCFRPHTandHMlaminatesaregivenfor:
0/0laminatesinFigure14.9.2.
32.5/0/32.5laminatesinFigure14.9.3.
+30/0/30(HM)and+45/0/45/90(HT)laminatesinFigure14.9.4.
+30/0/30(HM)and90/90(HT)laminatesinFigure14.9.5.
ThematerialdataisgiveninTable14.9.3.
Table14.93Materialdata
Material/ Ex Ey Gxy xy
KT
StackingSequence
(a)914CMS440
(HighModulusCFRP=HM)
0/0 221000 7300 5300 0.330 8.21
+30/0/30 116956 12874 30686 1.572 3.59
32.5/0/32.5 84413 14279 38663 1.523 3.00
+45/0/45/90 80507 80507 30686 0.312 3.00
+30/30 58376 10260 43379 1.771 2.60
90/90 7300 221000 5300 0.011 2.31
(b)914CTS440
(HighTensionCFRP=HT)
0/0 128000 8850 5400 0.300 6.54
32.5/0/32.5 59248 13148 23596 1.134 3.12
[See:Figure14.9.2,Figure14.9.3,Figure14.9.4,andFigure14.9.5]
163
ECSSEHB3220Part2A
20March2011
[See:Table14.9.3formaterialdata]
Figure14.92Stressdistributionataholefor0/0laminates
164
ECSSEHB3220Part2A
20March2011
Figure14.93Stressdistributionataholefor32.5/0/32.5laminates
165
ECSSEHB3220Part2A
20March2011
Figure14.94Stressdistributionataholeforlaminates
166
ECSSEHB3220Part2A
20March2011
Figure14.95Stressdistributionataholeforlaminates
Thepointofintersectionofthestresscurvewiththeordinateindicatesthestressconcentrationfactor

K T .
167
ECSSEHB3220Part2A
20March2011
Fordistancesy2R,thenetsectionstress,x,becomes:
1.2

x x

Finitewidthcorrectioncanbeappliedwhennecessary,[See:14.8].
14.10 Stress distribution around circular holes
14.10.1 General
The strength of composite plates with holes is less sensitive to stress concentration than those for
isotropic materials. A number of authors have investigated the problem of the stress concentrations
aroundholesinorthotropicplates,Ref.[141],[145],[1421],[1422],[1423].
TheresultspresentedherearebasedontheworkofLekhnitski,Ref.[1415],andaregivenforplates
containingcircularholessubjectedtotensileorshearloading.
Theanalysisisbasedontheassumptionsthatthe
Platewidthismuchgreaterthanthediameterofthehole,
Behaviourofthematerialislinearelastic,
Holeisopenandunloaded,
Laminateisorthotropic.
14.10.2 Stress concentration due to tensile load

Figure14.10.1showsthenotationfortensileloadcase.
Figure14.101Stressconcentrationduetotensileload
168
ECSSEHB3220Part2A
20March2011
Onthebasisoftheassumptions,thetangentialstressAisdefinedby:

E cos 2 m n sin 2 m cos 2 1 n cos 2 m sin 2
A sin 2 n1 m n sin cos sin cos

[14.101]
Ex
With:
sin 4 1 2 xy 2
sin cos 2 cos
4
1

E E G
x xy E x Ey
m Ex
Ey

n 2 E x Ex
xy
E y G xy
Themaximumtangentialstress(=90)gives,for=0,theconcentrationfactorKt:
A 90

Kt

E x E x
K t 1 2 xy
Ey G xy
14.10.3 Stress concentration due to shear load

Figure14.10.2showsthenotationforshearloadcase.
169
ECSSEHB3220Part2A
20March2011
Figure14.102Stressconcentrationduetoshearload
Usingthesamebasicassumptionsasfortensileloading,thetangentialstressAisdefinedby:
2EE 1 m n n cos 2 sin 2 1 mcos 2 m 1sin 2

A

With:
m E x

E y

n 2 E x Ex
xy
E y G xy
In contrast to the tensile load case, the location of the maximum tangential stress for shear loading
dependsonthelayup.
For=0,theconcentrationfactorbecomes:
max A
Ks

170
ECSSEHB3220Part2A
20March2011
14.11 Interlaminar fracture mechanics
14.11.1 Nomenclature
a delaminationlength
A delaminationarea,andconstantindelaminationgrowthexpression
B,D1,D2 exponentindelaminationgrowthrateexpression
G strainenergyreleaserate
Ginit strainenergyreleaserratetoinitiateadelamination
GI,GII,GIII individualmodesofstrainenergyreleaserate
GIc,GIIc,Gc modeI,modeIIandtotalinterlaminarfracturetoughness
Gmax maximumcyclicstrainenergyreleaserate
GR valuesoffracturetoughnesswhenanRcurveispresent
Gth cyclicthresholdofstrainenergyreleaserate
Gtot totalstrainenergyreleaserate
G amplitudeofcyclicstrainenergyreleaserate
N elapsedfatiguecycles
Ninit fatiguecyclestodelaminationinitiation
U strainenergy
14.11.2 Delamination and fracture mechanics overview
14.11.2.1 Delamination initiation and growth

Delamination, or interlaminar cracking, is a common type of damage in composite materials.
Delaminations are debonds in the resin rich regions between the laminated plies of a composite.
Delaminationcanbeasourceofprimaryfailureincompositestructures.
Delaminations can arise from interlaminar stresses due to geometric or material discontinuities at
designfeatures,suchasedges,holes,jointsordroppedplies.Theycanalsooccurfrommatrixcracks
orfrominterlaminarstressescausedbystructuralloading,suchasinacurvedlaminate,orbyforeign
bodyimpacts.
Itisalsopossibletoinadvertentlyintroducedelaminationsintostructuresatthemanufacturingstage.
The delamination, once initiated, extends under fatigue loading. During delamination growth, the
structural loads can be redistributed such that other delaminations occur elsewhere. The
delaminationsthencontinuetogrowandaccumulateuntilastructuralfailureoccurs,suchasbuckling
or fibre failure. Individual delaminations can cease to grow due to changes in overall stress
distribution.
171
ECSSEHB3220Part2A
20March2011
Theprogressivedamagecanresultinfinalstructuralfailure,whichisdependentonthelocalloading.
Iftheloadingiscompressive,sublaminatebucklingcanoccur,ortensilefibrefailureiftheloadingis
tension.
Mechanicalloadingisoftenthemaincauseoffailure.However,theeffectsofenvironment,including
temperature and fluid exposure under normal operating conditions, also need to be accounted for.
Generally,ahostileenvironmenthasanegativeeffect,reducingthefatiguelifeofthestructure.The
effectsofmechanicalandenvironmentalloadacttogetherandcannotbeinvestigatedinisolation.
14.11.2.2 Linear elastic fracture mechanics and the strain energy release rate
approach
Efforts to predict the initiation and growth of delamination have focused on interlaminar fracture
mechanics.Forthis,thechangeinstrainenergyperunitareaofdelaminationgrowth;knownasthe
strainenergyreleaserate,G,canbecalculatedusingtheformula:
U
G [14.11-1]
A
where:
U:strainenergyofthebody,
A:thecracksurfacearea
Calculated values of G for a delamination in a structure, are compared with the critical values
obtained from materials tests to determine whether the delamination is likely to grow. The critical
valuesofGdependuponthewayloadingisapplied.Thetotalstrainenergyreleaserateisconsidered
intermsofthreecomponentscalledfracturemodes,asshowninFigure14.11.1.
Mode I Mode II Mode III z

Opening, Slid ing, Tearing
Cleavage Sh ear
x
Figure14.111Fracturemodes:opening,shearing,andtearing
172
ECSSEHB3220Part2A
20March2011
Thefracturetoughnessmodesaredescribedas:
ModeIopening,cleavage,
ModeIIsliding,shear,
ModeIIItearing.
Thetotalstrainenergyreleaserateisgivenbysumofthecomponents:
GT=GI+GII+GIII
ThecriticalvalueofG,inpureModeIiscalledGIc.ThecriticalvalueofGinpureModeIIiscalledGIIc.
ThecriticalvalueofGinpureModeIIIiscalledGIIIc.Thethreecomponentsofloadingarenormally
appliedincombinationcreatingamixedmodeloadingcondition.
Coupontestshavebeendevelopedtomeasurefracturetoughnessinbothpureandcombinedmodes.
Strain energy release rate (SERR) can be calculated for a given delamination using finite element
analysis(FEA)ofacompositestructure.TheprincipalmethodusedforcalculationofSERRusingFEA
isthevirtualcrackclosure(VCCT)method.
14.11.2.3 Using fracture mechanics to identify the critical delamination

location
A flow diagram showing the overall framework for modelling damage using interlaminar fracture
mechanicswithrespecttostructuralintegrityisgiveninFigure14.11.2.
Identify areas of high interlaminar Identify damage threats and

stress/strain gradients for potential potential extent of damage from
delamination initiation sites threat
Conduct fracture analysis to

determine if delamination will
initiate or begin to grow
N Does
delamination
STOP
initiate or begin to
grow?
Y
Determine the extent to

which the delamination will
grow
As crack grows, identify load

redistribution in other areas of
the component and identify
further damage or failure
Figure14.112Outlineofnodamagegrowthmethodology
173
ECSSEHB3220Part2A
20March2011
The methodology begins by investigating areas of concern or hotspots with high interlaminar
stressesandstrainsthatresultinthematrixbeinghighlyloaded.Theseinterlaminarstressescanarise
fromgeometryorfromdamagethreats.
Once identified, a fracture analysis is conducted to determine whether a delamination initiates and
grows. If a delamination does grow, the analysis continues to assess how far it grows and if that
growth is likely to lead to structural failure. Component failure is dependent on the loading in the
area of interest, hence this needs to be defined. In many situations a significant reduction in the
stiffnessofthecomponentcouldbedeemedtodefinefailure.
If component failure does not occur, then the analysis can be stopped, or another potential
delaminationsiteexamined.Ifthedelaminationarisesfromastructuraldiscontinuity,thenthatlocal
area can be redesigned to prevent initiation at that site. If the delamination grows in a stable,
predictable,manner,theninspectionintervalscanbeestablishedtoensurethatthedelaminationhas
notgrownbeyondthatexpected.
Identification of interlaminar stresses and fracture properties need accurate modelling of through
thicknessproperties.Oftenthesemodelsarealocalmodeloftheareaofconcernwitheitheridealised
boundaryconditionsapplied,ortheactualboundaryconditionsfromtheglobalmodelusingglobal
localmodellingtechniques.Selectionoftheareaofconcernisoftendonebyengineeringjudgement,
especially if global finite element models include 2D shell elements which do not identify areas of
highinterlaminarstress.
Thestressoutputsfromthelocalmodelsenabletheidentificationofinterlaminarstressesawayfrom
anystresssingularitiesassociatedwithgeometricdiscontinuities.Thesestressesorstrainscanbeused
with suitable failure criteria, suchas maximum strainstress or interaction criteria, or withthe other
fracturecriteriadescribedinthischapter.Ifaninterlaminarfailureoccurs,itdoessoasadelamination
and the extent to which it grows is determined using interlaminar fracture mechanics. If a
delamination is not predicted to initiate, then other structural discontinuities are examined. The
stressesattheselocationsareFEAmeshsizedependentand,hence,areindefinite.Thus,thenumerical
values are not used to predict failure in or near these locations. However, the trend of the data is
important.Delaminationsprimarilyoccurwhentheinterlaminartensionstressesarehigh.Theycan
also grow where there is a high interlaminar shear stress. Therefore, these areas are identified for
fracture analysis. In a damaged structure, invariably, the damage region is where further
delaminations can grow. For a damaged component and for a component where delamination is
expected to initiate, a fracture analysis is conducted for delaminations growing either from the
damageorfromthepotentialinitiationsite.
14.11.2.4 Fracture mechanics prediction procedure

Interlaminarfracturemechanicsanalysis,usesvaluesofthestrainenergyreleaserateforagivenload
at different delamination lengths. These values are then compared with the materials interlaminar
fracturetoughnessandfatiguecrackgrowthproperties.
ThefracturemechanicspredictionprocedureissummarisedinFigure14.11.3.
174
ECSSEHB3220Part2A
20March2011
G-N DATA
G
FRACTURE ANALYSIS Gc
G/2 Initiation PREDICTION

Growth Cycles
Gc c
Initiation
c2
ac
Crack length da/dN DATA Cycles

da/dN
To structural
failure
Gc
G
Figure14.113Fracturemechanicslifepredictionmethodology
Thefirststepistoanalysethemodeltoobtainarelationshipbetweenstrainenergyreleaserate,Gand
delaminationlength,a,foragivenstressorstrain.Foraninitialdelaminationlength,acriticalvalue,
Gc,isobtained.TheGcanalyticalvalueiscomparedwiththematerialdelaminationdatatogivestatic
failureloadandlocationandthenumberofcyclesnecessarytoinitiatethedelamination.
FromthisvalueofGc,thedelaminationgrowthrate,da/dNisobtainedusingthematerialfatiguedata.
For an incremental number of cycles, the increment in delamination length is determined. This
providesanewvalueofGfromtheGarelationshipandanewvalueofda/dNforeachincrement.This
continuesuntileitherthedelaminationgrowstoacriticallengthorstopsgrowing,soprovidingthe
fatiguelifeatthegivenload.
14.11.3 Standard test methods (static and fatigue)
14.11.3.1 General
The discussion of interlaminar fracture toughness testing methods given here is more detailed than
theoverviewoftestmethodsforcomposites,[See:Chapter7].
14.11.3.2 Interlaminar fracture toughness testing

Coupon tests are used to measure the interlaminar fracture toughness. The majority of the tests
developed to date are designed to individually measure Mode I, Mode II or Mode III fracture
toughness.However,tousefracturetoughnessdatafordelaminationinitiationandgrowthstudies,it
needstobedeterminedusingrepresentativeloadingconditionsforadelaminationinthecomposite
structure.Fracturetoughnesschangeswiththeloadingmode.
175
ECSSEHB3220Part2A
20March2011
Itislikelythatthedelaminationwillbesubjecttomixedmodeloadinginarealstructuralsituation.
Themixtureofmodescouldalsochangeasthedelaminationgrows.Nosingletestmethodhasbeen
developedtomeasureanarbitrarymixtureofModeI,ModeIIandModeIII.
MixturesofModeIandModeIIarethemostcommonloadingconditionsseenatdelaminationfronts
in composite structures. Test methods are available to measure arbitrary mixtures of Mode I and
ModeII,mostnotablytheMMBmixedmodebendingtest.
There are limitations on how fracture toughness tests can be used, such as the limitations on
unidirectional layups and fibre bridging. For most of the tests, initiation and precrack data can be
measured and a crack growth resistance (Rcurve) obtained. The initiation data often provides the
mostconservativevalueforuseinfracturemechanicsbaseddesignstudies.
14.11.3.3 Types of test methods

Varioustestmethodshavebeendevisedtoassessfracturebehaviour.Theseinclude:
DCBdoublecantileverbeam,usedforModeI,asshowninFigure14.11.4.
CNcentrenotchforModeI.
DTdoubletorsionforModeI.
ENFendnotchflexureforModeII,showninFigure14.11.4.
4ENF4pointbendendnotchflexureforModeII.
ENSendloadedsplitforModeIIusedforthinlaminates.
ELSendlapshearforModeII.
ENCBendnotchedcantileverbeamusedforthicklaminates.
CBENcantileverbeamenclosednotchforModeII.
EDTedgedelaminationtensiontestformixedModesIandII.
CLScrackedlapshearformixedModesIandII.Thiswasdevelopedforassessingadhesives.
DSDCBdoublesplitdoublecantileverbeamforModeIII.
ArcantestarrangementformixedModeIandII.
ECTedgecracktorsiontestforModeIII
MMBmixedmodebendingforModesIandII
Someofthesetestsaremorewidelyacceptedthanothers.Themostcommonlyusedtestmethodsare
giveninTable14.11.1.
176
ECSSEHB3220Part2A
20March2011
25
10
Double Cantilever Beam Specimen
Typical Dimensions in mm
Drawings not to scale
25 140
65
25
End Notched Flexure Specimen

25 25
Figure 14.11-4 - DCB and ENF specimens
Table14.111Standardtestmethods
Testmethod(Standard) Comments
ModeIassessmentbasedonESISdocumentfromUSA.
ISO/CD13,586 (ModeIIbeingstudiedinVAMASleadproject).
ASTMD552894a: DoublecantileverbeamModeI
ASTMD6671 MixedmodebendingModeI+II
AirbusprEN6033 DeterminationofModeI(GIc)forCFRP
AirbusprEN6034 DeterminationofModeII(GIIc)forCFRP
14.11.3.4 Fatigue delamination characterisation

Thereareessentiallytwoapproachesforcharacterisingfatiguedelamination.Thesemethodsarethe
delamination initiation method and the delamination growth method. Both methods are needed in
ordertoaccuratelypredictthestructuralintegrityofacomponent.
The DCB and ENF specimens, as shown in Figure 14.11.4, have been used to characterise fatigue
delaminationbymonitoringthedelaminationgrowthperfatiguecycle,da/dN.
Alternatively,a4ENF,Ref.[1424],specimen(fourpointbendendnotchedflexurespecimen)canbe
usedtodetermineGIIth.The4ENFhasadecreasingcrackgrowthratewithcracklength.
Expressionsrelatingtheappliedcyclicstrainenergyreleaserate(GmaxorG)withda/dNasapower
law are given in Equation [14.112]. This equation can be further extended to incorporate the
accelerationofdelaminationgrowthasGapproachesthequasistaticvalueandthedecelerationasthe
delaminationgrowthratereachesathresholdvalue,givingEquation[14.113].
AGmax B
da
[14.112]
dN
177
ECSSEHB3220Part2A
20March2011
G D1
1 th
Gmax
AGmax
da B [14.113]
dN D1
1 Gmax
Gc

Forcompositeshowever,thevaluesoftheexponentBinthesepowerlawsaregenerallyhigh.They
canrangefromapproximatevaluesof3toinexcessof15.Thus,usingonlythefatiguegrowthpartof
the curve, any small deviation from the anticipated service load can lead to large errors in the
predicted delamination growth rate. This is because G is proportional to the square of the load and
hencethedelaminationgrowthrateisproportionaltotheloadtothepowerof2B.Thispotentialerror,
and the high rates of delamination growth, have led to the adoption of a no growth approach to
design and damage tolerance characterisation of composite structures in fatigue. The no growth
approach needs the threshold value of strain energy release rate, Gth to be determined. Then, if a
delamination is present, and the strain energy release rate is below the Gth value, the delamination
doesnotgrow.
Forhighcyclefatigueissues,thevalueofGthisevaluatedforcrackgrowthratesintheorderof108
mm/cycle sothat threshold valuesareguaranteed to be below108 mm/cycle or 1mmin108 cycles.
Thiscriteriaisconfirmedwiththeallowabledamageintheapplication.
In generating the da/dN data, the delamination needs to be grown during the fatigue test. Fibre
bridgingcanoccurduringthisgrowthgivingartificiallyhighvaluesofGthandexponentB.Therefore,
for more representative values, the G data, is normalised by the Rcurve This is achieved by
determining GImax/GIR for each of the fatigue data points where GIR is the value of the static fracture
toughnessatthecorrespondingdelaminationlengthofthatfatiguedatapoint.
Owing to fast delamination growth, emphasis is often placed only on the cycles to produce
delaminationinitiation,delaminationgrowthisnotdetermined.Aconservativeapproachistoassume
thatonceadelaminationhasinitiated,itgrowstoitsfullextent,onagivenpath,inthenextloadcycle.
AcompleteGNdelaminationonsetcharacterisationcanbeobtainedforagivenmodemix.
Hence,bytestingmodeI,mixedmodeandmodeIIspecimens,acompleteGNcharacterisationcanbe
developedforallmodes,asshowninFigure14.11.5.Bychoosinganappropriatenumberofcyclesfor
the application, Na, such as the number of cycles between inspections, or the fatigue life, a design
valueofGcanbespecifiedsuchthatthedelaminationdoesnotgrowwithinthosecycles.
178
ECSSEHB3220Part2A
20March2011
Figure 14.11-5 - Representative G-Nonset data
14.11.4 Calculation of strain energy release rate in structural

analysis
14.11.4.1 Global energy comparisons

Strain energy release rate, G can be calculated using a global energy balance between two finite
elementanalysissolutionsofdifferentdelaminationareas(AiandAi+1):
dU U a U ai
G i1 [14.11-4]
dA Aai1 Aai
Where: UisthestrainenergyandAisthedelaminationareafortheithsolution
Thisgivesavalueforthetotalstrainenergyreleaserate,Gtot.Thelimitationtothisapproach,however,
isthatthecomponentparts,GI,GII,andGIIIcannotbedeterminedusingthismethod.Othermethods
areavailableforcalculationofGsuchas:
Virtualcrackclosuretechnique(VCCT)
Jintegralmethod
Cracktipelementmethod
179
ECSSEHB3220Part2A
20March2011
14.11.4.2 Virtual crack closure technique (VCCT)
Thevirtualcrackclosuretechniqueisusedtocalculatestrainenergyreleaserateinamixedmodefor
compositematerials.ItiseasytoimplementinFEApackages,Ref.[1425].Thetechniqueisbasedon
the work done to close the crack tip. An assumption can be made which allows values of G can be
calculated from the results of one finite element solution rather than two. If the increment of
delaminationgrowth,a,issmallincomparisonwiththeoveralldelaminationlength,thentheforces
aheadofthedelaminationcanbeassumedtobesimilartothosewhenthedelaminationhasgrown.
Gvaluesareplottedattheexistingdelaminationlength.Generally,theglobalchangeinenergyand
VCCT approaches give similar results. If they do not, and there is no error, then either there was a
significantchangeofgeometrybetweenthetwodelaminationlengthsusedtocalculatedU/dA,orthe
delaminationlengthisshortandtheVCCTassumptionofabeingsmallcomparedwiththeoverall
delaminationlength,hasbeenviolated.
Owing to the numerical oscillatory nature of the stresses at the delamination tip between different
materials,orpliesofdifferentorientation,themoderatioisdependentonthesizeoftheelementsat
thedelaminationfront,Ref.[1426],[1427].Whentheelementsizeisreduced,theGI/Gtotratiochanges
andthevaluesdonotconvergewithcontinueddecreaseinelementsize.Onesolutiontothisproblem
is to incorporate a resin layer between the plies and to grow the delamination within this. The
disadvantageofthismethodisthatthemodelingcanbecomeintensivebecauseofthesmallthickness
oftheresinlayer.Inaddition,itisnotalwayspossibletodeterminethesizeoftheresinlayerbetween
pliesand,hence,anassumptiononitsthicknessisoftenapplied.
An alternative method to determine the mode mix ratio is to modify one of the material properties
(normallyaPoissonsratio)toafictitiousvalueinordertoforcethebimaterialconstantinthenear
cracktipstressfieldequationtozero,Ref.[1427],[1428].
14.11.4.3 2D VCCT
In 2D using 4noded quadrilateral elements the values of GI and GII are calculated by using the
equations:
1
GI FyF (V B V C ) [14.115]
2A
1
GII FxF (U B U C ) [14.116]
2A
In 2D using 8noded quadrilateral elements the values of GI and GII are calculated by using the
equations:
GI
1
2A

FyF (V B V C ) FyG (V D V E ) [14.117]
GII
1
2A

FxF (U B U C ) FxG (U D U E ) [14.118]
180
ECSSEHB3220Part2A
20March2011
Where:
A istheelementlength
FxF , FxG , FyF , FyG arexandycomponentsofforceatnodesFandG

UB,UC,UD,UE arexcomponentsofdisplacementatnodesB,C,DandE
V ,V ,V ,V areycomponentsofdisplacementatnodesB,C,DandE
B C D E
B D
F G
C E
A A
x
Thexydirectionsareinalocalcoordinatesystem.ThelocationsofnodesBtoDrelativetothecrack
tipareshowninFigure14.11.6.
Figure 14.11-6 - Node definitions for 2D VCCT
14.11.4.4 3D VCCT
In 3D, values of G are calculated along the entire crack front. The equations used to calculate the
componentsofGusingVCCTforboth8nodedand20nodedbrickelementsare:
Foran8nodedbrickelement:
1
GI FzH (W B W E ) [14.119]
2A
1
GII FxH (U B U E ) [14.1110]
2A
181
ECSSEHB3220Part2A
20March2011
1
GIII FyH (V B V E ) [14.1111]
2A
Fora20nodedbrickelement:
1 H FzG FJ
GI Fz (W W ) Fz (W W )
B E
(W A W D ) z (W C W F ) [14.1112]
2A 2 2
1 H B FxG FJ
G II
Fx (U U ) Fx (U U )
E
(U A U D ) x (U C U F ) [14.1113]
2 A 2 2
1 H B
FyG A FyJ C
G III Fy (V V ) Fy (V V )
E
(V V )
D
(V V F ) [14.1114]
2A 2 2
Where:
A isthecrackareaassociatedwiththeGcalculation
FxiAi isthecomponentofforceforxi=x,y,zandAi=nodesG,H,J,
UAi isthexcomponentofdisplacementforAi=nodesA,B,C,D,E,F,,
V
Ai istheycomponentofdisplacementforAi=nodesA,B,C,D,E,F,,
W Ai isthezcomponentofdisplacementforAi=nodesA,B,C,D,E,F,,
The definition of the nodes relative to the crack front position at which G is calculated is shown in
Figure14.11.7.
Forthe8nodedbrickcase,nodesA,C,D,F,G,J,,anddonotexist.Also,forthecaseofthe20
noded element where the crack is at a free edge, there are only four elements surrounding node H
(insteadofeight)andhencethetermsinvolvingnodesC,FandJareignored.
182
ECSSEHB3220Part2A
20March2011
A
C
B F
J
A E
H
D
G
Figure14.117Nodedefinitionsfor3DVCCT
14.11.4.5 Using interlaminar fracture mechanics in design

AnexampleofaGaplotforadelaminationunderastiffenerisdescribedinRef.[1429]andshownin
Figure14.11.8.
TheplotshowsGtotcalculatedusingEquation[14.114]andGIandGIIusingVCCTina2Dplanestrain
modelanda3Dmodelusingplateelements.Thistypeofplotgivesthreepiecesofinformationthat
help predict delamination initiation and delamination extent in the component. Firstly, the value of
Gtotcontainsapproximately60%GI.Inmostmaterials,themodeIinterlaminarfracturetoughness,GIc,
is significantly less than the mode II interlaminar fracture toughness, GIIc. For brittle epoxies, this
differencecanbeashighasanorderofmagnitude.Significantdifferencescanalsobeseeninfatigue
initiationandcrackgrowthtesting.Hence,ifthegeometryofthepart,thelayup,orthematerialcan
bealteredinthedesigntohavealowerGIcomponent,thepanelbecomesmoredamageresistantand
damagetolerant.
183
ECSSEHB3220Part2A
20March2011
1000
Plane strain model
G (J/m2) Plate model
800
Total
600
Mode I
400
Mode II
200
0 4 8 12 16 20
Debond Length (mm)
Figure14.118Gacurveforastiffenerdebond
The second piece of information from Figure 14.11.8, is the general slope of the Ga plot. For all
delamination lengths the slope is positive, indicating that delamination growth is unstable under
quasistaticloading,orthefatiguedelaminationgrowthrateincreaseswithdelaminationgrowth.
SchematicsforothertypesofGaplotsareshowninFigure14.11.9.

G G Initially positive slope
Negative slope (Stable
delamination growth)
Gcrit
Gcrit
Crack arrest
a a
Figure 14.11-9 - Schematics of possible G-a curves
InFigure14.11.9a,theslopeoftheGaplotisalwaysnegative,thusunderquasistaticloadingmore
loadisneededtogrowthedelaminationfurther.Infatigue,thedelaminationgrowthratedecreases
untilthedelaminationeventuallystopsgrowing.
In Figure 14.11.9b, the slope of the Ga plot is initially positive but then becomes negative. This
indicatesthatthedelaminationgrowthisunstableatfirstandthenstops.Thisbehaviourcanbeused
instructuraldesigngeometries.
184
ECSSEHB3220Part2A
20March2011
The third piece of information from Figure 14.11.8 is the value of G necessary to initiate the
delamination.Astresssingularityexistsbecausethereisasuddenterminationofthedelaminationat
the start of the crack. In the close vicinity of the stress singularity, the Ga curve probably turns
sharply towards zero. However, if the Ga curve isextrapolatedto a zero crack length ignoring this
sharpturn,itispossibletousetheextrapolatedvalueofGtocomparewiththeinterlaminarfracture
toughness,Gc,ofthematerial.Again,localdesignoptimisationcanincreasetheanalyticalvalueofG,
givingamoredurablepart.
This methodology has been validated against many different configurations, including: Edge
delamination in a flat multidirectional laminate, Ref. [1430], a tapered laminate, Ref. [1431] and a
curvedlaminate,Ref.[1432].
14.12 References
14.12.1 General
[141] J.Awerbuch
NotchedStrengthofCompositeLaminates
InternalReportIB13185/04,DFVLR(Germany)
ESACR(P)2017(Vol.4,AppendixH),January1985
[142] E.M.Wu
FractureMechanicsofAnisotropicPlates
CompositeMaterialsWorkshop,TechnomicPub.Co.,Inc.Stamford,
Conn.,1968
[143] M.E.Waddoups,J.R.Eisenmann&B.E.Kaminski
MacroscopicFractureMechanicsofAdvancedCompositeMaterials
J.CompositeMaterials,Vol.5,1971
[144] J.M.Whitney&R.J.Nuismer
StressFractureCriteriaforLaminatedCompositesContainingStress
Concentrations
UniaxialFailureofCompositeLaminatesContainingStress
Concentrations
FractureMechanicsofComposites,ASTMSTP593
AmericanSocietyofTestingandMaterials,1975
[146] R.F.Karlak
HoleEffectsinaRelatedSeriesofSymmetricalLaminates
ProceedingsofFailureModesinComposites,IV
TheMetallurgicalSocietyofEIME,Chicago,1977
[147] R.B.Pipes,R.V.Wetherhold&J.W.Gillespie,Jr.
NotchedStrengthofCompositeMaterials
185
ECSSEHB3220Part2A
20March2011
[148] R.B.Pipes,J.W.Gillespie,Jr.,&R.V.Wetherhold
SuperpositionoftheNotchedStrengthofCompositeLaminates
PolymerEngineeringandScience,Vol.19,No.16,1979
[149] R.B.Pipes,R.V.Wetherhold&J.W.Gillespie,Jr.
MacroscopicFractureofFibrousComposites
MaterialsScienceandEngineering,Vol.45,1980
[1410] K.Y.Lin
FractureofFilamentaryCompositeMaterials
Ph.D.Dissertation,DepartmentofAeronauticsandAstronautics
M.I.T.,Cambridge,Mass.,January1976
[1411] J.W.Mar&K.Y.Lin
FractureMechanicsCorrelationforTensileFailureofFilamentary
CompositeswithHoles
J.ofAircraft,Vol.14,No.7,July1977
[1412] C.C.Poe,Jr.&J.A.Sova
FractureToughnessofBoron/AluminiumLaminateswithVarious
Proportionsof0and45Plies
NASATechnicalPaper1707,November1980
[1413] O.L.Bowie
AnalysisofanInfinitePlateContainingRadialCracksOriginating
fromtheBoundaryofanInternalCircularHole
JournalofMathematicsandPhysics,Vol.35,1956
[1414] G.R.Irwin
FractureDynamics
FractureofMetals,A.S.M.,Cleveland,1948
[1415] S.G.Lekhnitski
AnisotropicPlates
SciencePublishers,NewYork,1968
[1416] R.E.Peterson
StressConcentrationFactors
JohnWiley&Sons,NewYork,1974
[1417] W.F.Brown,Jr.&J.E.Srawley
PlaneStrainCrackToughnessTestingofHighStrengthMetallic
Materials
ASTMSTP410,AmericanSocietyofTestingandMaterials,1966
[1418] C.S.Hong&J.H.CrewJr.
StressConcentrationFactorsforFiniteOrthotropicLaminateswitha
CircularHoleandUniaxialLoading
NASATP1469,May1979
[1419] S.G.Lekhnitski
TheoryofElasticityofAnisotropicElasticBody
186
ECSSEHB3220Part2A
20March2011
HoldenDay,Inc.(SanFrancisco),1963
[1420] MBB/ERNOFinalReport
IntegrityControlofCarbonFibreReinforcedPlasticsStructural
Elements
ESAContract4442/80/NL/AK(SC),1985
[1421] S.P.Garbo&J.M.Ogonowski
StrengthPredictionsofCompositeLaminateswithUnloadedFastener
Holes
AIAAJournal,Vol.18,No5,May1980
[1422] L.B.Greszczuk
StressConcentrationsandFailureCriteriaforOrthotropicand
AnisotropicPlateswithCircularOpenings
ASTMSTP497,1972
[1423] ESDUDataSheets,85001,Nov.1985
[1424] R.H.Martin&B.D.Davidson
ModeIIfracturetoughnessevaluationusingafourpointbendend
notchedflexuretest
4thInternationalConferenceonDeformationandFractureof
Composites,March2426,1997,Manchester
[1425] K.N.Shivakumar,P.W.Tan&J.C.NewmanJr.
Avirtualcrackclosuretechniqueforcalculatingstressintensity
factorsforcrackedthreedimensionalbodies
InternationalJournalofFracture,Vol.36,1988,pp.R43R50
[1426] C.T.Sun&C.J.Jih
Onthestrainenergyreleaseratesforinterfacialcracksinbimaterial
media
EngineeringFractureMechanics,28:1327,1987
[1427] R.H.Martin,P.Sriram&S.J.Hooper
Usingamixedmodefatiguedelaminationcriterion
Presentedatthe12thASTMSymposiumonCompositeMaterials:
TestingandDesign,Montreal,Canada,May1994
[1428] B.D.Davidson
Predictionofenergyreleaserateforedgedelaminationusingacrack
tipelement
CompositeMaterials:FatigueandFracture,(FifthVolume),ASTM
STP1230,R.H.Martin,Ed.,ASTM,Philadelphia,1995,pp.155175
[1429] R.H.Martin
Localfracturemechanicsanalysisofstringerpulloffand
delaminationinapostbuckledcompressionpanel
ProceedingsoftheICCM10,Whistler,BritishColumbia,August1418
1995VI253
187
ECSSEHB3220Part2A
20March2011
[1430] T.K.OBrien
Mixedmodestrainenergyreleaserateeffectsonedgedelamination
ofcomposites
EffectsofDefectsinCompositeMaterials,ASTMSTP836,D.J.
Wilkins,Ed.,ASTM,Philadelphia,1984,pp.125142
[1431] G.B.Murri,S.A.Salpekar,andT.K.OBrien
Fatiguedelaminationonsetpredictionintaperedcomposite
laminates
CompositeMaterials:FatigueandFracture(3rdVolume),ASTMSTP
1110,T.K.OBrien,Ed.,ASTM,Philadelphia,1991,pp.312339
[1432] R.H.Martin&W.C.Jackson
Damagepredictionincrosspliedcurvedcompositelaminates
CompositeMaterials:FatigueandFracture,FourthVolume,ASTM
STP1156,W.W.Stinchcomb,N.E.Ashbaugh,Eds.,ASTM,
Philadelphia,1993,pp.105126.
188
ECSSEHB3220Part2A
20March2011
15
Prediction of dynamic characteristics
15.1 Introduction
The important distinction between material and structural damping is often not appreciated and
consequent misinterpretation of data results in errors. It is therefore necessary to clarify this matter
andshowthecorrectapplicationofpredictionmethodsfordamping,[See:15.3;15.4and15.5].
Definitionsofthemostwidelyuseddampingexpressionsandtheirinterrelationships,Ref.[155],are
alsogiven,[See:15.2].
ArigorousanalyticaltreatmentofthetheoryofdampingisgiveninRef.[151],[152],[153].
15.2 Definition of damping terms
15.2.1 General terms
15.2.1.1 Specific energy

Dampingcanbeexpressedquantitativelyoneitheranabsoluteorarelativeenergybasis.Inabsolute
terms,thespecificenergyUistheenergydissipatedperunitvolumeofmaterialduringonecycleof
vibrationatagivenstressamplitude.
Thebehaviourofmanymaterialscanberepresentedapproximatelybytherelationship:
U J n
[15.2-1]
Where: Jandnarematerialconstants,Ref.[152].
Theenergydissipatedbydampingisrepresentedbytheareawithinthestress/strainhysteresisloop.
189
ECSSEHB3220Part2A
20March2011
15.2.1.2 Loss factors
Awellestablishedfundamentalratiowithwhichtoassessdampingonarelativeenergybasisisthe
lossfactor,where:
1 Energy dissipated during one cycle

2 Maximum strain energy stored during cycle
1 U
[15.2-2]
2 U
Forauniformblockofmaterialsubjectedtoauniaxialsimpleharmonicstressofamplitude,misthe
materiallossfactor,where:
J JE
n n2
1
m 2 [15.2-3]
2E
2
Ifthematerialhaslinearmaterialdamping,i.e.,n=2,misconstant:
JE
[15.2-4]
m

Forthegeneralsituation,amemberhavinganonuniformstressdistribution,thestructurallossfactor
isdeterminedby:
JE V dv
V dv
n
V
[15.2-5]

s 2
V
Or,intermsofstrain:
JE n1 V dv n
V [15.2-6]
V dv
s 2
V
And,inshear:
J 1 G1n 1 V dv n
V [15.2-7]
V dv
s 2
V
Therefore, loss factors determined from experiments on specimens having a nonuniform stress
distributionsarestructurallossfactors.
Ifthematerialhaslineardamping,i.e.n=2,thestructurallossfactorisequaltothemateriallossfactor
andisindependentofstructuralgeometryorstressdistribution.
190
ECSSEHB3220Part2A
20March2011
15.2.2 Complex modulus model
15.2.2.1 General
Experimentalevidencesuggeststhatthedampingofcompositematerialsislinearatsmallamplitudes.
Althoughthisapproachhasnumerousdeficiencies,Ref.[153],themostwidelyusedmodeloflinear
dampinginstructuraldynamicsisthecomplexmodulusmodel.Thismodel,whichhastheadvantage
of incorporating stiffness and damping properties in a complex modulus, is used throughout this
review,Ref.[154].
15.2.2.2 Energy dissipation under steady state sinusoidal excitation

Thisisusuallydefinedintermsofthespecificdampingcapacity,whichistheratiooftheenergy
dissipated,Ud,inaunitvolumepercycletothemaximumstrainenergystoredperunitvolume,U,
i.e.:
U 2 d
[15.28]
U
15.2.2.3 Bandwidth of half power points in steady state sinusoidal excitation

Onemethodofspecifyingthisisintermsofthedimensionlessdampingratio,,whichisdefinedin
termsoftheratiooftheactualtocriticaldampingcoefficients,Ref.[151].canbecalculatedfrom:

2 1
[15.29]
2 n
where1and2arethefrequenciesatwhichtheenergydissipatedinthesystemisonehalfofthatat
resonancen.
15.2.2.4 Quality factor

Q 2 1 [15.2-10]
15.2.2.5 Loss tangent under sinusoidal excitation

Applyingthecomplexstiffnessapproachtothematerialstiffness(elasticmodulus):
E E iE E 1 ig
R I R [15.2-11]
WhereEisthecomplexmodulus,ERandEI,aretherealandimaginarycomponentsofErespectively
andgisthelosstangent:
gE R

E I
WhereERandEIareusuallycalledstorageandlossmodulirespectively.
191
ECSSEHB3220Part2A
20March2011
15.2.2.6 Decay of free vibration
Themostcommonmeasureofthisphenomenonisthelogarithmicdecrement,definedas:
ln a i
[15.2-12]
a 1
i
Whereaiistheamplitudeoftheithcycle,iisaninteger.
Formostmaterials,atleastatrelativelysmallamplitudes,issmallandindependentofamplitude.A
morepracticalmethodofcalculatingfromexperimentaldataisthen:
1 a
ln n
[15.213]
n a 1 n
wherenisanarbitraryinteger.
15.2.2.7 Interrelationships
The interrelationships between the parameters can be assumed for low damping complex modulus
materials:
1
g 2 2 [15.214]
Q
15.3 Prediction methods for damping

Ifcompositesaretobeusedinprimaryloadbearingstructures,itisessentialthattheirresponsesto
static and dynamic loads are predicted accurately. Such predictions involve the use of not only
stiffnessandstrengthpropertiesbutalsodampingcharacteristics.Indeed,dampingistheparameter
whichgovernsresonantresponse,acrucialconsiderationinstructuressubjectedtoacousticloads.
Some simple first approximation theories can be applied to predict the damping of a composite
materialfromthedynamicpropertiesofitsconstituentsandtheirconfiguration.Suchtheories,[See:
15.4],arebasedontheassumptionsthat,Ref.[155]:
Nosignificantmacroscopicorinterfacialslipoccurs.
Fibres and matrix are characterised by their elastic behaviour and associated loss coefficients,
Ref.[154].
[Seealso:15.2foracompilationofdefinitionsofdampingterms;15.5fordataondampingproperties]
192
ECSSEHB3220Part2A
20March2011
15.4 Determination of damping characteristics
15.4.1 Unidirectional characteristics

AccordingtoHashin,Ref.[156]:
Em
c mV m [15.4-1]
E c
Where:
c=flexurallossfactorofthecompositeinfibredirection
m=flexurallossfactorofthematrix
Vm=volumefractionofthematrix
Em=modulusofthematrix
Ec=compositemodulus;determinedusingthelawofmixtures
15.4.2 Off axis characteristics

Thisanalysisisbasedonthestrainenergyintheply,Ref.[158].Theenergydissipatedintheplyis
calculated from the compliance matrix of the ply, the stress in the ply, and the specific damping
capacitiesoftheplyrelativetothefibreaxis,asfunctionsofstress.Thiscomplexmethod,involving
extensivenumericalintegration,canbeavoidedbyassuminglineardamping.
The complex moduli of the ply relative to the fibre direction can be substituted into modulus
transformationequationsforaplyofarbitraryorientation,Ref.[159],providingplycomplexmoduli,
fromwhichthedampingcanthenbecalculated,i.e.:
1
2 12 sin 2 cos 2 1 sin 4

1 1
cos 4
[15.4-2]

E1

E x E1 G12 E2
1

1 1
sin 4
2 12 1
sin 2 cos 2 cos 4 [15.4-3]

E y E1 G12 E1 E2
2 4 12 sin 2 cos 2 sin 4 cos 4

2
1
1

2

1
[15.4-4]
G xy E1 E 2 E1 G12 G12
1

2 1 2
xy

4 4
2
E x 12 sin cos sin cos
2
[15.4-5]
E1 E1 E 2 G12 G12
Where:
193
ECSSEHB3220Part2A
20March2011
E , E , G ,

x y xy xy
arethelongitudinal,transverseandshearcomplexmoduli
andPoissonsratiorelativetotheplyaxis
E , E , G ,

1 2 12 12
arethelongitudinaltransverseandshearcomplexmoduli
andPoissonsratiorelativetothefibreaxis
15.4.3 Laminate characteristics

Adams and Bacon, Ref. [158], calculated the damping of an FRP laminate by summing the energy
dissipatedintheindividualply,[See:UnidirectionalcharacteristicsandOffaxischaracteristics].
Thespecificdampingcapacityofthelaminateinaspecifiedmodeisthencalculatedbydividingthe
totalenergydissipatedbythatstored.Againthecomplexityofthenumericalintegrationmakesthis
methodunattractive.
By assuming linear damping, Schultz and Tsai, Ref. [1510], produced a more attractive method of
predicting the damping of CFRP laminates. Essentially, the complex moduli of the laminate are
substituted directly into the welldocumented transformation equations for laminated composites
providing the complex moduli of the laminate. Consequently, the shear or flexural damping of the
laminatesisfoundfromtherealandimaginarypartsoftherelevantcomplexmoduli.
A computer program, developed by ESDU, can be used to predict the damping of laminates, [See:
16.29].
15.5 Approximate data on damping

To verify the dynamic behaviour of a composite structure at the preliminary design stage, the
approximatedataonlossfactorspresentedinFigure15.5.1cangiveinitialguidance,Ref.[155].
[Seealso:Referencesformoredetaileddataondampingproperties]
Figure 15.5-1 - Summary of loss factors for various fibres / epoxy composites
194
ECSSEHB3220Part2A
20March2011
15.6 References
15.6.1 General
[151] C.W.Bert
MaterialDamping:AnintroductoryReviewofMathematicalModels,
measuresandexperimentalTechniques
J.SoundVib.29(2),1973
[152] B.J.Lazan
DampingofMaterialsandMembersinStructuralMechanics
PergamonPress,1968
[153] R.E.D.Bishop
TheTreatmentofDampingForcesinVibrationTheory
J.RoyalAero.Soc.59,1955
[154] C.W.Bert
CompositeMaterials:ASurveyoftheDampingCapacityofFibre
ReinforcedComposites
DampingApplicationsforVibrationControl
Proc.WinterAnnualMeeting,1980
[155] T.A.Palmer
ADesignGuidetotheDampingofCompositeMaterials
SALReport1046forESTEC,April1985
[156] Z.Hashin
ComplexModuliofViscoelasticComposites:IIFibreReinforced
Materials
Int.J.SolidStructures,6,1970
[157] H.L.Cox
TheElasticityandStrengthofPaperandotherFibrousMaterials
Brit.J.Appl.Phys.13,1952
[158] R.D.Adams&D.G.C.Bacon
EffectofFibreOrientationandLaminateGeometryontheDynamic
PropertiesofCFRP
J.CompositeMaterials,7,1973
[159] R.M.Jones
MechanicsofCompositeMaterials
McGrawHill,1975
[1510] A.B.Schultz&S.W.Tsai
MeasurementsofComplexDynamicModuliofLaminatedFibre
ReinforcedComposites
J.CompositeMaterials3,1969
195
ECSSEHB3220Part2A
20March2011
16
Computer analysis of composites
16.1 Introduction
Anoverviewofcommonlyavailablesoftwareproductsispresented.
NOTE The software packages described are not a comprehensive list of
thoseavailable.Potentialusersareadvisedtocontactsuppliersfor
particulardetailsoftheirsoftware.
Theprogramsaredividedintothreedifferenttypes,[See:16.2]:
Finiteelementprogramswithspecialcompositefeaturesincluded.
Laminateanalysisprogramsperforminglaminatedplatetheory.
Programsforspecialapplications.
16.2 Computer programs: Analysis of composites
16.2.1 General
Table16.2.1listssomecommonlyavailablesoftwarepackages,alongwithlinkstosupplierswebsites.
Thesoftwareislistedbyprogramtype.
Asummarydescriptionforeachpackageisgivenfor:
FiniteelementinTable16.2.2.
LaminateanalysisinTable16.2.3.
SpecialapplicationsinTable16.2.4.
[See:16.3forESDU;16.39forESAComp]
196
ECSSEHB3220Part2A
20March2011
Table16.21Commonlyavailablesoftwarefortheanalysisofcomposites
Program Type
Program Name Supplier (web site link) Platform Finite Laminate Special
Element Analysis Application
Abaqus Inc.
ABAQUS
[Hibbit Karlsson and Sorenson Inc.]
ACOMAP C.A.S.A Space Division
ANSYS Inc.
ANSYS
[Swanson Analysis System Inc.]
ARGUS Merlin Technologies Inc.
BEAM
Stanford University *
BOLT
CLASS Materials Science Corp.: ASM Intl.
COALA Cranfield Institute of Technology
COMPOS Technical University of Linkping
COSMOS/M Structural Research and Analysis Corp.
CURE Stanford University *
PC-
ESAComp Windows;
Componeering Inc
[See: 16.39] Linux;
Unix
ESDU A7916

A8039
ESDU A7303

A8147
ESDU Engineering Data Service
ESDU A8335
[See: 16.03]
ESDU A8418
ESDU A8501
ESDU A8336
ESDU A8512
GENLAM
S.Tsai; Think Composites
LAMRANK
ICAN COSMIC , University of Georgia
IMPACT Stanford University *
LAMA Dornier System
MSC.NASTRAN MSC Software
NISA Engineering Mechanics Research Corp.
SAMCEF SAMTECH S.A; CISI Engineering
SAP Computer Software Inc.
STRESS
W8GAIN Stanford University *
WIND
197
ECSSEHB3220Part2A
20March2011
16.2.2 Finite element programs
Table16.22Computerprogramsforfiniteelementanalysis
Name Application Composite Features Optimise
* General purpose finite element * Anisotropic material models in all elements,
program. including non linearity.
ABAQUS
* Linear, non-linear and coupled * 3-D solid elements.
analysis.
* General purpose finite element * 2-D and 3-D plate/shell and solid elements
stress and heat transfer analysis with orthotropic and anisotropic material
program. properties.
* Non-linear capabilities, but not * Laminated shell elements.
ANSYS
generally implemented for composite * Crack-tip solid elements.
materials. * Thick-shell elements.
* Pre- and post-processing. * Max strain, max stress, Tsai-Wu and user
defined failure criteria.
program. with orthotropic temperature dependent
* Extensive capabilities for non- properties.
linear analysis and for composite * Layered composite plate/shell and doubly
ARGUS materials. curved elements
* Buckling load and mode processor. * Shell element with anisotropic temperature
* Modal analysis processor. dependent properties.
* Pre- and post-processing * Biaxial max. stress and max. shear stress
failure criteria
program. with orthotropic temperature-dependent
* Linear/non-linear static, properties.
linear/non-linear dynamic analysis. * Layered composite shell elements.
COSMOS/M
* Buckling and heat transfer * Tsai-Hill and mod. Tsai-Wu failure criteria.
analysis.
* 1-D, 2-D and 3-D structural and
thermal models.
* General purpose finite element * Plate/shell elements with orthotropic
program. temperature-dependent properties (input via
* Static, dynamic and PCOMP, MAT8 card).
MSC
thermodynamic analysis. * Ply stresses.
.NASTRAN
* Non-linear analysis. * Tsai-Hill, Hoffmann and Tsai-Wu failure
* Pre- and post-processing by criteria
MSC.PATRAN
* General purpose finite element * Solid elements.
program. * Thick-shell elements (material properties may
* Pre- post-processing (digitiser change within an element).
NISA input of nodal data for 2-D and 3-D * Laminated shell elements.
structures). * In-plane and interlaminar stresses.
* Max. stress, max. strain, mod. Hill, von Mises,
Tsai-Wu and delamination failure criteria
* General purpose finite element * 2-D and 3-D plate/shell elements
program * Thick shell elements
* Linear/non-linear stat and dyn. * Multilayer solid elements
* Thermal analysis * Sandwich modelling
SAMCEF
* Buckling * Transverse shear deformation
* Optimisation * In-plane and interlaminar stresses
* Pre-, post-processing * Max. stress, max. strain, Tsai-Hill Tsai-Wu
and Sandhu failure criteria
* General purpose finite element * Only SAP7 has 2-D and 3-D elements with
program orthotropic material properties.
SAP
* SAP7 has non-linear capability. * Multi-layer plate/shell elements.
* Pre- and post-processing. * Tsai-Hill and Tsai-Wu failure criteria
Optimise:softwarewithoptimisationroutinesforlaminateelementstobeadded.
198
ECSSEHB3220Part2A
20March2011
16.2.3 Laminate analysis programs
Table16.23Computerprogramsforlaminateanalysis
Name Application Special Features
* Interactive input/output
* Matrix/fibre data: ply properties; hybrid laminates.
* Stiffness and compliance matrix.
* Engineering constants (membrane)
Laminate analysis program based on
ACOMAP * Mechanical and thermal loads.
laminated plate theory
* Ply stresses in ply co-ordinates.
* Ply stresses and strains in laminate co-ordinates.
* Max. stress, max. strain and Tsai-Hill failure criteria.
* Last ply failure.
* Matrix/fibre data: ply properties
* Stiffness and compliance matrix engineering constants
CLASS * Thermal expansion coefficient
* Ply stresses and strains
* First ply/last ply failure
* Input/output files.
* Ply properties; hybrid laminates.
* Reduced input for symmetric stacking.
* Stiffness and compliance matrix
* Engineering constants (membrane and bending).
* Mechanical and hygrothermal loads.
Laminate analysis program based on * Variation of hygrothermal loads across the laminate.
COALA
laminated plate theory * Ply stresses and strains in ply co-ordinates.
* Ply stresses in laminate co-ordinates.
* Stresses and strains at top and bottom of each ply.
* Max. stress, max strain, Hoffmann, Tsai-Hill and Tsai-Wu
failure criteria
* First ply failure.
* Index of margin of safety.
* Ply properties (material data base).
* Stiffness and compliance matrix
* Engineering constants (membrane anti-bonding).
COMPOS * Mechanical and hygrothermal loads.
* Ply stresses and strains in ply and laminate co-ordinates.
* Stresses/strains at top and bottom of each ply.
* Tsai-Hill and Tsai-Wu failure criteria
* First ply failure.
* Fiber/matrix micromechanics
* Analyses for constitutive and thermal/moisture expansion
behavior of plies
* Laminates
- 2.5D behavior - classical lamination theory (CLT)
- laminate strength in principal loading conditions
- load response - laminate and layer level response, effects due to
thermal/moisture loads
- first ply failure (FPF) and degraded laminate failure (DLF)
Analysis and design of composite
- several commonly used failure criteria and a possibility to add
laminates and laminated structural
ESAComp user specified criteria
elements
- wrinkling of sandwich face sheets, core shear failure
- interlaminar shear failure
- constant and variable load approach in determining margins of
safety
- failure and design envelopes
- sensitivity studies for 2.5D behavior and FPF analyses -
tolerances for a ply property or layer orientations
- notched laminate analysis of circular and elliptic holes
- layer drop-off - exterior or embedded drop-off in a solid
199
ECSSEHB3220Part2A
20March2011
Name Application Special Features

laminate or a face sheet of a sandwich
- straight free edge analysis using built-in FE solver
- laminate through-the-thickness temperature distribution
- laminate through-the-thickness moisture distribution and
ESAComp moisture content as a function of time
(continued) * Plates and stiffened plates
- Mindlin plate analysis using built-in FE solver
- rectangular plates with any combination of clamped, simply
supported, and free edges
- load response and failure due to transverse loads
(combinations of a point load, line loads, and distributed load),
- stability and failure under in-plane loads, natural frequencies
* Beams and columns
* Bonded joints, [See: Table 16.02.4]
* Mechanical joints, [See: Table 16.02.4]
Note: Other features include: design tools; FE import/export
interfaces; batch run capability (integration with other software);
databank.
* Ply properties.
ESDUpac
Stiffnesses of laminated plates * Stiffness and compliance matrix.
A8335
* Engineering constants.
* Ply properties; hybrid laminates
* Mechanical and thermal loads.
ESDUpac Failure of composites based on
* Damage history up to failure.
A8418 laminated plate theory
* Damage caused by constrained unbalanced laminates.
* Max stress and mod. Puck failure criteria.
* Fibre/matrix and ply properties input from database
* Hybrid laminates
* Hydrothermal analysis
* Stress concentration around holes
Micro-mechanic, macro-mechanic
ICAN * Delamination locations around holes
and laminate analysis
* Free edge stresses
* Transverse shear stresses
* Failure loads based on max. stress
* First ply failure by fibre breakage
* Ply properties (material data base); hybrid laminates.
* Stiffness and compliance matrix.
* Engineering constants (membrane and bending).
Laminate analysis program and
GENLAM/ * Ply stresses and strains in ply and laminate co-ordinates.
laminate ranking based on laminated
LAMRANK * Stresses/strains at top and bottom of each ply.
plate theory
* Tsai-Hill and Tsai-Wu failure criteria
* First ply failure/ last ply failure.
* Indication of critical load case.
* LAMRANK performs a ranking of all possible lay-up for a given
load case.
* Interactive input/output ply properties (material data base).
* Stiffness matrix (also 3-D).
* Engineering constants (membrane).
* Stress concentration factor for circular holes.
* Optimised stacking sequence with respect to free edge
LAMA Laminate analysis program based on delamination.
(continued) laminated plate theory * Equivalent plate model for symmetric sandwiches.
* Ply stresses and strains in ply and laminate co-ordinates.
* Stresses/strains at top and bottom of each ply.
* Hashin, Tsai-Hill and Tsai-Wu failure criteria
* First ply failure/last ply failure.
200
ECSSEHB3220Part2A
20March2011
16.2.4 Special applications programs
Table16.24Computerprogramsforspecialcompositeapplications
Name Application Features
Laminates, plates and stiffened
* Thermal and moisture distribution
panels, beams and columns.
predictions.
[See also: Table 16.02.3].
* Bonded joints: single and double lap; single
and double strap; single and double sided
scalfed lap, bonded doubler. Beam and
plate models for adherends, linear and
nonlinear adhesive models. Combinations
of axial, bending, in-plane and out-of- plane
ESAComp shear loads. Joint deflection forces and
* Bonded joints moments in adherends, adhesive stresses,
* Mechanical joints margins of safety for cohesive failure of
adhesive and laminate failure due to
in-plane and bending loads.
* Mechanical joints: single and double lap
joints under axial loads. Fastener and by-
pass loads, laminate stresses and strains on
fastener holes. Margins of safety for
failure. Prediction of failure mode.
* Tensile or shear stresses in the adherends.
* Inelastic shear stresses or strains in the
adhesive of a balanced bonded single or
ESDUpac multi-step lap joint.
A7916 Strength of bonded joints * The adherends may have different material
A8039 properties.
* Program ESDU 2020 does not consider the
stresses in the adherend, but the peel
stresses.
* Buckling problems for laminated plates and
ESDUpac sandwich panels with composite facings.
Buckling of rectangular
A7303 * Clamped and simply supported edges.
orthotropic plates
A8147 * Compression and shear loads (also in
combination).
* Stress/strain distribution around holes in
ESDUpac Stress concentration in composite composite plates.
A8501 plates * Tensile, compression or shear loads.
* Ply properties or laminate stiffnesses.
* Lower natural frequencies of initially
unstressed, specially orthotropic,
ESDUpac Natural frequencies of laminated
rectangular, flat or slightly curved plates.
A8336 plates.
* Any combination of simply supported or
clamped edges possible.
* Material damping in rectangular, multi-
layered composites vibrating in one of their
lower frequency natural modes.
ESDUpac * Only balanced symmetric laminated plates,
Damping of composites
A8512 simply supported or all edges clamped.
* Output for the specific damping capacity and
viscous damping ratio for each specified
mode.
201
ECSSEHB3220Part2A
20March2011
Name Application Features

* Calculation of forces and moments induced
on a composite beam by external loads.
* Different types of loads and supports.
* Output for deflection curvature.
BEAM Design of composite beams.
* Stresses/strains.
* Margin of safety for the plies.
* Only symmetric laminates or sandwich with
equal face sheets.
* Prediction of failure and failure mode.
* Three types of bolted joints:
1) joints with a single hole
Design of pin-loaded holes in
BOLT 2) joints with two identical holes in a row
composites.
3) joints with two identical holes in tandem.
* Only uniform tensile loads and symmetric
laminates.
* Calculation of optimal cure cycle.
* Influence of material properties and process
Analysis of curing process of
CURE variables on cure cycle.
composites.
* Unsymmetric but no hybrid laminates
possible.
* Calculation of delamination area induced in
a composite plate by a foreign object impact.
Calculation of impact damage
IMPACT * Variable characteristics of impactor.
for composite plates.
* Stress history using transient finite element
formulated for a 3-D element mesh.
Calculation of hygrothermal
STRESS * Only hygrothermal loads.
stress/strain in composites.
* Calculation of:
- cylinder temperature,
- viscosity,
Analysis of filament winding of
WIND - degree of cure,
composites.
- fibre position,
- fibre tension,
- residual stresses/strains after curing.
* Moisture content as a function of
Calculation of moisture content in temperature and time.
W8GAIN
a composite. * Heat conduction and moisture diffusion are
described by Fourier and Fick formulae.
202
ECSSEHB3220Part2A
20March2011
16.3 ESDU data for composite analysis
16.3.1 General
ESDUInternationalplc(UK)producesdesigndataandinformationforengineersworkinginarange
of engineering fields. A number of these documents, referred to as Data Items, have particular
applicabilitytodesignwithcomposites.
AlistoftheindividualDataItemsisgiveninTable16.3.1,whichincludesalinktoasynopsisofeach
one,alongwithdetailsofrelevantESDUpacs.
16.3.2 ESDU data items

IndevelopingESDUDataItems,afulltechnicalassessmentismadeoftheavailabledatabothinterms
of the reliability of experimental techniques and by comparison with theoretical predictions. The
qualified staff of ESDU work with the guidance and assistance of committees whose members are
drawnfromindustry,researchandeducationalestablishments,andincludetypicalusersofthedata.
By this means, the accuracy of the work can be assured, the relevant and practical experience of a
numberofindividualscanbeincludedtosupplementthequantitativeresultsandtheformofthefinal
presentationcanbechosentosuitthetypicalusers.
203
ECSSEHB3220Part2A
20March2011
Table16.31ESDUdatasheetsforcompositeanalysis
SMH ESDU No.
Title
Topic
Laminated Composites - General analysis
16.7 Laminate stress analysis 74039
16.8 Plate stiffnesses (in-plane) 75002
16.16 Lay-up arrangements for special orthotropy 82013
16.19 Plate stiffnesses and apparent elastic properties 83035
16.24 Stresses and strains around circular holes 85001
16.27 Selection of reinforcement around circular holes 86003
16.34 Plate through-the-thickness shear stiffnesses 89013
Buckling of balanced laminated composites
16.4 Buckling of orthotropic plates 73003
16.5 Flexural stiffness of flat strips 73007
16.6 Metallic skin stiffeners reinforced by composite - local buckling 73015
16.12 Buckling of specially orthotropic composite plates 80023
16.15 Buckling of orthotropic plates 81047
16.31 Buckling of curved composite panels 87025
Buckling of unbalanced laminated composites
16.28 Buckling of unbalanced composite plates 86020
Sandwich panels with composite face plates
16.30 Sandwich column and beam face plate wrinkling 87013
16.32 Sandwich panel face plate wrinkling 88015
Bonded joints
16.9 Bonded joints - 1 78042
Failure criteria
16.17 Failure modes of laminated composites 82025
16.18 Failure criteria for layers of a laminated composite 83014
16.22 Failure analysis 84018
16.36 Delamination and free edge stresses 90021
16.37 Delamination at termination of plies 91003
Natural modes of vibration
16.20 Natural frequencies of laminated flat plates 83036
16.26 Sandwich panel natural frequencies 85037
16.33 Vibration of singly-curved laminated plates 89011
16.35 Vibration of plates with in-plane loading 90016
Damping and response to acoustic loading
16.21 Strains in skin panels under acoustic loading 84008
16.23 Endurance under acoustic loading 84027
16.25 Damping in composite plates 85012
16.29 Sandwich panel response to acoustic loading 86024
204
ECSSEHB3220Part2A
20March2011
16.3.3 ESDUpac
Manycalculationsforfibrereinforcedlaminatedstructuresaremoreconvenientlymadewiththeaid
ofcomputerprograms.
[See:ESDUInternationalplcwebsiteforfurtherdetails]
16.4 Buckling of orthotropic plates
16.4.1 Title
Bucklingofflatrectangularorthotropicplates.
DataItemNo.73003[Seealso:ESDUInternationalwebsite].
16.4.2 Usage and scope

Laminatedcompositeplatesandsandwichpanelsfacedwithlaminatedcompositescanbemodelled
as homogeneous orthotropic plates. For such plates overall buckling is an important condition.
Buckling is a function of the plates inplane and flexural stiffnesses, its throughthethickness shear
stiffnesses,itsshape,boundaryconditionsandtheappliedloading.Thethroughthethicknessshear
stiffnessesareofparticularimportancesince,iftheyareverylow,theycandominatebuckling.
16.4.3 Analysis
Theplateistreatedashomogeneous,andcanbeisotropicorspeciallyorthotropic.Thethroughthe
thickness shear stiffnesses in the principal plane of orthotropy and orthogonally to that plane are
included. The elastic buckling load is derived for symmetrical combinations of clamped and simply
supported edge conditions for rectangular plate subject to combinations of inplane compression,
tension and shear loading. The solution is presented as a computer program. Buckle wave forms
havingupto7halfwavesalongoracrosstheplateareexamined.
16.4.4 ESDUpac A7303
16.4.4.1 General
PreviouslyESDUpacE1001
Theprogramcalculatestheelasticoverallbucklingloadsfororthotropicrectangularplates.
16.4.4.2 Input
Boundaryconditions
Alledgesclamped.
Alledgessimplysupported.
Symmetricalcombinationsofclampedandsimplysupportededges.
Loading(canbetensile)
205
ECSSEHB3220Part2A
20March2011
Giventwoofthethreeinplaneloadsdeterminetheremainingcauseofbuckling.
Giventheratiooftheappliedloadsdeterminealltheirvaluesatbuckling.
Geometryandstiffnesses.
Platepropertiesandgeometry
Aspectratio.
Parametersderivedfromflexuralandshearstiffnessesofplate.Poissonsratio.
16.4.4.3 Output
Buckling loads or ratios of loads and numbers of buckle half wave examined. Statement if
appliedloadscausebucklingalone.
16.4.5 Notes
Theaccuracyofthecalculationsisopentovariationwiththenumberofiterationsintherootsolving
routines.
[See:16.3foraccesstoprogramsanddocumentation]
16.5 Flexural stiffness of flat strips
16.5.1 Title
Stiffnessofloadedflatstripsundersinusoidallydistributedbendingcouplesattheiredges(forusein
localbucklingcalculations).

The behaviour of isotropic or specially orthotropic flat strips subject to sinusoidally distributed
couplesalongtheiredgescanbeusedtoprovidethelocalbucklingstressofmostcommonstructural
engineeringsections.AnexampleoftheiruseappearsinDataItemNo.73015,[See:16.6].
16.5.3 Analysis
The elastic behaviour of thin strips of specially orthotropic material has been analysed to produce
graphswhichplottherotationalstiffnessattheedgesforarangeofstiffnessvaluesandarangeofhalf
wavelengthsoftheappliedsinusoidalcouples.Stripswithonefreeedgeandacoupleactingonthe
otheredgeorstripswithcouplesalongbothedgeseitheropposedorinthesamesense,areanalysed.
Graphsarealsoprovidedforcalculationofthenecessarystiffnessvalues.
[See:16.3foraccesstodocumentation]
206
ECSSEHB3220Part2A
20March2011
16.6 Metallic skin stiffeners reinforced by composite -

local buckling
16.6.1 Title
Estimation of the local buckling stress under biaxial compression of an isotropic skin with fibre
reinforcedintegralunflangedstiffeners.

Laminatedfibrereinforcedcompositestripscanbebondedtostiffenersinmetallic(isotropicmaterial)
structures to increase their stiffness. The compressive strength of such structures can be limited by
local(shortwavelength)bucklingofthestiffenerscoupledtothestructurethattheystiffen.Wherethe
structure is effectively composed of flat strips, as for example a sheet reinforced by Zsection
stiffeners,thestiffnessunderoverallcompressionwithdestabilisingsinusoidallydistributedcouples
acting along the junctions of the strips can be calculated. This is an unstable condition and the
compressive end load where the couples balance and instability, or buckling, occurs without the
necessity of externally applying destabilising couples can be found and this represents a buckling
condition.Thecalculationcanberepeatedforarangeofwavelengthsandthewavelengthyieldingthe
lowestbucklingloadbeascertained.
16.6.3 Analysis and data

An example of calculation using an integrally machined panel with simple unflanged stiffeners
reinforcedbycarbonfibrebondedtothestiffenersisworkedthroughindetailandthesearchforthe
minimumbucklingloadissetoutintabularform.
ThestiffnessesrequiredareobtainedfromthegraphsofDataItemNo.73007,[See:16.5].Themethod
involves setting up a simple equation relating the stiffnesses of the strips forming the structure and
theformulationofthegoverningequationisexplained.
207
ECSSEHB3220Part2A
20March2011
16.7 Laminate stress analysis
16.7.1 Title
Stressanalysisoflaminatedflatplates.

This item presents the formulae for the elastic analysis of flat laminated plates built up from thin
orthotropic layers which can have different material elastic and thermal properties, thicknesses and
orientationsoftheirprincipalaxesoforthotropy.Stressesinlaminatedplatesariseeitherasaresultof
externally applied loading or as a result of the method of manufacture. Current methods of
productionoflaminatedplatesusuallyrequiretheassemblytobecuredatatemperatureabovethe
working temperature range of the plate. Upon cooling to working temperatures, the different
orientationsofthelayerstogetherwiththeiroftendifferentelasticandthermalpropertiesleadtothe
developmentofasystemofselfequilibratingstresses.Similarly,internalstressescanbegeneratedin
platescuredatroomtemperaturebutwhicharesubjectedtoawiderangeofworkingtemperatures.
16.7.3 Analysis
Theformulaepresentedrelatetouniforminplanestresssystemsawayfromtheplateboundariesor
othergeometricdiscontinuities.Noallowanceismadeforthedeformationsassociatedwiththrough
thethickness shear stiffnesses. The formulae deal with eight cases and provide the stressstrain
transformationrelationships:
Balancedlaminates Unbalancedlaminates
Externallyappliedloads(nobuiltin Externallyappliedloads(nobuiltin
stresses) stresses)
Builtinstresses(noexternallyapplied Builtinstresses(noexternallyapplied
loads) loads)
Externallyappliedloadswithbuiltin Externallyappliedloadswithbuiltin
stresses stresses
Plateheldflat(nootherexternallyapplied
loads)
Plateheldflatandexternalinplaneloading
applied
Externallyappliedloadscanbedirectorshearinplaneloadingsoroutofplanebendingortwisting.
A fully worked example is provided. It is important that thesign conventionsof thestiffnesses and
stresses are in accordance. The stiffnesses required for the calculation are set out in Data Item No.
75002,[See:16.8].
208
ECSSEHB3220Part2A
20March2011
16.8 Plate stiffnesses (In-plane)
16.8.1 Title
Stiffnessesoflaminatedflatplates.

Tocorrectlyanalyselaminatedcompositeplates,theinplaneelasticityshouldalsobeknownsothat
thefollowingattributescanbedetermined:
Stress
Strain
Deflection
Stability
Vibration
Moststructurallaminatesarecomposedofhighlyorthotropiclayersanditisimportanttounderstand
howthearrangementofthelayerscaninfluencethestiffnesses.Itisalsoimportanttounderstandhow
theplatestiffnessesrelatetothegeneralequationforasolidbody.Afurtherknowledgemetbythe
data,isinformationontheeffectofaddinglayersandtheirinfluenceonplatesymmetry.
16.8.3 Analysis and methods

The fundamental equations used to calculate inplane stiffness, is set out and the calculation of the
stiffnessesforanexamplelaminateisworkedthrough.Theeffectoflayerarrangementisdescribedfor
the common case where all the layers are specially orthotropic, have the same properties and
thicknessbutarelaidupatdifferinganglestotheplateaxes.Specialcasesareidentifiedandthenthe
influenceoflayuparrangementonthedirect,bendingandcouplingtermsisexplained.Theeffectof
repositioningtheplatereferenceaxis(aboutwhichthestiffnessesarecalculated)isexplainedandthe
relevantequationsaresetout.Tablesofthevariousformsofstiffnessmatricesthatcanbeconstructed
aresetouttogetherwithexamples.
209
ECSSEHB3220Part2A
20March2011
16.9 Bonded joints - 1
16.9.1 Title
Shearstressesintheadhesivesinbondedsinglestepdoublelapjointsloadedintension.

In order to design bonded lap joints, it is necessary to determine the stress distribution in the
adhesive.Theshearstressesareoneofthemajorcomponentsofthestressdistributionandassessment
of their maximum value with account of their inelastic behaviour is therefore necessary. This Data
ItemiscomplementarytothecomputerprograminDataItemNo.79016,[See:16.10],andexamines
thesinglestepjoint.Amajordesignobjectiveisthatofminimisingstressesbycarefulselectionofjoint
geometryandadhesive,[See:16.13forpeelordirectnormalstresses].

A shear lag analysis is used to determine the distribution of shear in the adhesive resulting from
transmission of the inplane tension or shear loads from one adherend to the other through the
adhesive, taking account of the nonlinear stressstrain shear characteristics of the adhesive. It is
assumedthattheadherendsarerigidinflexureandtheanalysisdoesnotthereforeconsidertheeffect
ontheadhesiveortheadherendsofnormaltensileloads(peel)orbendingeffectsresultingeitherfrom
thetransmissionoftheloadfromoneadherendtotheotherorfromexternalbendingmoments.The
adherend is considered to behave elastically. The joint stresses are calculated with the aid of the
program in ESDUpac A7916, [See: 16.10], and are presented graphically for three different adhesive
materials; with properties stated. The graphs show how the maximum stress is influenced by the
geometryandstiffnessesofthejointandaidwiththeuseoftheprogram.Cleartrendsarediscernible
andthefactorsinfluencingthestressesandtheirrelativesensitivitycanbeseen.
210
ECSSEHB3220Part2A
20March2011
16.10.1 Title
Inelasticshearstressesandstrainsintheadhesivesbondinglapjointsloadedintensionorshear.

Inordertodesignbondedlapjointsitisnecessarytodeterminethestressdistributionintheadhesive.
Inordertosmooththestressdistributionmostjointsaresteppedanddetailsofthestressdistribution
at each step are calculated . The shear stresses are one of the major stress components of the stress
distributionandanassessmentoftheirmaximumvaluewithaccountoftheinelasticbehaviourofthe
adhesivematerialandadherendstiffnessesateachstepisthereforenecessary.Adesignobjectiveisto
selectthecombinationofstepgeometriesandadhesivestiffnesstominimisetheshearstressesandso
obtainanoptimumjoint,[See:16.12forpeelstresses].

A shear lag analysis is used to determine the distribution of shear in the adhesive, resulting from
transmission of the inplane tension or shear loads, from one adherend to the other through the
adhesive,takingaccountofthenonlinearstressstrainshearcharacteristicsoftheadhesive.Itassumes
thattheadherendsarerigidinflexureanddoesnotthereforeconsidertheeffectontheadhesive,or
theadherendsofnormaltensileloads(peel),orbendingeffectsresultingeitherfromthetransmission
of the load from one adherend to the other, or from external bending moments. The adherend is
considered to behave elastically. Each step is divided up into small increments and the stress
distributionovereachincrementistakentobeuniform.
211
ECSSEHB3220Part2A
20March2011
16.10.4 ESDUpac A7916
16.10.4.1 General
PreviouslyESDUpacE1007.
Thisprogramcalculatestheshearstressesandstrainsintheadhesivesinmultisteplapjoints.
16.10.4.2 Input
Numberofstepsinjoint.
Geometry and elastic properties of adherends over each step. Elastic shear modules for
adhesive.
Shearstressinadhesiveatelasticlimit.
Shearstresstowhichadhesivestress/straincurveisasymptotic.Numberofincrementsineach
step.
Requirementfortensionortransverseshearloadingonjoint.
16.10.4.3 Output
Adhesiveshearstressesandstrainsateachincrementoneverystep.
Adherendnominalstressesateachincrementoneverystep.
Allinputdata.
[See:16.3foraccesstoprogramanddocumentation]
16.11.1 Title
Elasticstressesintheadhesiveinsinglestepdoublelapbondedjoints.

In order to design bonded joints it is necessary to determine the stress distribution in the adhesive.
The direct (normal throughthethickness direction) stresses and the shear stresses are the major
stressesgoverningjointdesign.Thedirect(normal)stressespeakattheendsofthejoints.ThisData
ItemiscomplementarytothecomputerprograminDataItemNo.80039,[See:16.12]andexaminesa
singlestep joint. A major design objective is minimising the stresses by careful selection of joint
geometryandadhesive.
212
ECSSEHB3220Part2A
20March2011

The data arebased on the elastic flexible joint analysis, [See: 16.12]. The joint stresses are calculated
with the aid of the program of ESDUpac A8039 and are presented graphically, for a range of non
dimensionalparameterratiosofjointcomponentstiffnessesandgeometry.Thegraphsshowhowthe
stresses are influenced by the geometry and stiffnesses and therefore aid design of this and more
complex types of joints using the program. Clear trends are desirable and the relative sensitivity of
stressestothefactorsaffectingthemcanbeseen.
16.12 Buckling of rectangular specially orthotropic plates
16.12.1 Title
Bucklingofrectangularspeciallyorthotropicplates.

Overall flexural buckling of laminated composite plates, subject to biaxial inplane compressive
loadingorinplaneshear,canbeadesignlimitation.
For speciallyorthotropic materials, laminated composite plates in which thelayupis such that end
load and shear, and bending and twisting moments, are all uncoupled, [See also: 16.16], the elastic
flexural buckling loads can be ascertained from the boundary conditions and the four flexural
stiffnesses. Flexural buckling ignores the through thethickness shear stiffness and therefore is an
upperlimittothebucklingload,[Seealso:16.15].

The elastic analysis of orthotropic plates subject to inplane loadings and with clamped or simply
supported edges shows that, with suitable manipulation, the buckling solutions for isotropic plates
can be used if the terms are appropriately modified by the orthotropic properties. By this means
solutionsarepresentedgraphicallyforthecases:
Plateloading Conditionsofplateedges
Uniaxial Allsidessimplysupported
Allsidesclamped
Onepairofoppositesidessimplysupportedtheotherpair
clampedagainstrotation

Biaxial Allsidessimplysupported
Onepairofoppositesidessimplysupportedtheotherpair
clamped
213
ECSSEHB3220Part2A
20March2011
Longplateswiththelongsidessimplysupported

Shear Allsidessimplysupported
Longplateswiththelongsidesnormallysupportedand
elasticallyrestrainedagainstrotation

Shearplusuniaxial Allsidessimplysupported
16.13.1 Title
Elasticadhesivestressesinmultisteplapjointsloadedintension.

In order to design bonded joints it is necessary to determine the stress distribution in the adhesive.
The direct (normal, throughthethickness direction) stresses and the shear stresses are the major
stresses governing joint design. The direct (normal) stresses peak at the ends of the joint and where
steps in the adherend are introduced with the maximum usually occurring at the extreme ends. A
design objective is the selection of joint geometry and adhesive to prevent failure under the normal
stressesdeveloped.

Theprogramisbasedontheelasticflexiblejointanalysisdescribedhere.Theanalysisdeterminesthe
distributionofshearanddirectnormalstressesintheadhesiveresultingfromthetransmissionofin
plane tension loads from one adherend to the other through the adhesive, zero external bending
momentsareassumed.
Theassumptionsmadeare:
Adherendstressesnormaltotheplaneofthejointandadhesivedirectstressesintheplaneof
thejointareneglected.
Also,thejointanditsadjacentsheetaretakentobehaveaccordingtothetheoryofcylindrical
bendingofplatesofsteppedcrosssectionandneutralplane.
The adhesive layer behaves elastically, is negligibly thin and provides a perfect shear
connectionbetweenthetwoadherends.
Consequently,atbothendsofthebondedlength,andtheinterstepboundaries,thetotaljointdepthis
immediatelyeffective in bending and in tension. Thisleads to an overestimate of the stress, but has
littleinfluenceonthestressesinthecaseofmultistepjoints.
214
ECSSEHB3220Part2A
20March2011
16.13.4 ESDUpac A8039
16.13.4.1 General
PreviouslyESDUpacE1020.
Thisprogramcalculatestheelasticshearanddirect(normal)stressesinabondedlapjoint.
16.13.4.2 Input
Numberofstepsinthejointandjointconfigurationtypenumber.
Elasticpropertiesofadherends.
Elasticpropertiesofadhesive.
Geometryandappliedloading.
16.13.4.3 Output
Shear and direct (normal) stress distribution across the joint and their ratios to their average
valueinthejoint.
Jointgeometryandelasticpropertiesoftheadhesiveandeachadherend.
Appliedloading.
Allinputdata.

16.14.1 Title
Guidetotheuseofdataitemsinthedesignofbondedjoints.

In order to make the most efficient use of data for the analysis of bonded joints it is necessary to
understand their limitations and helpful to know their relationship to the other analytical methods
available.Theinfluenceofthedataontheresultingdesignsisalsoofinterest.Thescopeofthedata
availableinESDUDataItemsonbondedjointsisaddressedtogetherwiththeavailabledataonjoint
design,[Seealso:16.9,16.10,16.11and16.12].
16.14.3 Information and guidance

Themajortheoriesonanalysisofthestrengthofbondedjointsarebrieflynoted.Thisisfollowedbya
comparison of the analytical methods and their relevance to joint design. This commences with the
jointtypesandthenrelatesjointfailuremodestothedataavailableforassessingfailure.
215
ECSSEHB3220Part2A
20March2011
Amajorfactorindesigningbondedjointsistheavailabilityofthestrengthandstiffnesspropertiesof
the adhesive. Where these are not available they should be determined experimentally and suitable
test methods are described. Data for fourteen adhesives are listed and shear stressstrain curves
presented. The major factors influencing the strength data are discussed and the influence of some,
e.g.temperatureandhumidity,isillustrated.
16.15 Buckling of orthotropic plates
16.15.1 Title
Bucklingofflatrectangularorthotropicplates(Applicationtolaminatedcompositeplates).

Laminatedcompositeplatesandsandwichpanelsfacedwithlaminatedcompositescanbemodelled
as homogeneous orthotropic plates. For such plates overall buckling is an important condition.
Bucklingisa function of the plates inplaneand flexural stiffnesses,its through thethickness shear
stiffnesses,itsshape,boundaryconditionsandtheappliedloading.Thethroughthethicknessshear
stiffnessesareofparticularimportancesince,iftheyareverylow,theycandominatebuckling.

Theplateistreatedashomogeneous,andcanbeisotropicorspeciallyorthotropic.Thethroughthe
thickness shear stiffnesses in the principal plane of orthotropy and orthogonally to that plane are
included. The elastic buckling load is derived for symmetrical combinations of clamped and simply
supported edge conditions for rectangular plates subject to combinations of inplane compression,
tension and shear loading. The solution is presented as a computer program. Buckle wave forms
havingupto7halfwavesalongoracrosstheplateareexamined.
16.15.4 ESDUpac A8147
16.15.4.1 General
PreviouslyESDUpacsE1001andE1031.
Theprogramcalculatestheelasticoverallbucklingloadsfororthotropicrectangularplatesfromthe
propertiesofthelayersfromwhichitisconstructedandtheplatesizeandboundaryconditions.The
program of ESDUpac A8147 can deal with any combination of the plate materials, loadings and
boundaryconditionsdescribed.
16.15.4.2 Plate materials

Coversanyplatewhichisspeciallyorthotropic,[See:16.16].Someexamplesofplatetypesthatcanbe
withinthislimitationandcanbedealtwithbytheprogramare:
216
ECSSEHB3220Part2A
20March2011
isotropicmaterials
sandwichpanels,withisotropicororthotropiccoresandfaceplates.
laminatedfibrereinforcedplates,composedoflayershavingdifferentproperties,thicknessesor
orientations of their principal axes of orthotropy. Such plates can conform exactly to the
conditions of special orthotropy if suitable ply stacking sequences are chosen. However,
approximatesatisfactionismorecommonlyencountered.
16.15.4.3 Plate loadings

Anycombinationofdirectbiaxialloadsandshearloadinginwhichanytwoloadsarespecified.Also
calculationswhereallthreeloadsareinafixedratiotooneanother.
16.15.4.4 Plate boundary conditions

Alledgeseitherclampedorsimplysupportedoranoppositepairofedgesclampedandtheotherpair
simplysupported.
16.16 Lay-up arrangements for special orthotropy
16.16.1 Title
Laminatestackingsequencesforspecialorthotropy(Applicationtofibrereinforcedcomposites).

In most structural laminates analysis is simplified, stability is maximised, and distortion during
manufactureisminimisedifstackingsequencesareusedwhichensurethatendloadandshear,and
bending and twisting moments, are all uncoupled. The uncoupled condition is termed special
orthotropy.ThisDataItemprovidesthedefinitivelistofspeciallyorthotropicstackingsequencesfor
laminates of up to 21 layers. In addition, notes are given on combining the stacking sequences
providedandonstackingarrangementsforlaminateswithlargenumbersoflayers.
The provision of tabulated lists of layups that are specially orthotropic enables easy selection of
arrangementsthatmeetbothdirectandflexuralloadings.

Aformaldefinitionofspecialorthotropyisgiven.Layupsareseparatedintofourgroups:symmetric
laminateswithevenoroddnumbersoflayersandantisymmetricgroupswithevenoroddnumbers
oflayers.Foreachgroupthesequencesof0(or90)andangledlayerswhichgivespecialorthotropy
forplatesofupto21layersarelisted.Ofparticularimportanceisknowledgeofthesmallestnumber
of layers in each group with which special orthotropy can be achieved. The tables list 48 layups of
symmetric (even), 27 layups of symmetric (odd), 380 layups of antisymmetric (even) and 273 lay
upsofantisymmetric(odd)allofwhicharespeciallyorthotropic.
217
ECSSEHB3220Part2A
20March2011
The tables also list coefficients which enable easy calculation of the terms of the flexural stiffness
matrixfromthestiffnessmatrixofoneoftheidenticallayersfromwhichtheplateiscomposed.
Rulesforextendinglayupsbeyondthelimitof21layerslistedareprovided.
16.17 Failure modes of laminated composites
16.17.1 Title
Failuremodesoffibrereinforcedlaminates.

In long fibre laminated composites, the fibres and matrix are designed to perform complementary
functions and therefore react to the various loading conditions in different ways. Additionally, the
orientationofthefibresusuallyvariesfromlayertolayerthroughthelaminatethickness.
Consequently there are many more potential failure modes for such laminates than for metallic
materials.Therearetwoprincipalwaysofconsideringthefailureofsuchlaminatesbasedonthe:
a. Behaviouroftheindividuallayersandtheinterfacesbetweenthem.
b. Laminateasawhole.
DataItemNo.82025dealswiththefirstoftheseapproaches.
16.17.3 Analysis and failure modes

In each layer of a laminate the principal function of the fibre is to transmit loads in its longitudinal
direction,althoughitalsotransmitstransverseandshearloads.Theresinmatrixconnectsthefibres,
maintainingtheirrelativepositionswhiledistributingtheloadamongthefibres,ensuringcontinuity
of transverse and shear loading in the laminate. The typical laminate is composed of a number of
individuallayers.Thelaminatecanbesubjectedtoanycombinationofloading,thoughthelayersare
most efficient at taking inplane longitudinal direct load. The failure modes are best understood by
dealing first with the modes for a single layer, under simple loading and then relating these to the
modesformorecomplexconfigurations.
Twelvedifferentmodesoffailureofalayerofalaminatedplatereinforcedbyunidirectionalfibresare
identified,illustratedanddiscussed.Theprimaryfactorsandsecondaryfactorsaffectingeachmode
arelisted.
Inadditiontotheindividuallayerfailuresmultidirectionallaminatescanfailinupto6furthermodes.
Thesearelistedanddescribedtogetherwiththeprimaryandsecondaryfactorsaffectingthem.
218
ECSSEHB3220Part2A
20March2011
16.18 Failure criteria for layers of a laminated composite
16.18.1 Title
Failurecriteriaforanindividuallayerofafibrereinforcedcompositelaminateunderinplaneloading.

Many different criteria have been developed for predicting the failure of the layers of long fibre
laminated composites, subject to the full range of inplane loadings. Many are identical in some
regions and some show discontinuitiesin theirfailure envelope,whilst otherenvelopesare smooth.
The provision of the appropriate strength data, and the choice of a suitable criterion are linked and
shouldbeconsideredtogether.BothaspectsareexaminedinDataItemNo.83014.

Eight of the commonly used failure criteria are described in detail and the failure envelopes
illustrated. The equations for a further three criteria are given. The types of failure criteria are
identified and tabulated in a coherent notational system. This highlights the similarities and
differences between the criteria. Comparisons of the failure predictions with experimental data are
shown for four of the criteria and failure envelopes are compared for all the criteria. The choice of
criterionappropriatetothedatacommonlyavailableisdiscussed.
Strengthtestdatausedinthevariouscriteriaarenotedtogetherwithillustrationsofthetypicaltests
thatwereusedtoobtainthem.Referencesareprovidedtothesestandardtests.
16.19 Plate stiffnesses and apparent elastic properties
16.19.1 Title
Estimationofthestiffnessesandapparentelasticpropertiesoflaminatedflatplates.

Stiffnessesandelasticproperties,thatcanbeestimatedusingItemNo.83035,whichcanbeusedfor
strength and stability analyses of structures. The item provides a computer program, referred to as
ESDUpac A8335, for the estimation of the stiffness and apparent elastic properties of composite
material laminated plates. The plate stiffnesses generated are those required for the strength, or
stability, analysis of thin plates and the apparent elastic properties are those required when using
orthotropic plate analysis methods. The program is applicable to plates composed of layers having
differentproperties,thicknessesandorientationsoftheprincipalaxesoforthotropy.Individuallayers
areassumedtobeorthotropicandhomogeneousandthebondingbetweenlayersisconsideredtobe
219
ECSSEHB3220Part2A
20March2011
perfect.Withintheitem,approximaterelationshipsaregivenforlaminaelasticpropertiesintermsof
individualfibreandmatrixpropertiesandtheirvolumefractions.Theserelationshipscanalsobeused
to estimate changes in known elastic properties of an individual lamina at different fibre volume
fractions.
16.19.3 Analysis for stiffnesses and elastic properties

TheplanestressconstitutiveequationforindividuallayersisformedusingHookeslaw.Thenusing
the straindisplacement relations in the expressions for the stress and moment resultants the
constitutiverelationsyieldtheinplanestiffnesssubmatrixA,thecoupledinplaneflexuralstiffness
submatrixBandflexuralstiffnesssubmatrixD.Theapparentorthotropicplateelasticpropertiesare
derived from elements of the inplane stiffness submatrix. These latter values are applicable to
multilayeredplateshavingsixormorelayers.
16.19.4 ESDUpac A8335
16.19.4.1 General
Thisprogramestimatesthestiffnessesandapparentelasticpropertiesoflaminatedflatplates.
16.19.4.2 Input
Laminaelasticpropertiescanbeeitherinputorestimatedfrominputvaluesoftheconstituentmatrix
and fibre elastic properties and the fibre volume content. Other inputs are the thickness and layup
angleofeachlayerandthelayerstackingsequence.
16.19.4.3 Output
Anyestimatedlaminaepropertiesarefirstoutputfollowedbytheplatelayuparrangement.A,Band
Dstiffnessmatricesarethenoutputfollowedbytheplateapparentelasticproperties.
16.19.4.4 Limitations
Maximumof5laminaeforwhichelasticpropertiesarecalculatedandamaximumof50layers.
16.20 Natural frequencies of laminated flat plates
16.20.1 Title
Naturalfrequenciesofrectangularspeciallyorthotropiclaminatedplates.
220
ECSSEHB3220Part2A
20March2011

Naturalfrequenciesshouldbeusedfordynamicresponsecalculationssuchastheanalysisforacoustic
fatigue.Theitemprovidesacomputerprogram,referredtoasESDUpacA8336,fortheestimationof
the natural frequencies for initially unstressed, rectangular, specially orthotropic plates. Specially
orthotropic plates are balanced laminates having layers of equal thickness, elastic properties and
orientationsthataresymmetricallyarrangedabouttheplatecentralplane.Theprogramenablesthe
estimationofthenaturalfrequenciesofflatplateshavinganycombinationofsimplysupportedand
clampededgeconditions.Theitemalsoprovidesanapproximateprocedurefortheestimationofthe
fundamentalnaturalfrequencyofslightlycurvedplateshavingalledgesclamped.
16.20.3 Calculation of natural frequencies

Forspeciallyorthotropicplates,theinplaneandflexuralstiffnesssubmatricesareuncoupledsothat
theflexuralstiffnesssubmatrix(Dmatrix)onlyisusedfortheestimationofnaturalfrequencies.The
method of analysis is based on the RayleighRitz method in which the plate displacements are
represented by a single beam characteristic function in each of the x, y orthogonal directions. In
calculating natural frequencies shear deformation and rotary inertia effects are neglected. Also
displacementsareassumedtobesmallrelativetotheplatethickness,andtheplateisassumedtobe
thinrelativetotheplatedimensions.
16.20.4 ESDUpac A8336
16.20.4.1 General
The computer program evaluates natural frequencies for initially unstressed, rectangular, specially
orthotropicplates.
16.20.4.2 Input
ElementsoftheDmatrixcanbeinputorevaluatedfromtheindividuallayerproperties,theirrelative
orientationsandlayupsequence.Theplategeometryandthemodenumbersofinterestareinput.
16.20.4.3 Output
Input data are output, and if the elements of the Dmatrix are calculated, these are output together
withtheplateconstruction.Naturalfrequenciesareoutputforthemodenumbersrequired.
IftheelementsoftheDmatrixarecalculated,amaximumnumberof50withintheplate.
221
ECSSEHB3220Part2A
20March2011
16.21 Strain in skin panels under acoustic loading
16.21.1 Title
Estimationofr.m.s.straininlaminatedskinpanelssubjectedtorandomacousticloading.

Laminatedcompositeplates,onthesurfaceofaerospacevehicles,canbesubjectedtointenseacoustic
loading during part, or all, of their normal working cycle. Knowledge of the response of such
structuresunderrandomacousticloadingisneededinordertoassesstheiracousticfatiguelife.The
item provides a computer program, referred to as ESDUpac A8408, for estimating the r.m.s. surface
strains in rectangular laminated skin panels subjected to random acoustic loading. The data are
applicable to specially orthotropic plates having principal axes of orthotropy parallel to the plate
edges. The computer program evaluates r.m.s. strains on the centreline of the plate parallel to the
shorter(curved)sideattheplateedgeandtheplatecentre.
16.21.3 Calculation of surface strains

Theassumptionofspecialorthotropyensuresthattheinplaneandflexuralstiffnessesareuncoupled
and enables the plate bending moments to be simply related to the elements of the plate flexural
stiffnesssubmatrix(matrixD).Thisrelationshipisusedtoevaluatetheratioofstressatapointonthe
surface of the plate to the applied static pressure. This ratio is used in a simplified normal mode
methodwhichassumesthatthepredominantformofskinvibrationisoneinwhichindividualplates
withinthestiffenedpanelvibrateindependentlyintheirfundamentalfixededgemode.Furtheritis
assumed that the pressure is uniform and inplane over the range of frequencies close to the
fundamental natural frequency of the panel. Empirically derived correction factors are used to
improvethecorrelationbetweenestimatedandmeasuredsurfacestrains.
16.21.4 ESDUpac A8408
16.21.4.1 General
PreviouslyESDUpac1042
Thecomputerprogramevaluatesther.m.s.strainresponseofthick,quasiisotropicplatesunderthe
actionofacousticloading.Thethickcriterionisusuallysatisfiedifplateshavenotlessthansixlayers.
16.21.4.2 Input
ElementsoftheDmatrixandplateorthotropicelasticpropertiescanbeinputorevaluatedfromthe
individual layer properties, their relative orientations and layup sequence. The plate geometry
togetherwiththeresponsefrequency,dampingratioandsoundpressurelevelareallinput.
222
ECSSEHB3220Part2A
20March2011
16.21.4.3 Output
Inputdataiscomputed,theplatestiffnesssubmatricesandelasticpropertiesarecalculated,theyare
output together with the plate construction and apparent elastic properties. For each set of data the
r.m.s.strainsaretabulated.
Ifplatestiffnesssubmatricesandelasticpropertiesareestimated,amaximumnumberof50layersis
possible.
16.22 Failure analysis
16.22.1 Title
Failureanalysisoffibrereinforcedcompositelaminates.

Thedesignoflongfibrelaminatedcompositesinvolvesananalysisofthefailureofthevariouslayers
(attheirvariousorientations)undertheloadingappliedtothecomposite.Thisisachievedbyastress
analysisoftheplateandthelayersthereinandthentheapplicationofafailurecriterion.Thefailureof
the matrix material supporting a fibre does not necessarily constitute total failure and it is often of
value to follow the various matrix failures that occur as the load increases up to the point where
collapse occurs. This makes it possible to take account of the fact that minor failure of the matrix
materialatoneloadisnotcompletecollapse.
16.22.3 Analysis
The failure assessment is based on a linear elastic layerbylayer plate analysis of the laminate, in
whichthestressesineachlayerinturnarecomparedwiththePuckModifiedCriterion,[See:16.18],
which differentiates between fibre and matrix material failures. If a layer matrix failure occurs, the
matrixmaterialpropertiesofthatlayerareremovedbyassumingthatitstransversedirect(inplane)
andshearstiffnessesarezero.Itisassumedthatwhenthishappensthefibrecanstilltakeloadinits
direction and that the laminate throughthethickness shear integrity is retained. Therefore, during
furtheranalysis,thelayerlongitudinal(fibre)stiffnesspropertiesareretainedasistheallowancefor
the thickness of the layer. Because these layer matrix failures can lead to the progressive failure of
otherlayerstheirdevelopmentistracedthroughthegradualapplicationofaloadingsysteminwhich
matrix material failures are present to provide a damage history under that system. This continues
untileither total matrixfailure has occurred throughout the laminate orasinglelongitudinal(fibre)
layerfailureoccurs,atwhichpointtotallaminatefailureisassumed.
Analysis is performed by a Fortran computer program. A laminate is analysed under consecutive
systems of combined inplane, moment and thermal loading. The program differentiates between
individuallayermatrixmaterialfailure,termeddamaged,anindividuallayerfibrefailureormatrix
223
ECSSEHB3220Part2A
20March2011
materialfailureofallthelayers,bothtermedlaminatefailure.Theconsecutiveloadingsystemscanbe
appliedwitheitherthedamageofeachloadingsysteminturncarriedovertothenextorwitheach
new loading system applied to the laminate as originally undamaged. If laminate failure does not
occurunderaparticularloadingsystemanassessmentofthelevelatwhichfailurewouldoccurfor
thatsystemisprovidedbeforetheanalysisreturnstothenextloadingsystem.
The program applies to balanced or unbalanced laminates that are initially unrestrained but also
includes the effect of holding the unloaded laminate flat (for storage), if required. While thermal
effectshavebeenincludedintheanalysismoistureeffectshavenotbecauseofthelesscertain,time
dependent,longtermnatureofsucheffects.However,moisturedegradationcanbeincorporatedfor
eachindividuallayerbyprovidingreducedvaluesoflayermaterialpropertiesandallowablestresses,
particularlyforthematrixmaterial.
16.22.4 ESDUpac A8418
16.22.4.1 General
Thisprogramtracestheinplanedamagehistoryofalaminatethroughtofailure.
16.22.4.2 Input
Propertiesofeachtypeoflayerused:
stiffness
Poissonsratio
coefficientsofthermalexpansion
thickness.
Allowablestressesintension,compressionandshearofeachtypeoflayerused.
Layupstackingarrangementandlayerorientations.
Loadingconditions:
appliedloads
bending
direct
shearinorthogonaldirections
workingtemperature
curingtemperature
restraintsatmanufacture
indicatorfordamagetobepassedtosuccessiveloadcasesornot.
16.22.4.3 Output options

Dataforlayerswherechosenfractionoffailurecriterionisexceededonly.
Extractoutputonly.
224
ECSSEHB3220Part2A
20March2011
16.22.4.4 Output
Fulloutputgivesstressesineachlayertogetherwiththevalueofthefailurecriterionatevery
stepofthecalculation.
Astatementofdamagedlayersateverystage.
Allinputdata.
16.23 Endurance under acoustic loading
16.23.1 Title
Endurance of fibrereinforced composite, laminated structural elements subjected to simulated
randomacousticloading.

Often a final requirement from an acoustic fatigue calculation is an assessment of the life of the
structure.Thiscanbemadefromplotsofr.m.s.strainagainstequivalentendurance,takenashalfthe
numberofzerocrossingstofailureofthestraintimefunction.ThesedataareoftenreferredtoasSN
plots. This item gives the results of fatigue tests on fibrereinforced composite specimens, typical of
aircraft structural elements, subject to narrowband excitation to produce random amplitude
responses,withzeromean,tosimulateresponseinaresonantmodetoacousticloading.Theresultsof
thesetestsareplottedasr.m.s.surfacestrains,atareferencelocation,againstequivalentendurance.
The data presented in this item are intended to be used only for guidance since, in some cases,
unusualconfigurationsweretestedandnotallrelevanttestdatawerereported.Endurancedataare
specifictothetypesofelementtested,themethodoftestingandthemethodoffailureidentification.
Testdataarewithintheendurancerange103to108cycles.
Thetypesofspecimenconsideredareplainspecimensandspecimensrepresentativeofbothbonded
andrivetedskinribconstruction,typicalofaircraftstructuralelements.Bothcarbonfibreandboron
fibre structural elements are considered and details are given of the element tested, the method of
testingandthemethodoffailureidentification.
Factorsthatcanaffecttheenduranceoffibrereinforcedcompositestructuralelementsarediscussed.
Newdataareaddedtothisitemastheybecomeavailable.
225
ECSSEHB3220Part2A
20March2011
16.24 Stress and strain around circular holes
16.24.1 Title
Elastic stress and strain distributions around circular holes in infinite plates of orthotropic material
(applicabletofibrereinforcedcomposites).
DataItemNo.85001Computerprogram[Seealso:ESDUInternationalwebsite].

Ameasureofthestressesandstrainsactingaroundtheperimeterofaholeinalaminatedcomposite
subjecttoacombinationofinplaneloadsisofinterestiffailureoredgedelaminationistobeavoided.
Thestressesandstrainsareafunctionofboththeloadingappliedandtheorthotropyofthematerial
andtheirrespectiveorientations.Thepointswherethestressesareamaximumarenotalwayscritical
becausetheycanalsobethepointswherethestrengthishigh.Itisthereforenecessarytosurveythe
perimeter to locate those points where the local stress exceeds the local strength in its direction of
action.Becauseofsymmetry,analysisofthestressesoveronlyhalftheperimeterisnecessaryandthis
isprovidedbytheESDUpacA8501programdescribed.
16.24.3 Analysis
The equations for the elastic stress field acting around the perimeter of a circular hole in a
homogeneous orthotropic plate subject to any combination of inplane loadings are set out.
Additionally the stressstrain relationships in both the circumferential direction and the chosen
Cartesian coordinates are set out. In application to laminated composites, the analysis is more
appropriatetoholeswhosediameterisgreaterthanabout6thicknesses.
16.24.4 ESDUpac A8501
16.24.4.1 General
PreviouslyESDUpac1048
Thisprogramcalculatestheelasticstressesandstrainsactingaroundacircularholeinanorthotropic
material.
16.24.4.2 Input
TheinplanemoduliandPoissonsratioforthelayers
Thelayuparrangement
Theappliedloading
Alternatively,theelementsoftheinplanedirectstiffnessmatrixortheplateeffectivemodulicanbe
input.
226
ECSSEHB3220Part2A
20March2011
16.24.4.3 Output
Thevaluesgivenat5intervalsaround180oftheholecircumferenceare:
twodirectandshearstressesintheplateaxes
twodirectandshearstrainsintheplateaxes
circumferentialandradialandshearstresses
inplanestiffnessmatrix.
inputdata.
16.25 Damping in composite plates
16.25.1 Title
Estimationofdampinginlaminatedandfibrereinforcedplates.

Knowledgeofthedampingofthestructureinthemodeofresponseisnecessaryfortheestimationof
the dynamic response to an applied loading. Fibrereinforced laminated structures are often
constructed with adhesive joints, thus eliminating sources of energy dissipation associated with
discrete joint attachments and consequently increasing the material damping contribution to the
overall damping. This item provides a computer program, known as ESDUpac A8512, for the
estimationofthematerialdampinginarectangular,multilayeredlaminatethatisvibratinginoneof
itslowerfrequencynaturalmodes.
The estimation procedure is applicable only to plates in which laminae having the same material
properties, and orientations of their principal axes of orthotropy, are symmetrically arranged about
theplatemidplane,i.e.balancedsymmetriclaminates.
The program estimates damping in plates having either all edges simply supported or all edges
clamped.Withintheitemapproximaterelationshipsaregivenforthespecificdampingcapacityofan
individual lamina in terms of components in longitudinal flexure, L, transverse flexure, T, and
longitudinalshear,LT. These damping components are necessary inputsforthe program. The item
alsogivesguidanceonthefactorsaffectingdamping.
16.25.3 Calculation of damping

The total energy dissipated is estimated by summing, over each layer of the plate, the energy
dissipated in longitudinal tensioncompression, transverse tensioncompression and longitudinal
shear. The specific damping capacity is then given by the ratio of total energy dissipated to the
maximumstrainenergystoredintheplate.
227
ECSSEHB3220Part2A
20March2011
Beamcharacteristicfunctionsareusedtodefinethemodeshapesanditisassumedthattheeffectof
inplaneforceonbendingisnegligible.Furtheritisassumedthat,iftheamplitudesofvibrationare
small,dampingcanbeconsideredtobeindependentofcyclicstressamplitude.
16.25.4 ESDUpac 8512
16.25.4.1 General
Theprogramestimatesthedampinginrectangular,multilayeredlaminates.
16.25.4.2 Input
Elasticpropertiesareinputtogetherwiththethreespecificdampingcapacities1,2and12foreach
layer. Other inputs are the thickness and layup angle and the layer stacking sequence, the plate
geometryandthelistofmodenumberstobeconsidered.
16.25.4.3 Output
Theplatelayerpropertiesandthelayerstackingarrangementareoutputfollowedbyatabulationof
dampingvaluesforeachmodeofvibrationconsidered.Dampingisoutputbothintermsofspecific
dampingcapacityandviscousdampingratio.
Maximumof30layersandmodenumberslessthan6.
16.26 Sandwich panel natural frequencies
16.26.1 Title
Naturalfrequenciesofsimplysupportedsandwichpanelswithlaminatedfaceplates.

A knowledge of the natural frequencies of a structure is often required before a dynamic response
calculation is made. The item provides a computer program, called ESDUpac A8537, for the
estimationofnaturalfrequenciesofinitiallyunstressedrectangularsandwichpanelshavinglaminated
compositefaceplates.Thepanelsareassumedtobesimplysupportedonallboundariesandcanbe
flat or singly curved with shallow curvature. In practical applications simply supported boundary
conditions are seldom achieved, so effective dimensions are used to estimate panel natural
frequencies.Guidanceonthechoiceofeffectivepaneldimensionsisgivenintheitem.Foreachpairof
mode numbers, the program enables five natural frequencies to be evaluated. These frequencies
correspond with modes having predominant deflections in the x, y and zdirections (the central
228
ECSSEHB3220Part2A
20March2011
plane of the plate being in the xyplane) and predominant rotations in the xz and yxplanes. An
optionisgiventocalculatethelowestnaturalfrequencyonly.

For the calculation of the natural frequencies and associated normalised panel deflections, it is
assumedthatthefaceplatesareorthotropicandhaveprincipalaxesoforthotropyparalleltothesides
of the panel. The face plates can be made of different materials and thicknesses. It is assumed that
bending is carried only by the face plates and transverse shear stresses only by the core. The
displacement forms used in the solution of the equations of motion are appropriate for simply
supportedboundaryconditions.Themethodofanalysisisbasedonlinearsandwichpaneltheoryand
isapplicabletopanelshavingthinfaceplatesandsmallvibrationamplitudes.
16.26.4 ESDUpac A8537
16.26.4.1 General
The computer programestimates the naturalfrequenciesand mode shapes offlat, or singly curved,
sandwichpanelshavinglaminatedcompositefaceplates.
16.26.4.2 Input
Theorthotropicfaceplatepropertiescanbeinputorevaluatedfromtheindividuallayerproperties,
their relative orientations and layup sequence. The core properties, panel geometry and mode
numbersofinterestarealsoinput.
16.26.4.3 Output
Inputdataareoutputandiffaceplateorthotropicpropertiesareevaluatedtheseareoutputtogether
with the face plate construction. Natural frequencies are output and mode shapes are tabulated if
requested.
Iffaceplateorthotropicpropertiesareestimated,amaximumof20layersispossibleineachfaceplate.
16.27 Selection of reinforcement around circular holes
16.27.1.1 Title
Example of the use of Data item No. 85001. Choice of reinforcement for a circular hole in a fibre
reinforcedlaminatedplate(Datarelatingtooneparticularsetoflaminateproperties).
229
ECSSEHB3220Part2A
20March2011
16.27.1.2 Usage and scope
In composite plates, the properties of the material vary with direction. This means that some
compositeshaveahighstressatapointaroundtheholeboundaryyetthisstressisinadirectionin
which the composite is particularly strong. Alternatively, another composite can give a lower stress
but the maximum stress is in a direction in which the material is less strong. The addition of
reinforcement is often anoption where the strength is poor and a method of deciding the optimum
reinforcementisprovided.ThestressesandstrainsareobtainedfromtheprogramofESDUpacA8501
[See:16.24].
16.27.1.3 Analysis and data

Agraphicaltechniqueisdescribedandillustrated,byexample,fordeterminingthecriticallocations
around the boundary of an unreinforced hole using data generated by ESDUpac A8501. Where the
strength is low, hole reinforcement is considered. The calculation of the reinforcement thickness is
describedtogetherwithitsdesirableorientation.Strictly,themethodonlyrelatestoholeswellaway
from edges or other discontinuities. The extent of the reinforcement is discussed and a
recommendationmade.
16.28 Buckling of unbalanced composite plates
16.28.1 Title
Elasticbucklingofunbalancedlaminatedfibrereinforcedcompositeplates.
(RectangularplatesofAsBtDstype,alledgessimplysupported,underbiaxialloading)

Incomplexlaminatedstructures,itisoftennotpracticaltoreducelaminatethicknessexactlyinline
withreducingloadandmaintainabalancedlaminate.Layercontinuityoverseveraladjacentpanelsis
often preferred for strength considerations even though it results in unbalanced plates within the
structure.
Theformofthestiffnessmatrixforanunbalancedplatecansignificantlyinfluencethebucklingload.
Wheresomechoiceinthelayuparrangementisopen,itispreferabletoselectalayupsuchthatthe
inplaneandflexuralstiffness matricesare populated in the same way as fora specially orthotropic
plate but the coupling stiffness matrix terms are zero except for the terms relating direct and shear
loads to curvature and twisting. Plates of that type are preferable because they exhibit bifurcational
buckling.(Mostotherunbalancedformsexhibitoutofplanebendingupontheapplicationofasmall
inplaneloadandthereafterbehaviourbecomesnonlinear.)
Although the result of employing an unbalanced plate that exhibits bifurcational buckling is a
reduction in buckling load this is still preferable to the nonlinear behaviour of other unbalanced
230
ECSSEHB3220Part2A
20March2011
plates.Reductionsupto30%ofthebalancedplatevaluecanoccurinextremecases,butreductionsof
10%or20%aremorecommon.

Thesolutionisbaseduponelasticthinplatesmalldeflectiontheoryinwhichtheplateisassumedto
behaveasahomogeneousorthotropicmaterialwhoseaxesoforthotropyarealignedwiththeedgesof
the plate. The theory only takes account of direct inplane, flexural and twisting deformations. No
accountistakenofthroughthethicknesssheardeformations.Theplateisassumedtobeinitiallyflat,
rectangular and simply supported along all four sides with the loadings applied inplane and
uniformlyalongallfouredges.Theappliedloadsareassumedtolieinthemidplaneoftheplateand
thisistheplanetowhichtheflexuralandcouplingstiffnessesarerelated.
16.28.4 ESDUpac A8620
16.28.4.1 General
Thisprogramcalculatesthebucklingloadofaparticularclassofunbalancedlaminatedflatplates.
16.28.4.2 Input
Platedimensions.
Transverseloadorlongitudinaltotransverseloadingratio.
Materialpropertiesofeachtypeoflayerusedintheplate.
Stackingsequenceoflayersandtheirorientations.
16.28.4.3 Output
Bucklingloadswithnumberofbucklehalfwavesalongandacrosstheplate.
Inplane,flexuralandcouplingstiffnessmatrices.
Plateconstruction,layerpropertiesandgeometry.
16.29 Sandwich panel response to acoustic loading
16.29.1 Title
Estimationofr.m.s.straininlaminatedfaceplatesofsimplysupportedsandwichpanelssubjectedto
randomacousticloading;includingasimplifiednaturalfrequencypredictionmethod.
231
ECSSEHB3220Part2A
20March2011

Sandwich panels are widely used in the design of lightweight structures for aerospace applications.
Wheresuchstructuresareusedinahighintensityacousticenvironmentitisnecessarytoinvestigate
theirresponsetoacousticloading.Theitemprovidesacomputerprogram,calledESDUpacA8624,for
theestimationofther.m.s.surfacestrainsinthelaminatedfaceplatesofsandwichpanelssubjectedto
random acoustic loading. The data are applicable to flat, or shallow cylindrically curved, sandwich
panels having thin laminated composite face plates. One stage within the analysis procedure is the
estimationofnaturalfrequenciesofvibrationsothattheprogramcanbeusedtoestimatethelower
naturalfrequenciesofthepanelbendingmodes.Thepanelsareassumedtobesimplysupportedon
all edges so that the reference location for surface strains is chosen to be the centre of the panel. In
practical applications simply supported boundary conditions are seldom achieved so that effective
dimensionsareusedtoestimatepanelnaturalfrequenciesandstrains.Guidanceisalsogivenonthe
choiceofeffectivedimensions.
16.29.3 Calculation of natural frequencies and surface strains

Forthecalculationofnaturalfrequenciesandr.m.s.surfacestrainsofsandwichpanels,thefaceplates
areassumedtobeorthotropicandtohaveprincipalaxesoforthotropyparalleltothepanelsides.The
upperandlowerfaceplatescanbedifferentmaterialsandthicknessesandthecoreisassumedtohave
zeroflexuralstiffness.Fortheestimationofr.m.sstrains,itisassumedthatthepanelrespondsonlyin
its fundamental mode and that the loading spectral density is constant over a range of frequencies
includingthepredominantresponsefrequency.
16.29.4 ESDUpac A8624
16.29.4.1 General
The computer program evaluates the bending mode natural frequencies, and strain response under
acousticloading,ofsandwichpanelswithcompositefaceplates.
16.29.4.2 Input
Theelasticpropertiesofthefaceplatescanbeinputorevaluatedfromtheindividuallayerproperties,
their relative orientations and layup sequence. The core properties and panel dimensions are also
input. For the calculation of natural frequencies the mode numbers of the frequencies of interest
shouldbeentered.Forr.m.s.straincalculations,thedampingandsoundpressurelevelsareinputfor
eachcase.
16.29.4.3 Output
Input data are output and if faceplate properties are calculated these are output together with the
faceplate construction. For each set of data natural frequencies are output (if required) and r.m.s.
strainsatthepanelcentreareoutputforboththeupperandlowerfaceplates.
Iffaceplateelasticpropertiesareestimated,amaximumof20layersispossibleineachfaceplate.
232
ECSSEHB3220Part2A
20March2011
16.30 Sandwich column and beam face plate wrinkling
16.30.1 Title
Elasticwrinklingofsandwichcolumnsandbeamswithunbalancedlaminatedfibrereinforcedplates.

One possible mode of failure of a sandwich column or beam with thin face plates is the localised
buckling(wrinkling)ofthecompressionfaces.Forfibrereinforcedfaceplatesmadeupfromanumber
ofthinlayers,itispossiblethatthefaceplateseachformanunbalancedlayupabouttheirlocalmid
planes.Thislackofbalanceisfurtheraccentuatedbythelayerofadhesivethatisusedtobondeach
face plate to the core material and which can have a significant effect on the wrinkling loads. It is
thereforenecessarytoexploretheinfluenceofthesefactors.
Twoformsofwrinklingcanoccur,symmetricalandantisymmetrical,andbotharetobeinvestigated
to determine the minimum buckling loading. Compressive loadings of interest are those applied
directlyandthoseoccurringasaresultofbending.
16.30.3 Analysis
The solution is based upon elastic, thinplate, smalldeflection theory, in which the laminated face
plates are assumed to behave as homogeneous orthotropic plates, whose axes of orthotropy are
aligned with the edges of the beam or column. For columns, the face plates are assumed to form a
balancedlayupaboutthecentralplaneofthesandwich.Inbeams,thetensionfaceplateisassumed
toprovidearigidboundaryforthecoreagainstwrinklingofthecompressionfaceplate.
Thetheoryonlytakesaccountofdirectinplaneandflexuraldeformationsasgovernedbythedirect
andflexuraltermsoftheinplane,flexuralandcouplingstiffnesssubmatrices.Noaccountistakenof
throughthethicknesssheardeformationsofthefaceplates.Thefacesareassumedtobeinitiallyflat,
andtobeheldflatbythecoreuntiltheywrinkle.Thecoreisassumedtogivecontinuoussupportto
thefacesandthepossibilityoflocalisedbucklingwithinhoneycombtypecells(oftencalledintercell
buckling)isnotconsidered.Also,thecoreistakentohavezerolongitudinalstiffnessbutfinitedirect
and shear stiffnesses in the throughthethickness direction. These assumptions make the analysis
moreapplicabletohoneycombtypecoresthantosolidfoamcores,butgiveconservativeresultsfor
the latter. Half the applied load is assumed to lie in the midplane of each face plate (which might
includealayerofadhesive)andthisistheplanetowhichthestiffnessesarerelated.
233
ECSSEHB3220Part2A
20March2011
16.30.4 ESDUpac A8713
16.30.4.1 General
Theprogramcalculatestheloadonasandwichcolumnorbeamatfaceplatewrinkling.
16.30.4.2 Input
Sandwichcorecompressiveandshearstiffnesses.
Sandwichthickness.
Stiffnesses,Poissonsratioofeachlayertypeusedinfaceplates.
Laminatedfaceplatelayerstackingsequenceandorientation.
16.30.4.3 Output
Loadatfaceplatewrinklinginsymmetricandantisymmetricmodes,andonbending.
Wavelengthsofwrinkles.
Faceplatestiffnesses,andeffectivestiffnessterms.
Allinputdata.
16.31 Buckling of curved composite panels
16.31.1 Title
Elastic buckling of cylindrically curved laminated fibre reinforced composite panels with all edges
simplysupportedunderbiaxialloading.

Curved panels can be subject to buckling under axial compressive loading. Additionally, curved
panels are often loaded by pressure on the convex face reacted at the sides parallel to the axis. The
resultingbucklingbehaviouriscomplex.Forrelativelyshortpanels,allbucklewavelengthsgivethe
sameminimumbucklingload.Forintermediatelengthsofpanel,usuallyoneoftwocombinationsof
wavelengthsdominatesand,forverylongpanels,anothercombinationdominates.Thebucklingload
decreases as the length increases for those longer panels. A method investigating all wavelength
combinationsisthereforenecessarytoensurethattheminimumbucklingloadisidentified.

The panels considered are of rectangular form and uniform thickness having cylindrical curvature
acrossthewidth.Alledgesofthepanelareassumedtobesimplysupported.Thepanelissubjectedto
directloadingapplieduniformlytotheendsofthepanelinthedirectionofthecylindricalaxis,which
234
ECSSEHB3220Part2A
20March2011
can be combined with similar loading along the sides of the panel applied in the circumferential
directioninordertoexertapressureonthepanelsurface.
Theanalysisisbaseduponelastic,thinplate,smalldeflectiontheoryinwhichthepanelisassumedto
behaveasahomogeneousorthotropicmaterialwhoseaxesoforthotropyarealignedwiththeedgesof
theplate.
Thetheoryonlytakesaccountofdirectinplane,flexuralandtwistingdeformationsasgovernedby
the inplane and flexural stiffness matrices. Throughthethickness shear deformations are
disregarded.
16.31.4 ESDUpac A8725
16.31.4.1 General
Theprogramcalculatestheaxialbucklingloadwithorwithoutpressureloadingapplied.
16.31.4.2 Input
Materialpropertiesofeachtypeoflayerused.
Stackingsequenceandorientationsofeachlayer.
Panelgeometry.
Circumferentialloading(reactingpressureontheconvexface).
16.31.4.3 Output
Axialbucklingloadandnumberofbucklehalfwavesalongandacrosstheplate.
Flexural,directandcouplingstiffnessmatrices.
Plateconstruction,loadingandgeometry.
Circumferentialloadingapplied.
16.32 Sandwich panel face plate wrinkling
16.32.1 Title
Elasticwrinklingofsandwichpanelswithlaminatedfibrereinforcedfaceplates.
235
ECSSEHB3220Part2A
20March2011

A possible mode of failure of a sandwich panel with thin face plates that is subjected to inplane
biaxialloadsisthelocalisedbuckling(wrinkling)ofthefaces.Forfibrereinforcedfaceplatesmadeup
fromanumberofthinlayers,itispossiblethatthefaceplateseachformanunbalancedlayupabout
theirlocalmidplanes.Thislackofbalanceisfurtheraccentuatedbythelayerofadhesivethatisused
tobondeachfaceplatetothecorematerialandwhichcanhaveasignificanteffectonthewrinkling
loads.Twomodesofbucklingoccur,symmetricalandantisymmetricalforms.
16.32.3 Analysis
The solution is based upon elastic, thinplate, smalldeflection theory, in which the laminate face
plates are assumed to behave as homogeneous orthotropic plates whose axes of orthotropy are
aligned with the panel loading directions. The face plates are assumed to form a layup about the
centralplaneofthesandwich.
No account is taken of throughthethickness shear deformation of the face plates. The faces are
assumedtobeinitiallyflat,andtobeheldflatbythecoreuntiltheywrinkle.Thecoreisassumedto
givecontinuoussupporttothefacesandthepossibilityoflocalisedbucklingwithinhoneycombtype
cells(oftencalledintercellbuckling)isnotconsidered.Also,thecoreistakentohavezerostiffnessin
thexandydirectionsbutfinitedirectandshearstiffnessesthroughthethickness.Theseassumptions
maketheanalysismoreapplicabletohoneycombtypecoresratherthantosolidfoamcores,butgives
conservativeresultsforthelatter.Halftheappliedloadisassumedtolieinthemidplaneofeachface
plate (which might include a layer of adhesive) and this is the plane to which the stiffnesses are
related.
16.32.4 ESDUpac A8815
16.32.4.1 General
Theprogramcalculatestheloadsonasandwichpanelatfaceplatewrinkling.
16.32.4.2 Input
Sandwichthickness
Coreshearstiffnessesandthroughthethicknesscompressivestiffness.
Faceplatepropertiescanbeinputaseither:
propertiesofindividuallayersandtheirthicknesses,layupstackingsequenceandlayer
orientations,or
theelementsoftheinplane,flexuralandcouplingstiffnessmatricesforthecompleteface
plateassembly.
Loadingarrangementsare:
fixedvalueoftransverseload,
longitudinaltotransverseloadingratiofixed.
236
ECSSEHB3220Part2A
20March2011
16.32.4.3 Output
Longitudinal load at wrinkling or longitudinal and transverse loads in prescribed ratio at
wrinkling.
Wavelengthofwrinklesalongandacrossthepanel.
Calculatedorinputstiffnessmatrices.
Allinputgeometryandmaterialsdata.
16.33 Vibration of singly-curved laminated plates
16.33.1 Title
Naturalfrequenciesofsinglycurvedlaminatedplateswithsimplysupportededges.

Curved laminated composite plates are used in various applications, particularly in the aerospace
industry. The item provides a computer program, called ESDUpac A8911, for the estimation of the
lower natural frequencies of singlycurved laminated plates. The data are applicable to thin,
rectangular plates of shallow curvature having all edges simply supported. Static inplane forces,
uniformlydistributedalongtheplateedges,canbeconsideredbutinplaneshearforcesareassumed
to be zero. Within the item, the effects of both curvature and inplane loading on composite plate
naturalfrequenciesareillustrated.

The method of analysis is applicable to thin laminated plates of shallow cylindrical curvature. The
program is not restricted to symmetric laminates but the shear coupling terms within the stiffness
matricesarenecessarilynegligible.Thelaminatedplateconstitutiverelationsaresubstitutedintothe
equations of motion and, using the straindisplacement and curvaturedisplacement relationships
obtainedfromclassicaltheoryofelasticity,asetofthreedisplacementequationsisderived.
Substitution of appropriate displacement relationships into the displacement equations leads to a
cubicequationinfrequency.Thelowestofthethreenaturalfrequencies,forwhichthepredominant
displacementisnormaltothesurfaceoftheplate,isevaluated.
16.33.4 ESDUpac A8911
16.33.4.1 General
The computer program estimates the natural frequencies of curved, laminated composite plates
havingsimplysupportededgeconditions.
237
ECSSEHB3220Part2A
20March2011
16.33.4.2 Input
The elements of the inplane stiffness submatrix (Amatrix), the coupled inplane flexural stiffness
submatrix (Bmatrix) and the flexural stiffness submatrix (Dmatrix) are necessary. These matrix
elements can be directly input or evaluated from the individual layer properties, their relative
orientations and layup sequence. The plate geometry, the inplane applied loads and the mode
numberlistarealsoinput.
16.33.4.3 Output
The input data are output and, if the stiffness submatrices are evaluated, the plate construction is
outputfollowedbytheA,BandDmatrices.Foreachloadingcasethenaturalfrequenciesarelisted.
Ifthestiffnesssubmatricesareevaluatedthemaximumnumberofplatelayersis20.
16.34 Plate through-the-thickness shear stiffnesses
16.34.1 Title
Transverse(throughthethickness)shearstiffnessesoffibrereinforcedcompositelaminatedplates.

The transverse (throughthethickness) elastic properties of continuous fibre reinforced composite
laminatedplatesare,toalargeextent,dependentuponthepropertiesofthematrixmaterialusedto
bondthefibrestogether.Formanymatrixmaterials,thesepropertiescanbeverylowcomparedwith
thoseinthedirectionofthereinforcingfibres,withtheresultthatsignificantdeflectionscanbecaused
bythroughthethicknessshearstresseswhichariseinplatebending.Itisnecessarytoobtainvalues
fortheshearstiffnesstermswhichrelatethetransverseshearforceresultantstothetransverseshear
strains. ESDUpac A8913 provides a FORTRAN computer program for the determination of these
stiffnessesforlaminatedplates,withoutrestrictiononthetypeoflayup.Theplatesareconsideredto
bebuiltupfromthinorthotropichomogeneouslayerswithperfectbondingbetweenthem.Thelayers
can have different material elastic properties, thickness and orientations of their principal axes of
orthotropy.Theanalysisisalsoapplicabletothicksandwichpanels.
16.34.3 Analysis
Mindlins plate theory is applied to orthotropic plates. The assumption implies that the transverse
strains are independent of the throughthethickness dimension, with the result that the
corresponding shear stresses are constant in each layer of the laminated plate if they are calculated
fromthesimpleshearstressstrainequations.Thisconstantstrainstressapproachleadstotransverse
shear stiffnesses which are usually too high. An improvement is obtained by using the stress
equilibriumequationsineachlayertoevaluatethetransverseshearstressdistribution(nottakentobe
238
ECSSEHB3220Part2A
20March2011
constantineachlayer).Thesestressesarethenrelatedtothetransverseshearforcesperunitwidth,
usingtheoverallplateequationsofcylindricalbending.Theshearstrainenergydensityisevaluated,
using the transverse shear stressstrain equations for each layer. This is used to obtain a flexibility
matrixandhencetheshearstiffnessmatrixgivingtherelationbetweenthetransverseshearforcesper
unitlengthandthetransverseshearstrains.
16.34.4 ESDUpac A8913
16.34.4.1 Input
Material properties of each layer type used; moduli and Poissons ratio for the three orthogonal
directions.Thicknessandlayupangleofprincipalaxes.Stackingarrangementofthelayers.
16.34.4.2 Output
Throughthethicknessshearstiffnessesoflaminate.
Inplane,bendingandcouplingstiffnessmatrices.
Inputpropertiesandlayuparrangement.
Limitedtolaminatesof500layersmaximumcomposedofupto50differentmaterialsororientations.
16.34.5 Notes
Alllayersarecheckedtoensurethattheinputpropertiesrelatetoarealmaterial.
16.35 Vibration of plates with in-plane loading
16.35.1 Title
Natural frequencies of isotropic and orthotropic rectangular plates under static loading; including
shearloading.

Oftenplatesaresubjectedtostaticinplaneloadingthatsignificantlyaffectsnaturalfrequencies.The
item provides a computer program, called ESDUpac A9016, for the estimation of the lower natural
frequencies of isotropic, flat rectangular plates under such loading. The static inplane loading,
uniformlydistributedalongtheplateedges,canbeanycombinationofdirectandshearloadingthat
doesnotcausetheplatetobuckle.Theprogramisapplicabletospeciallyorthotropicplatesforwhich
the principal axes of orthotropy are parallel to the plate edges. Plates having all combinations of
239
ECSSEHB3220Part2A
20March2011
clamped,simplysupportedandfreeboundaryconditionscanbeconsidered.Generalguidanceisalso
givenontheeffectofinplaneloadingonnaturalfrequenciesofplates.

TheRayleighRitzmethodofanalysisisusedinwhichtheplatedisplacementsarerepresentedbya
finite series of orthogonal polynomial functions. The set of orthogonal polynomials used are
sequentiallyderivedfromastartingpolynomialfunctionthatsatisfiestheboundaryconditionsofthe
plateunderconsideration.Platesareassumedtobethinandeffectsofsheardeformationandrotary
inertiaareneglected.
16.35.4 ESDUpac A9016
16.35.4.1 General
The program estimates the natural frequencies, gives a tabulation of associated plate displacements
andplotsnodallinesofisotropicandorthotropicplatesunderstaticloading.
16.35.4.2 Input
Theisotropicororthotropicmaterialproperties
Plategeometry
Inplaneappliedloading
Boundaryconditions
Thenumberoffrequenciesofinterestorthemaximumfrequencyofinterestcanbespecified.
16.35.4.3 Output
Inputdataarecalculatedforeachsetofinplaneappliedloads,naturalfrequenciesandifrequested
mode shapes and nodal line plots are output. If the applied loading exceeds the buckling load no
modaldataareoutput.
The maximum number of modes that can be evaluated is controlled by the chosen number of
orthogonalfunctionsusedtodefinethemodeshapes.
16.36 Delamination and free edge stresses
16.36.1 Title
Delaminationandfreeedgestressesinlaminatedplatessubjectedtoconstantuniaxialstrain.
240
ECSSEHB3220Part2A
20March2011

At free edge regions, classical plate theory cannot account for the three dimensional nature of the
stress field in which the unattached free edge has zero values of direct stress and shear stresses. At
free edges the stresses in the throughthethickness direction can be sufficiently high to cause local
delamination.ESDUpacA9021providesaFortrancomputerprogramfortheapproximatecalculation
ofthesestressesforplateswhicharesubjectedtoaconstantuniaxialstrain.Also,afailurecriterionis
usedtopredictedgedelaminationandtheinterlayerinwhichitoccurs.Theplatesarebuiltupfrom
thin orthotropic homogeneous layers and these can have different material elastic properties,
thicknessesandorientations.Themethodiscapableofanalysingplateswithahundredorsodifferent
layers. This gives it a major advantage over finite element methods which have difficulty in coping
withmorethan10orsolayers.
16.36.3 Analysis
Theanalysisofthefreeedgestressesisbasedonlinearelastictheoryandtreatsaplatesubjectedtoa
constantuniaxialstrainparalleltotheedge.Underthisconditionitisassumedthatthedependenceof
thestresscomponentsacrossandthroughtheplatecanbeseparated,andthatexponentialfunctions
can be used to represent them. The stressesare averaged over a chosen distance from the free edge
and then used in a quadratic failure criterion to predict when delamination is likely to occur.
Guidance on an appropriate value of the averaging distance is obtained from the available
experimental data. Not all the boundary conditions are properly met by analysis. However, those
whicharenotmetdonotappeartoaffectsignificantlythestressesobtained.
16.36.4 ESDUpac A9021
16.36.4.1 Input
Materialpropertiesofeachlayertype,moduli,Poissonsratiosandinterlaminarshearstresses
anddirectstressatfailure.
Layerthicknessandangleoforientation.
Layuparrangement.
Applieduniformstrainparalleltothefreeedge.
Averagingdistancetobeusedforedgestresses.
16.36.4.2 Output
Strainandequivalentloadtocausedelamination.
Stressesresultingfromappliedstrain.
Valuesofconstantsinstressequations.
Inplanedirectstiffnessmatrix.
Allinputdata.
Layerdatacanberequestedforallinterfacesoronlythosewithafailurecriterionaboveasetvalueor
forthosewithasetnumberofthehighestvaluesofthecriterion.
241
ECSSEHB3220Part2A
20March2011
Plateshouldbeofabalancedarrangementoflessthan500layerscomposedoflessthan50different
materialsandorientations.
16.36.5 Notes
Allmaterialsarecheckedtoensurethattheyrelatetoarealmaterial.Laminatesarecheckedtoensure
thattheyarebalanced.
16.37 Delamination at termination of plies
16.37.1 Title
Delaminationoftaperedcomposites.

Laminated composite plates can be manufactured with variable thickness by terminating, or
droppingoff,someofthepliesalongthelengthofthelaminate.Thus,discontinuitiesareintroduced
intheformofsteps,wherethepliesareterminated,andthisgivesrisetohighlocalstresseswhichcan
cause the dropped plies to delaminate from adjacent plies. ESDUpac A9103 provides a Fortran
computerprogramfortheapproximatecalculationoftheaxialloadtocausedelaminationintapered
laminatedplates,wheretheloadisappliedinthedirectionofthetaper.
16.37.3 Analysis
Delamination is predicted using linear elastic fracture mechanics combined with simple strength of
materialstheory.Anequationisobtainedfortheappliedloadtocausedelaminationintermsofthe
local droppedply geometry, the effective elastic moduli, and the critical strain energy release rate.
Thislatterquantitycanbeobtainedfromexperimentandatestmethodforobtainingitisdiscussed.
Also,guidelinesaregivenforgooddesignpractice.
Thefracturemechanicssolutionisapproximate.Althoughthetheoreticalpredictionsagreequitewell
with the available experimental results, there can be layup geometries and materials in which
agreementbetweentheoryandexperimentisnotsogood.Thisappliesmoretoasymmetrictapered
plates in compression where instability effects can lead to premature failure. Thus, the predicted
delaminationloadsareonlyusedasaguideforinitialdesignpurposes.Wherepossible,verificationof
final designs involves testing on specimens that are representative of the actual method of
manufacture,materials,geometryandloading.
242
ECSSEHB3220Part2A
20March2011
16.37.4 ESDUpac A9103
16.37.4.1 General
Thisprogramcalculatestheloadatwhichdelaminationislikelyataplyterminationpoint.
16.37.4.2 Input
Propertiesoflayersusedinlaminate:
stiffnesses
Poissonsratio
thickness
orientation.
Terminationarrangement:
layupandstackingsequenceoflayerseachsideoftheterminationpoint
configurationnumberforterminationarrangement.
Criticalstrainenergyreleaseratebetweenlayers.
16.37.4.3 Output
Meandelaminationstressesandloadsunderconditionsofbothzerotransversestrainandzero
transverseload.
Effectivemoduliofplateeithersideoftheterminationpoint.
Allinputdata.
16.38 Thickness selection to meet a loading combination
16.38.1 Title
Thicknessselectionforlaminatedplatessubjecttoinplaneandbendingmoments.

Using classical laminated plate theory, it is relatively straightforward to calculate the separate layer
stressesforgivenlayerthicknesses,anglesandmechanicalpropertiesoflaminatedplates.However,if
thelayerthicknessesarenotspecifiedattheoutsetthentheproblemoffindingasetofthicknessesto
withstandagivensetofloadsefficientlyisfarmoredifficult.
This procedure is provided for plates which are symmetrical about their midplane, with all layers
madeofthesamematerial,andthelayeranglesarelimitedto0,90,+45or45.Layersatangles
+45 and 45 occur as adjacent pairs. Also, all layers of the same angle have the same thickness. A
stackingsequenceforthelayeranglesisspecifiedatthebeginning,andadesignprocedureisusedto
243
ECSSEHB3220Part2A
20March2011
determine layer thickness for this layup. In this procedure, the thickness of layers is increased or
reducedaccordingtocertainrulesbasedonexpectationsofhowbestthelayersreinforceeachother.
Thesearereferredtoasheuristicdesignrules.Ifrequested,theprocedureprovidesalternativelayups
forcomparisonwiththatobtainedforthespecifiedlayup.Itcanalsotakeintoaccountpracticalply
thicknesses,whereeachlayerconsistsofmultiplesofaspecifiedstandardplythickness.
16.38.3 Analysis
Thelaminatedplateisassignedaparticularstackingsequenceforthelayeranglesandaninitialsetof
layer thicknesses is determined using a netting analysis, where only the fibres in certain specified
directionsareassumedtocarryagivensetofinplaneloadsandmoments.Theplateisthentestedto
failure. Depending on the mode of failure and where it occurs, specific heuristic redesign rules are
invoked which either increase or reduce the thickness of some layers to a certain extent. This new
designisthentestedforfailureandtheprocessisrepeateduntilthesetofthicknessesissuchthata
layer(orlayers)isjustonthepointoffailureattheprescribedloads.
Acorrespondingdesignhavingthesamelayupbutwiththelayerthicknessesroundedupordown
(asrequested)isthenprovided.Alternativeshavingthesamenumberoflayersorlessthanaspecified
layupcanalsobeobtainedascanafulllayerbylayerstressandstrainanalysisforeachdesign.
16.38.4 ESDUpac A9233
16.38.4.1 Input
Propertiesoflayersusedinlaminate:
stiffnesses,
Poissonsratio,
thickness,
allowablestrain.
Initialarrangement:layupandorientationoflayers
16.38.4.2 Output
Allinputdetails
Platelayercompositionofleastthickness
Overallmidplanestrains
Overallplatecurvatures
Strainsatthetopandbottomofeachlayer
Stressesatthetopandbottomofeachlayer.
244
ECSSEHB3220Part2A
20March2011
16.39 ESAComp
ESAComp is software for the analysis and design of composite laminates and laminated structural
elements.
The development work was initiated by the European Space Agency ESTEC, who envisioned open
softwarewhichcombinesallnecessarycompositesanalysisanddesigncapabilitiesunderoneunified
user interface. Although it originated in the aerospace field, ESAComp has been developed as a
generaltoolforpeopledealingwithcomposites,bothinindustryandinresearch.
The core of the ESAComp development work was conducted under an ESA/ESTEC contract by
Helsinki University of Technology, Laboratory of Lightweight Structures and its partners. The first
official release of ESAComp was in 1998. In 2000, the development work was transferred to
ComponeeringInc.(Finland),whichalsoservesasthesoftwaredistributorandproviderofESAComp
supportservices.
ESAComphasavastsetofanalysisanddesigncapabilitiesforsolidandsandwichlaminatesandfor
micromechanical analyses. It also includes analysis tools for structural elements: plates, stiffened
panels,beamsandcolumns,bondedandmechanicaljoints.
[See:Table16.2.3forESACompanalysisfeatures;Table16.2.4forspecialapplicationfunctions]
Owing to the ability to interface with widelyused finite element software packages, ESAComp fits
seamlesslyintothedesignprocess.
[Seealso:ComponeeringInc.websiteforaccesstoprogramanddocumentation]
245
ECSSEHB3220Part2A
20March2011
17
Composite adequate design
17.1 Introduction
Theapplicationoffibrereinforcedplastics,FRP,offersthepossibilitytotailoranoptimisedcomposite
structuretoagivensetofdesignrequirements.Unfortunatelythedesignedstructureisverysensitive
tooffdesignconditionsorimperfections.
Somebasicdesignrulesonhowtoapproachthedesignofcompositestructuresareprovided.
[See:17.5forbasicrules;17.6forfirststeps]
17.2 Anisotropy of composites

Thebehaviouroffibrereinforcedcompositesisdictatedbythedifferentpropertiesintheparalleland
perpendiculardirections,i.e.:
Paralleltothefibres,thepropertiesofthefibresdominate.
Transversetothefibres,matrix(resin)propertiesaremoreimportant.
Forthisreasonthedesignershouldconsidereither:
makeuseofthisanisotropy,byhavingthestrongestmaterialdirectioninthedirectionofgreatest
load,or
reducethedegreeofanisotropy,bylaminatingthematerialinvarieddirections.
Anoptimisedcompositestructurealsoshouldtakeaccountof:
Typeoffibres.
Typeofresin.
Volume fractionoffibres.
Thicknessofeachply.
Propertiesinfluencedbytheanisotropyofacompositeare:
Strengthandstresses.
Elasticbehaviour(Young's modulus).
246
ECSSEHB3220Part2A
20March2011
Poisson's ratio.
Coefficientofthermalexpansion.
Moistureconductivity
Electricalconductivity.
Thermalconductivity.
Figure17.2.1illustratestheanisotropyofCFRP.
Figure 17.2-1 - Anisotropic behaviour of carbon fibre reinforced plastic
Problems arising from anisotropy often occur when metallic design concepts are directly translated
intocompositedesigns.
247
ECSSEHB3220Part2A
20March2011
17.3 Stress-strain relationships
17.3.1 Reinforcing fibres

Fibresdonothaveadefinedyieldpoint,i.e.theyaremostlyelasticuntilfracture,asshowninFigure
17.3.1,Ref.[171].
[Seealso:17.4]
17.3.2 Stress concentrations

Compositesarenotabletoreducelocalstressconcentrationsbyplasticdeformation.Thisleadsto:
highnotch sensitivity
preventionorlimitingresidual stresses.
TheinfluenceofanotchonaluminiumandCFRPspecimensisshowninFigure17.3.2.Thenotched
CFRPspecimenshaveanimportantreductioninspecifictensilestrength,whereasnotchedaluminium
specimensshownosignificantdecreasewhenstaticallyloaded.
Asregardsfatiguestrength,CFRPtestspecimensseemtobelessaffectedbynotchesthanaluminium.
ThefatiguestrengthofnotchedandunnotchedCFRPapproachesthesamelevelafter107stresscycles.
248
ECSSEHB3220Part2A
20March2011
Figure 17.3-1 - Stress-strain response of CFRP (in tension and compression)
Thenotchsensitivityofcompositesneedscarefulconsiderationduringthedesignofstaticallyloaded
structures. Whereas for dynamic loading, composites appear to be less notch sensitive than
conventionalmetals.
249
ECSSEHB3220Part2A
20March2011
Figure 17.3-2 - Fatigue strength of notched and un-notched aluminium and carbon fibre
reinforced plastic
17.4 Fibre strength and stiffness
17.4.1 General
Fibrereinforcedplasticcompositescoverawiderangeofspecifictensilestrengthsandspecifictensile
modulus,asshowninFigure17.4.1,Ref.[172].
Thehighstrengthandstiffnessareonlyavailableinthefibredirection.Thereforefibresareoriented
alongloadpathstomakeuseoftheseproperties.
Definition of the laminate layup (ply angle, thickness) need careful stress analysis, including the
mechanicalandthermalstraincompatibilitybetweeneachlayer.
In aerospace engineering, comparisons between different materials are often made using specific
properties,where:
Property
Specificproperty=
Density
250
ECSSEHB3220Part2A
20March2011
Sogivinganindicationofthelevelofpropertyavailableperunitmassofmaterial.
Figure 17.4-1 - Specific tensile strength and modulus properties of reinforcing fibres
17.4.2 High stiffness applications
17.4.2.1 Glass fibres

ThefrequentlyusedSglassandEglassfibreshavehightensilestrengths,butthetensilemodulusis
comparativelylow.Structureswherehighstiffnessisimperativearethereforeexcluded.
17.4.2.2 Carbon fibres

For highstiffness applications, carbon fibrereinforced plastics, especially those with high modulus
(HM)fibresareused.
[Seealso:2.3,3.3,6.2]
17.4.2.3 Boron fibres

Boronfibresarealsousedforstiffstructures,buttheirhandlingisdifficultandcostsofproductionare
high.Theirusewithinvariousmetalmatrixcomposites,MMCs,enablesuseatoperatingtemperatures
above300C.
[Seealso:46.14,46.15]
251
ECSSEHB3220Part2A
20March2011
17.4.2.4 Aramid fibres
The polyamide fibres, also known as aramid, combine high strength and high modulus but suffer
from low compression strength, e.g. Kevlar 49 (DuPont) and Twaron HM (Teijin; previously
AZKO).
Kevlarisoftensubstitutedforglassfibresinreinforcedplasticsforcertainapplications.
[Seealso:2.3,3.3]
17.5 Basic design rules
17.5.1 General
Inprinciplethedesignofcompositesdoesnotdifferfromthatofmetal,i.e.mostoftheruleswhichare
employedformetalstructuresarealsoacceptableforcomposites.
17.5.2 Aspects of construction
17.5.2.1 Changes in thickness

Where variations in thickness of a laminate are necessary (stiffness changes), they are stepped
gradually;asshowninFigure17.5.1.
Figure 17.5-1 - Shaping changes in thickness
Anabruptchangeinthethicknessactsasastressconcentrationandshouldbeavoided.Suchareasare
susceptibletodelaminationandcanresultinstructuralfailure.
Equation[17.51]isasuggesteddesignruleforshapingthicknesschanges;illustratedinFigure17.5.2:
l t [17.5-1]
where:t=0.7mm
252
ECSSEHB3220Part2A
20March2011
Figure 17.5-2 - Design of shapes
17.5.2.2 Radii, curves and sharp corners

Theminimumradiusofasinglefibremeansthatlayupsarenotabletofollowasharpcorner,i.e.90
angle,withoutaradius.
Figure17.5.3showsthatthelaminatedoesnotfitintothecorner,orovertheedge.
Figure17.53Needforradiioncorners
Sharpcornersleadtostressconcentration,delaminationandfinallycrackingofthelaminate.
Toavoidthisproblem,therecommendationsinclude:
Uselargeradii.Radiiareinfluencedby:
fibretype,
fabricationtechnique,and
curingprocess.
253
ECSSEHB3220Part2A
20March2011
Use recommended minimum radius of curvature for fibres and laminate, as shown in Table
17.5.1.
Table 17.5-1 - Minimum radii for fibres and laminates
GRP: r = 0.2 to 0.5 mm

Fibre
CFRP (HT): r = 0.5 to 1.0 mm
Ri = 1 mm + n 0.1 mm
Laminate
Ri = 4.8 to 6.4 mm
Key: n:numberofplies;r:radius.
17.5.2.3 Openings
Openingsarecutoutsinthestructure,forpurposessuchas:
wiring,
fuelpipes,or
inspection.
Designguidelinesfortheshapeofcutouts,asshowninFigure17.5.4,include:
Asfaraspossible,useroundholesforcutouts(lowerstressconcentrations).
Whererectangularholesareunavoidable,allcornersareroundedwithalargeradiustoprevent
cracks.
The cutout is positioned so that only the smallest possible number of fibres are cut, as
illustratedinFigure17.5.4B.
A=BadC=Good
Figure 17.5-4 - Cut outs: Bad and good practice
Cuttingoutlargerareasofastructuremeansdisturbingloadpaths.Loadsdisturbancecanbeavoided
aroundthecutoutbyeither:
designingamouldedinflangearoundtheopening,or
bybondingonareinforcement.
254
ECSSEHB3220Part2A
20March2011
17.5.2.4 Other design considerations
Avoidundercuts,butthesecanbemadebyusingsplitorrubbermoulds.
Providedraftangleofbetween1and3(generally2)
Mouldedinholesonlyforlargediameter
Practicalthicknesses:
Minimum:0.6mmto0.8mm,becausehandlingbecomesdifficult
Maximum: Unlimited, but only 6mm to 8mm per cure, otherwise a critical exothermic
reactiontakesplace.
Nolimitingsizefactor,exceptmouldsizeandhandlingconsiderations.
If an angle section (or in general open profiles) is designed the spring back resilience of the
structure, produced by different properties parallel and perpendicular to fibres, should be be
considered.
Load transfer by bolted joints: Use of lugs and avoiding drilled holes prevent edge
delamination,stressconcentrationandearlyfailure.
17.5.3 Aspects of laminate lay up
17.5.3.1 Fibre forms

Differentkindsoffibresareavailable,invarioussemiproductformsassummarisedinTable17.5.2.
ThebasicdifferencesbetweentheseformsareillustratedinFigure17.5.5,whichshowsasketchofthe
tensile strength of a single layer for various directions. This determines the preferred loading
direction,i.e.maximisesthestrengthobtainedfromthefibres.
Table 17.5-2 - Basic and semi-product forms
Basic Fibre Form Semi-product

Continuous rovings UD prepreg
Woven fabrics and tapes Bi-directional prepreg
Short fibres Multi-directional prepreg
Key: : Examples: chopped strand mat (CSM), felts; UD : unidirectional
255
ECSSEHB3220Part2A
20March2011
Figure 17.5-5 - Directionality of tensile strength of various semi-product forms:

Preferred loading direction
17.5.3.2 Lay up
If alayupis not well designed, FRPare susceptible to residualstresses. To reacha mechanicaland
thermal compatibility between the plies and to prevent distortion after curing, the design guidance
shouldinclude:
Useasymmetricalstackingofpliesandaminimumofthreedifferentfibreangles,asshownin
Figure 17.5.6.Thisleadsto:
nonwarpingstructures,and
minimumloadingofresinmatrix.
The angle between two stacked plies should be as small as possible to prevent interlaminar
stresses,e.g.
0/+45/90/4545/90/+45/0
betterthan
0/+45/45/9090/45/+45/0
Whenrivetsorboltedjointsareinvolved,alaminatewithabout50%45layersisnecessaryto
havehighbearingstrength,[See:Chapter22].
Compositesofcarbonfibresare:
thermallyisotropicwhenonly0/90pliesareused.
thermallyandelasticallyisotropicwhen0/60pliesused.
If possible split design techniques are employed, as shown in Figure 17.5.7. This enables the
applicationofelementaryprofiles(onlyafewlayersinasimplestructure)whichareassembled
toamorecomplexproduct.
256
ECSSEHB3220Part2A
20March2011
Takeintoaccountthedifferentcoefficientsofthermalexpansionwhenalaminateisplacedinor
onametalmould.Figure 17.5.8summarisesdimensionalchangesfordifferentmaterials.
Figure 17.5-6 - Symmetrical lay-up: Avoids warpage and minimises loading the resin
matrix
Figure 17.5-7 - Split design technique
257
ECSSEHB3220Part2A
20March2011
Figure 17.5-8 - Dimensional change for various materials due to a temperature change
of 100
17.5.4 Fabrication aspects
17.5.4.1 Cost effectiveness

Although material prices, especially carbon fibres, have fallen in recent years, the designer of
compositestructuresshouldconsidercosteffectiveness.
Themainpartsoffabricationcostare:
Materialcosts,
Costofproductionequipment,
Costofassembly,
258
ECSSEHB3220Part2A
20March2011
Qualitycontrol,and
Waste.
Figure17.5.9illustratesthecontributingfactorstothecostoffabrication,Ref.[174].
Key:
Material costs
Cost of production equipment
Cost of assembly
Quality control
Waste
Figure 17.5-9 - Fabrication costs for a CFRP structure
Figure 17.5.10 compares the cost distribution for the fabrication of a typical structural component
designedinmetalorCFRP,Ref.[175].
259
ECSSEHB3220Part2A
20March2011
Figure 17.5-10 - Comparison of costs: Metallic and CFRP
17.5.4.2 Criteria for cost effectiveness

Generalguidelinesforcosteffectivenessinclude:
Construction:
Isthestructuremanufacturedbyoneshottechnique,wheneverpossible?Splitdesign
reduceswastebutraisesassemblycosts.
Ismaximumautomationused?
Material:Reducematerialcostbyoptimumutilisation.
Fabrication:
Isthenumberofcomponentsataminimum?
Hascompositeadequatedesignbeenfullyconsidered?
Chooseoptimumprocessingtechniquestoreducewaste.
Areassemblycostsatminimum?Useasfewjointsaspossible.
Makequalitycontrolsimple.
17.5.4.3 Processing technique

Variousmanufacturingmethodsareused,assummarisedinTable17.5.3,[Seealso:Chapter38].
260
ECSSEHB3220Part2A
20March2011
Table17.53Compositemanufacturingmethods
Moulding
Pressure Hand
Vacuum bag Semi-automatic
Autoclave Automatic lay-up
Resin injection
Filament Winding
Weaving
Automatic lay-up
Braiding
Pultrusion
17.6 First steps in designing a composite
17.6.1 General
A typical example for the preliminary design of composites is a Carpet Plot. This is a graphical
solution based on the classical lamination theory for 0/45/90 laminates and is used for the
determinationofmoduli,strengthorthermalexpansioncoefficients.Carpetplotsareonlyoflimited
use.
More effective methods were developed by S.W. Tsai, Ref. [176]; often available in the form of
software.
17.6.2 Carpet plots

A simplifieddesign method, where it is assumed that a laminatelayup only consistsof0/45/90
plies, is presented in Figure 17.6.1 (modulus) and Figure 17.6.2 (stiffness) for a carbon fibre/epoxy
composite,Ref.[172].
261
ECSSEHB3220Part2A
20March2011
Figure17.61Carpetplotcurvesofmodulusfor0/45/90laminates
Figure17.62Carpetplotcurvesofstrengthfor0/45/90laminates
262
ECSSEHB3220Part2A
20March2011
Thebasisfortheuseofthesediagramsisthatthe0/45/90familyisreasonablyclosetooptimum
withregardtodesign,fabricationandweight.Thediagramsshowhowtomakeafirstapproximation
ofstackingorder,[See:Useofcarpetplots].
Carpetplotsareoflimitedusebecausethebasisforthecalculationisrarelydefined.Thetransitionof
uniaxial strength from unidirectional to multidirectional laminates is often made by over
simplificationoftheory,i.e.itisassumedthatabiaxialstateofstresscanberesolvedinthemannerof
nettinganalysis.Thekindoffailureofthelaminate(firstorlastplyfailure)isnotalwaysdefined.
17.6.3 Use of carpet plots
17.6.3.1 General
TherearetwodifferentwaysofusingCarpetPlots:
Method1,whentheelasticmodulusofthecompositeisknown.
Method2,whenanoptimumstackingisassumed.
CarpetPlotuseisdescribedbyexamplewithFigure17.6.1andusingFigure17.6.2.
17.6.3.2 Method 1
Thisisappliedwhentheelasticmodulusofthecompositeisknown:
Selectapercentageshareof0,45and90plies:
Forexample,[See:Figure 17.6.1]:
ModulusEx=90GPa(13x106psi),andasshowninpointA,apercentagestackingof:
60%0
20%45
20%90
isanarbitrarychoice.
(Toobtainthepercentageof90plies,simplysubtractthepercentageofpliesat0and45from100
%).
Fromthestackingchoicetheallowablelongitudinaltensilestrengthcanbedetermined:
Forexample,[See:Figure17.6.2]:
WiththepercentagestackingofpointAanallowablestrengthx=724MPa(105ksi)results.
Ifthestrengthvalueisnotacceptable,chooseanewpercentagedistribution:
It can be necessary to know optimum stacking, when only modulus Ey is given. If this is true,
interchangethexandycoordinatesand0and90percentagedistributionstoobtaintheallowable
strengthy.
263
ECSSEHB3220Part2A
20March2011
17.6.3.3 Method 2
Thisisappliedwhenanoptimumstackingisassumed:
BywhichExcanbedetermined
40%0
40%45
20%90
AcombinationisshownbypointBandleadstoamodulusEx=69GPa,(10x106psi)
Fromtheassumedstacking,theallowablelongitudinaltensilestrengthcanbedetermined.
WiththepercentagestackingofpointBanallowablestrengthx=586MPa(85ksi)results.
17.6.3.4 Other parameters

Besidesdetermininglayup,modulusortensilestrengthothercharacteristicscanalsobedetermined
byasimilarprocedure,including:
shearmodulus
compressivestrength
Poisson's ratio
coefficientofthermalexpansion
Theseareoftenintheformofacomputervariationprogrambasedontheclassicallaminationtheory.
[Seealso:Chapters11,12,13,14,15and16]
17.7 References
17.7.1 General
[171] E.Heinze&S.Roth:DornierSystemsGmbH
Practicalfiniteelementmethodoffailurepredictionforcomposite
materialstructure.DornierInternalReport,1975
[172] L.M.Schwartz
CompositeMaterialHandbook
McGrawHill,1984
[173] GeorgeLubin(Ed)
HandbookofComposites
VonNostrandReinholdCompany,1982
264
ECSSEHB3220Part2A
20March2011
[174] M.Flemming:DornierSystemGmbH
LecturePaperatTechnicalUniversityBerlin,1980
[175] H.Conen,Kaitatzidis
HhenleitwerkausCFKfrdenAlphaJet
VerarbeitenundAnwendenvonKohlefaserKunststoffen
VDIVerlagGmbH,Dsseldorf,1981
[176] S.W.Tsai
CompositeDesign1986
ThinkComposites,1986
265
ECSSEHB3220Part2A
20March2011
18
Curing stresses: Effects and prediction
18.1 Introduction
Compositematerialsaremanufacturedby:
placingfibreandresinintoorontoamould,then
curingthelayup.
The fibre and resin are normally combined as a prepreg, although some nonprepreg processing
methodsareused.
Boththelaminatelayupandthecurecycleareselectedcarefullytominimiseresidualstresses.
Thischapterdescribeshowresidualstressesoccur,[See:18.4],howtheyarecalculated,[See:18.5],and
howtheycanbereduced,[See:18.6].
18.2 Cure process
18.2.1 Composite materials
18.2.1.1 Thermosetting resin

Compositecuringisaccomplishedbyexposingthelayupforapredeterminedlengthoftimeto:
elevatedtemperature,and
pressure
Itisapplicabletoallthermosettingresinsystems,althoughthepreciseconditionsvarybetweenresin
types.
18.2.1.2 Thermoplastic composites

Thermoplasticpolymersdonotcure,[See:6.16,6.17].
266
ECSSEHB3220Part2A
20March2011
18.2.2 Cure parameters
18.2.2.1 General
The magnitude and duration of temperature and cure pressures have a significant affect on the
performanceofthefinishedproduct.
AtypicalcurecycleforaCFRPcompositeisshowninFigure18.2.1.
Material:Carbon/Epoxy:T300/914
Figure 18.2-1 - Typical cure cycle for CFRP
18.2.2.2 Temperature
The heat applied during the cure provides initiation and promotion of the chemical reaction in the
resin,whichchangesthemolecularstructurebycrosslinking.
18.2.2.3 Pressure
Theappliedpressureprovidestheforceneededto:
squeezeexcessresinoutofthelaminate,
consolidateindividualplies,and,
compressvapourbubblesandvoids
267
ECSSEHB3220Part2A
20March2011
18.3 Analytical notation for residual stress
Aij Laminate stiffness

T
Ni Equivalent stress resultants (due to thermal strain)
Nk Load (where subscript k denotes the loading direction)
Qij Reduced stiffnesses (at the final temperature)
x; y Loading direction

j Mid-plane total strains (measured from the stress-free state)
j
T
Thermal strains (measured from the stress-free state)
i Ply stresses (i-th ply)
i
R
Residual stresses (i-th ply)
18.4 Residual stresses
18.4.1 General
Residualstressesinfibrereinforcedplasticcompositesareproducedduringprocessing.Thestresses
resultwhenthecompositecoolsfromthecuringtemperaturebecauseof:
differentcoefficientsofthermalexpansion,CTE,parallelandperpendiculartothefibre,and
differentCTEofthefibresandmatrix.
18.4.2 Types of residual stresses

Thetypesofresidualstressare:
Micro-residual stresses.
Macro-residual stresses.
18.4.2.1 Micro-residual stresses

Theseareassociatedwiththelocalgeometrywithintheplies.Microcrackingofthematrixleadstoa
decrease in residual bending strains (unsymmetrical laminates), reduced deformation and value of
CTE.Difficultiesalsooccurwhenmoistureentersmicrocracksandthestructureoperatesatcryogenic
268
ECSSEHB3220Part2A
20March2011
temperature.Ingeneral,alloftheseproblemsdependonthepropertiesoftheresin.Abadchoiceof
matrixmaterialcancausecatastrophicdamage.
18.4.2.2 Macro-residual stresses

Thesearecausedbygrossaveragestressthroughthecrossply.Theparameterswhichhaveaneffect
onthemoregeneraldistortionofthelaminateare:
variationinthickness.
stackingsequence.
fibreangleorientation.
Withoutfullconsiderationoftheseparameters,curvatureandwarpofthecuredstructureoccurson
cooling,asaresultof:
differences in the conductivity and coefficient of thermal expansion, CTE, between fibre and
matrix,and
differentthermalanisotropicbehaviourofthefibreitself;especiallycarbonfibre.
Other parts processed using a curing process, such as sandwich materials or block inserts, have a
major impact on residual stresses. To avoid reaching critical stress or to limit shape changes and
springbackoccurring,careisneededtoconsiderfullytheeffectsofsuchresidualstresses.
18.5 Calculation of curing stresses
18.5.1 Residual stresses after curing
18.5.1.1 General
Supposingthatthe0pliesandthe90pliescontractunconstrainedbyeachotherwhencoolingdown
fromthecuretemperature,thethermalstrainsaredefinedasshowninFigure18.5.1.
Since T and
arenotequal,thereisageometricalmismatchbetweenthe0and90plies.
T
x y
x and y areinternallyappliedtothe0
Toensuregeometriccompatibility,theresidualstresses R R
oT (laminatecuringstrain)whichdependson T ,
and90pliesresultinginfinalstrain
andthe
T
e x y
elasticmoduli.
18.5.1.2 Symmetric laminates

Assuming elastic material behaviour and symmetrical stacking, the ply constitutive relations can be
writtenas,Ref.[182],[183]:

1
Q Q
R T T
i ij A jk N K ij j
[18.51]
269
ECSSEHB3220Part2A
20March2011
Figure 18.5-1 - Residual stresses after curing
FromEq[18.51],thetotalplystressesrelatedtotheappliedlaminateresultants,Ni,canbewrittenas:
1
i Qij A jk N K i
R
[18.5-2]
Thetransverseresidualstressesinthepliesafterfabricationaretensile.
270
ECSSEHB3220Part2A
20March2011
18.6 Reduction of thermal stresses and distortions
18.6.1 General
Recommendations to reduce thermal stresses and thermal distortions resulting from cure cycles
include:
Symmetricallaminatestackingisthefirstdesignrule.
EvaluatecarefullythecombinationofCFRPwithothermaterials,e.g.steel,aluminium,because
measureslikeprestrainshouldbecarriedouttopreventcurvatureafterfabrication.
Avoid large variations in thickness because this leads to dissimilar cooling and cracks in the
areasofthicknesschanges.
Fibre angle orientations influence the residual stresses due to curing. An angle ply laminate
using45pliesonlyismostsusceptibletothermalloading.Suchanorientationresultsinhigh
interlaminarstresseswhichaffectthegenerationofmicrocracks,Ref.[18-4].
NOTE The situation is different if mechanical loads are also applied.
Figure 18.6.1 and Figure 18.6.2 show the principal behaviour of various angle ply laminates due to
thermal(curing)andmechanicalloading,Ref.[184].
Figure 18.6-1 - Normalised maximum curing stresses in []s laminates
271
ECSSEHB3220Part2A
20March2011
Figure 18.6-2 - Through thickness tensor polynomial distributions for curing stresses
and stresses at first failure in []s laminates
18.6.2 Stress relieving
18.6.2.1 Annealing
Theannealingofcuringstressesthroughapostcureattheoriginalcuretemperatureisnoteffective.
18.6.2.2 Low temperature thermal cycling

Thermal cycling to a low temperature shows no direct influence on the residual strength of the
laminate. However, it leads to a stress relief caused by microcracking of the resin. Therefore low
temperaturecyclingcanbeusedforunbalancedlaminates,e.g.compositeswithvariableresincontent
withinthelaminateormoregenerallydistortedlaminatesaftermanufacturing.
18.7 References
18.7.1 General
[181] DornierSystemGmbH
Finalreportofastudyonevaluationofinbuiltstressesduring
manufactureofelementsmadeofcarbonfibrereinforcedplastics
272
ECSSEHB3220Part2A
20March2011
ESTECContract2930/75/NL/PP(SC)
TechnomicPublication,1980
[183] R.Y.Kim
Effectofcuringstressesonthefirstplyfailureofcompositelaminates
JournalofCompositeMaterials,Vol.13,January1979
[184] C.T.Herakovich
Onfailuremodesinfinitewidthangleplylaminates
ICCM3,1980,Vol.1
273
ECSSEHB3220Part2A
20March2011
19
Manufacturing faults and service damage
19.1 Introduction
Composite structures, like many engineering structures, need to be able to function effectively
throughoutadefinedlifetimewhilstalsomeetingsafetyandeconomictargets.Structuresareexposed
to a series of events that include loading, environment, and damage threats. These events, either
individually or cumulatively, can cause structural degradation, which can affect the ability of the
structuretoperformitsfunction.
Oneoftheuncertaintiesfortheuserofcompositematerialsishowtodeterminewhetheradefectis
benignornot.Therearetwoseparatetypesofdefectthateffectthestrengthofacompositestructure:
Manufacturing threats: Imperfections or defects from manufacturing of the composite or
assemblyofthestructure.
Servicethreats:Damageresultingduringhandling,installationorinserviceuse.
Understandingthenatureofthedefectpresentenablesboththeeffectofthisdefecttobeascertained
andinsomeinstancesaforensicexaminationofafailedparttobecarriedoutforfailureanalysis.For
everycompositepart,acceptance/rejectioncriteriaareestablishedtobeusedduringinspectionofthe
part. This combines analysis and testing to help identify the effect of a defect on performance of a
compositestructure.Theapproachisverydependentonthepurposeofthestructure,e.g.thepresence
ofmicrocrackingcanbelessimportantforastructuralapplicationbutmoresoforafluidcontainment
application.
SomeexamplesofpossibledefectsareillustratedinFigure19.1.1.
274
ECSSEHB3220Part2A
20March2011
Figure19.11Examplesofdefectsincompositematerials
Thepropertiesofmaterialsandstructurescanbesignificantlydegradedthroughmanufacturingorin
servicedamage.Inthecaseofcompositematerials,thesituationisverydifferentfromthatofmetals.
Unlikemetals,fibrereinforcedcompositematerialsdonotundergoplasticdeformationunderimpact.
A local blow or impact event therefore either results in elastic deformation or damage. The energy
absorbedinanimpacteventdependsonthestrengthofthefibreresinbond,amongstotherthings.If
the bonding is strong, a crack can propagate through the material with minimal deviation. Given a
weakerbond,thecrackpathcouldbemuchmoreirregular,leadingtoextensiveinterfacialdebonding
butconsiderableenergyabsorption.
Thelocationorseverityofmanufacturingdefectsandinservicedamagecanbedifficulttoanticipate
and detect. Some manufacturing flaws are not easily detectable until the structure is exposed to the
service environment, e.g kissing bonds are not detectable until the joint is loaded in service. The
complex loading that astructure experiences inservice and the design of the structure canresult in
loadcasesthatarenotpredictedduringthedesignphase.
In this chapter a general overview of the manufacturing and inservice defects which can occur in
composite materials is presented. Means of preventing or minimising manufacturing faults are
discussedandpotentialinservicedamageeventsarehighlighted.
[Seealso:Chapter8foreffectsofmanufacturingpractices]
275
ECSSEHB3220Part2A
20March2011
19.2 Manufacturing defects in composite materials
19.2.1 General
Manufacturingdefectscanarisefromanumberofsources,including:
Impropercureorprocessing,
Impropermachining,
Contamination,
Mishandling,
Inadequatetooling,
Improperdrilling,
Tooldrops,
Mislocationofholesordetails,
Impropersanding,
Substandardmaterial.
Table19.2.1listsdefectsthatcanoccurincompositematerials,includinghoneycombinterfaceswitha
composite.Eachdefectisalsodescribed.Thetabledoesnotcontainquantitativeinformationbecause
thecomplexityofcompositematerialsproducesahighlyspecificinterdependenceof:
compositestructure
componentdimensions,and
defecttype,orientationandsize.
Consequently,itisimpossibletodefinedetectionlimits.
[Seealso:Chapter34]
Toavoiddefectsarisinginprepregmaterialsuppliedbymanufacturers,consultRef.[194].
Forinspectionandqualityassurance[Seealso:34.2,34.3and34.4].
276
ECSSEHB3220Part2A
20March2011
Table19.21Typesofdefectsincompositematerials
Porosity
Prepreg gaps
Contamination
(solvent, solid, prepreg backing sheet)
Fibre alignment
Lay-up order
State of cure of the matrix
Fibre/resin ratio local variations
Prepreg joints
Inter-ply delaminations
Skin to core debonding
Resin microcracking
Damaged honeycomb core *
Misplaced potting compound *
Edge member/shear connection debond *
Key:*Interfacedwithacomposite.
19.2.2 Description of manufacturing defects
19.2.2.1 Porosity
Porosity is also referred to as resin voids and void content. It usually occurs as a result of the
applicationofpressuretoolateinthecurecycle,causingentrapmentofairorvapour.Theresulting
porosityusuallyoccursthroughoutthelayupratherthanbeingconfinedtoarestrictedarea.
Contamination by dust particles or other foreign objects increases the possibility of void nucleation.
Resinrichzonesarealsopreferredsitesfortheformationofvoids.Ahighpopulationofvoidswithin
the layup causes some reduction in the loadtransfer capabilities of the layup, therefore affecting
resindominatedproperties.
Voidsactasreservoirsformoistureandcanhaveamarkedeffectonthemoistureinitiatedchangesin
theoverallpropertiesofthelayup.Porositycanoccurinfilamentwoundcomponentswhencorrect
proceduresarenotfollowed,forexampleinsufficientresinappliedtotheyarn,overrapidwindingor
incorrectcuring.Thepresenceofthesevoidscancausechangesinthemechanicalandenvironmental
propertiesofthecomposite.
19.2.2.2 Prepreg gaps

Asheetofunidirectionalprepregisproducedbyrollingandflatteningseveralbundlesoffibreswhich
arethenimpregnatedwithresin.Damageorinadequatemonitoringofthisprocesscanresultingaps
in a fibre sheet or, less seriously, disturbances in the fibre direction. These gaps in asdelivered
prepregareidentifiedbyqualitycontrolproceduresbothbytheproducerandtheuser,so,inpractice,
tendnottobeamajorproblem.
277
ECSSEHB3220Part2A
20March2011
19.2.2.3 Contamination
Alaminatecanbecomecontaminatedbytheincorporationof:
dust,[Seealso:Porosity],
oilsorgreases,
solvents,or
solids.
Such contaminants are normally controlled by strict cleanliness procedures. The two types of
contaminationthatneedparticularconsiderationare:
Sharp solid debris: Such as metal swarf which could, if incorporated into a laminate, cause
significantfibrecuttingandlocaldamage.
Prepregbacking sheet:Smallpiecescanbeleftwhentheprepregiscuttoshape.Wherebacking
sheet is present between plies, no bonding occurs, with a consequent drastic local change in
laminateproperties.
19.2.2.4 Fibre alignment

During prepreg manufacture and handling, fibres can become detached from the prepreg and form
either into whorls (during prepreg manufacture) or misoriented fibres (during layup or in the
prepreg).Theseaffectthelocalpropertiesofthelaminateandintroducestressraisersandresinrich
areas which provide preferential voidformation sites. Whorls are coils of fibre that form in the
prepreg sheet. They occur at the end of prepreg production runs and are rarely encountered. Mis
oriented fibres are usually caused by the prepreg tack or stickiness causing fibres to be dragged
duringlayupoperations.
19.2.2.5 Lay up order

Thenumberofpliesandtheirrelativeorientationsaredeterminedbythestipulateddesignstrength
andstiffnessofalaminate.Thesameistrueforthenumberandwindingangleoflayersinafilament
wounditem.Itispossibleforaparticularlayertobeeitherlaidorwoundattheincorrectangletothe
other layers, or even to be omitted altogether. If either of these faults occurs, the properties of the
laminatedonotmeetthosestatedinthedesign.
19.2.2.6 State of cure

Stateofcuredescribesthefinalstateoftheresin:
determinedbycure cycle,and
affectedbyvariationsinresinchemistry.
Bothofthesearesignificantinthefinalpropertiesofthecomposite.Thecurecycleisrarelythecause
of faulty laminate production, because procedures can be readily established and automatically
monitored. These procedures come from extensive development programmes using both non
destructiveandmechanicaltesting.Variationsofresinchemistrywithinasingleprepregsheetcanbe
assumedtobeverysmall;althoughsignificantvariationsbetweenbatcheshavebeenfoundforsome
types of resins. Prepreg properties are usually within the tolerances laid down by the prepreg
supplier. Such tolerances are taken into account when designing with the prepreg. Although the
chemistryofaresinchangesthroughoutstorageprepreglifethemostrapidchangesoccurduring
278
ECSSEHB3220Part2A
20March2011
thecurecycle.Consequentlytoobtainaconsistentchemistrythroughoutfabrication,itisnecessaryto
ensureevencuringintermsoftemperatureandpressuredistribution.
There is no practical nondestructive means of checking the state of cure. Furthermore, destructive
qualitycontrolsamplestakenfromperipheraltestareascannotensureuniformityacrossalaminated
sheet.
19.2.2.7 Local fibre/resin ratio variations

Themanufactureofprepregusestowsoffibresthatarecompoundedwiththeresin.Insomecasesan
uneven distribution of fibres throughout the resin can result. With a sample size of greater than 1
gramme,thisvariationisusuallynogreaterthan2%to3%ofthenominalvalue.
19.2.2.8 Prepreg joints

Thetypesofprepregjointthatcanoccurinanygivenlayerare:
Endtoendjoints,wherethejointisperpendiculartothefibredirection
Sidebysidejoints,whichareparalleltothefibredirection.
Endtoend joints are not advisable, except possibly in the most uncritically stressed area of a
component. There is a sharp discontinuity of stress transfer at the joint and the high mechanical
properties of the fibre are lost. Joints parallel to the fibre direction are much less serious to the
integrityofthefabrication.Theyarearrangedsothatthetwopiecesofprepregbuttexactlytogether:
Anoverlapcausesanextraplytobeformedlocallywithresinrichfillets.
Agapresultsinalocallackoffibresandaresinricharea.
19.2.2.9 Inter-ply delaminations

Delaminationdefectsaredefinedasthefailure,foranyreason,oftwoplieswithinalaminatetobond
together.Themostcommoncauseistheinclusionofprotectiveprepregbackingsheetduringlayup,
which results in a complete disbond or delamination, [See: Contamination]. Another cause can be
entrappedairbubblesformedduringlayupovercomplexcurvedsurfaces,[Seealso:Porosity].
Delaminationsareparticularlydeleterioustoalaminateasthe:
Whole basis of the strength and stiffness of the composite structure is modified, particularly
withrespecttocompressiveloads.
Resistanceofthelaminatetobucklingfailureisseverelyreduced.
Fatiguestrengthofthelaminateislikelytobesignificantlyaffected;thedelaminationisacrack-
likefatigueinitiator.
InaCFRPfacedhoneycombstructure,adelaminationinonefaceskinlocallyreducestheinterlaminar
shearstrengthtozero.Thisintroducesadegreeofasymmetrywhichaffectsthestabilityofthewhole
sandwichstructure.
279
ECSSEHB3220Part2A
20March2011
19.2.2.10 Skin to core debonding
This is a particularly deleterious defect that affects compositeskinned honeycomb sandwich panels.
The mechanical properties of the panel are seriously degraded in the region of the disbond,
particularlywithrespecttobucklingfailure.Adebondcanbecausedby:
Contaminationof,ordamageto,thefilmadhesiveusedtojointheface skinlaminatetothecore
Contaminationofthefaceskinlaminateitself
Contaminationofthebondareaofthecore
Mechanicaldamagetocorecellwalls,bycrushingormorelocaldamage.
Afaceskinbondedtoacrushedcore,suchthatanindentationisformedintheskin,producesashift
oftheneutralaxiswithconsequentchangesinstabilityandbendingstiffness.
19.2.2.11 Resin microcracks

Microcrackscanbecausedduringthecurecycleorbysubsequentthermalcycling.Resinrichzones
arepreferredsitesforthenucleationofcrackswithintheresinbuttheycanbecreatedthroughoutthe
wholecuredlaminate.OwingtothenatureofCFRPcomposites,adifferentialstressisinherentinthe
materialanditisbelievedthatmicrocrackingisthemechanismbywhichthisstressisrelieved.Asthe
resin is a plastic, the rate of loading affects the strain at which the resin breaks. Therefore rapid
temperaturechangesaremorelikelytocausemicrocracksthanslowertemperaturevariations.
Resin starvation can give rise to microcracking in the direction of the fibres. Whilst resin
microcracking is essentially a defect in the matrix, it can have either beneficial or adverse effects
depending on the intended use of the composite. There are instances where the presence of
microcracksprovidesstressreliefinthecomposite,thusprovidingabeneficialeffect,[See:18.6].
19.2.2.12 Damaged honeycomb core

Thetypesofdamagethatcanoccurinthehoneycombcoreare:
Crushing damage, where the cell walls are depressed such that there is a depression in the
honeycomb surface. The result of such a depression is at best a thickening of the bond line
betweentheskinandthecoreand,atworst,atotaldisbondinwhichskin,adhesiveandcore
nevermakecontact
Formingdamage,whenthehoneycombisshapedoveracurvedmandrel.Thedesiredgradual
curvaturesometimesbecomesdisruptedbytheexcessivedeformationofasinglelineofcells.
Thiscausesadiscontinuityinthesurface,whichagaincausesbondlinethickeningorpossiblya
disbond.
19.2.2.13 Misplaced potting compound

Wherefixingsareusedinahoneycombsandwichpanel,thecrushingstrengthofthepanelislocally
increased by filling the honeycomb cells with a potting compound. This is normally done before
integration of the face skins with the core. It is possible that the potted region can be wrongly
positionedsuchthat,intheextreme,thefixingmissesthestrengthenedareaaltogether.Itistherefore
necessarytodeterminethepositionofthepottingbeforefixingtakesplace.
280
ECSSEHB3220Part2A
20March2011
19.2.2.14 Edge member debond
Therearetwotypesofedgemember:
Structural,and
Nonstructural
Structuraledgemembersaredesignedtojoincompositehoneycombsandwichcomponentstoother
partsofthestructure.Thejointbetweenthehoneycombsandwichandtheedgemembercantransfer
theloadsfrombolts,rivetsoradhesivejointsanddistributethemthroughoutthefaceskins.Theshear
componentsoftheseloadsaretransferredfromtheedgemembertothehoneycombcorebymeansof
an adhesive bond lying parallel to the core cell walls. This bond is usually made by means of a
foaming adhesive. Defects in the bond usually involve large voids (or to a lesser extent distributed
porosity)intheadhesivelayer.Thelayercanbeseveralmillimetresinthicknessandseveralhundred
millimetresinlateralextent.Thereisthusthepossibilityforlargevoidstooccur.Thepresenceofthese
voidsaffectstheloadtransferandstressdistributionpropertiesofthecomposite.
Nonstructural edge member are provided for cosmetic and handling protection purposes and are
usedonfreeedgesonly.Commonlyknownasedgecloseouts,theyusuallyconsistofventedtapeor
foamedplastic.Theyhavelittleeffectontheperformanceofthecomposite.Thetapeedgemembercan
reducesurfacequalityandcanhaveadetrimentaleffectonapaintfinish.Thefoamedplasticavoids
boththeseproblemsbutdoesincuramasspenalty.
19.2.3 Detection of defects

Most manufacturing damage is detected by routine quality inspection. Acceptance/rejection criteria
are used during inspection of the part to determine whether it is fit for purpose. Damage that is
deemed acceptable is included in the test and analysis program of the structure to demonstrate
strengthanddurabilitywiththedamagepresent.
Roguedefectsordamagethatissmallerthanthemaximumacceptablelimitsarenotalwaysdetected.
The possible existence and size of such flaws or damage should be assumed as part of the damage
tolerantdesignprocess[See:Chapter33].
19.3 Service threats for composite structures

Besidesdefectsrelatedtomanufacturing[See:19.2],therearealwayspotentialrisksofdamagefrom:
handling,
assembly,
inserviceoperation,e.g.:
lowvelocityimpactdamagefromtooling,
highvelocityimpactbymicrometeoritesordebris,
environmentaldamageinflight,
structuraloverloadoroverstressing,
chemicalcontamination,
thermalandmechanicalfatigue.
281
ECSSEHB3220Part2A
20March2011
The most common installation and inservice damage events can be classified under impact. Such
risks for composite structures can be numerous with a many variables such as the geometry of the
impactor, the energy of the impact event, the frequency of impact events and the environmental
conditions.Theseparametersareoftennotwelldefineduntilservicedataiscollected.
Inservicedamageoftenoccursrandomlyandthetypeofdamage,itslocation,size,andfrequencyof
occurrencecanonlybepredictedstatistically.Inservicedamageistypicallydefinedasnondetectable
(nonvisible) or detectable (visible). Components or structures are designed so that nonvisible
damagecanbetoleratedforthelifeofthestructure.Themostcommoninservicedamageisdueto
impactevents.
Highspeed meteoroids and space debris are a growing hazard for orbiting space missions with the
possibility of hypervelocity impacts. The combination of these impact risks and the wide use of
compositesraisesconcernsfortheoperationalintegrityofcurrentspacecraft.
Thislargenumberofpotentialeventsarefurthercombinedwithaspectsofthestructuresstatus:
Theconditionofthestructureatthetimeoftheevent.
Material properties, e.g. fibre type, fibre forms, ply orientation, matrix material, laminate
configuration.
Theresponseofthestructuretotheevent.
Thehistoryofthestructure.
Together, these can greatly increase costs for validating structural integrity under all potential
conditions.
In composites, the main damage features of hypervelocity impacts are matrix and fibres breaking
undercompressiveandtensileloads,togetherwithpliesdelaminating(orpeeling)undershearloads
The detection of damage during launch preparation is more difficult than in manufacturing, so
damagetolerancemethodologiesshouldincludesuchevents.
[See:Chapter20forfurtherdescriptionsoftheoperatingenvironmentsforspaceequipment]
19.4 Impact behaviour of laminates and sandwich

constructions
19.4.1 General
Damage of composite components and structures due to impact can vary greatly in severity. The
damagemodealsovariesbetweenlaminatesandsandwichconstructions.
Inlaminates,damagecanoccurinthematrixorfibreuptotheleveloffailedstructuralelementsand
the failure of bonded or bolted attachments. The extent of the damage caused by the impact event
determinestheresidualstrengthandfatiguebehaviourandiscrucialtodamagetolerance.
During the impact of a projectile with a sandwich panel, the panel is subjected to structural
deformationaswellaslocaliseddeformationaroundthepointofimpact.
282
ECSSEHB3220Part2A
20March2011
19.4.2 Laminates
19.4.2.1 General
ThevariousformsofimpactdamageobservedinlaminatesareshownschematicallyinFigure19.06.1,
Ref.[193].Includedareexpressionsfortheenergyneededtocausethem,typicalenergyvaluesfor2
mmthickCFRPandthelikelyareaofdamagecaused.
Itcanbeseenfromtheeffectsofgeometrythat:
Delaminations are more likely with short spans, thick laminates or laminates with low
interlaminarshearstrength.
Flexuralfailuresaremorelikelywithlargespansorthinskins.
Penetrationismostlikelyforsmallprojectilesmovingatsuchahighvelocitythatthelaminate
cannotrespondquicklyenoughinflexureandhencegeneratinghighstressesclosetothepoint
ofimpact.
Forlaminates,animpacteventcancausedamagethatincludes:
Fibrebreakage
Matrixdamage
Delaminationanddebonds
283
ECSSEHB3220Part2A
20March2011
Figure 19.4-1 - The primary failure modes of composite laminates under impact loading
19.4.2.2 Fibre breakage

Fibre breakage can be a critical failure mode because structures are typically designed to be fibre
dominant(i.e.fibrescarrymostoftheloads).Inmanyinstancesfailureistypicallylimitedtoanarea
nearthepointofimpact,andisconstrainedbytheimpactobjectsizeandenergy,Ref.[196].
19.4.2.3 Matrix damage

Matrix damage, such as cracking, usually appears at the matrixtofibre interface or in the matrix
paralleltothefibres.Thesecrackscanreducemechanicalpropertiesbutarenotusuallycriticaltothe
structure.Theycanbecomecriticalwhenthematrixdegradationiswidespreadandtheaccumulation
ofcrackingcausesdegradationofmatrixdominatedproperties.
Forfibredominantlaminates,asmallreductioninpropertiesisobservedwhenthematrixisseverely
damaged.However, matrix cracks cansignificantlyreduce characteristics that are dependent on the
resinpropertiesorthefibre/resininterface,e.g.interlaminarshearorcompressionstrength.Thereis
alwaysthepossibilitythatmatriximperfectionscoulddevelopintodelaminations,whichareamore
criticaldamagetype.
19.4.2.4 Delamination and debonds

Delaminations form at the interface between the layers in the laminate and can form from matrix
cracksgrowingintotheselayersorfromlowenergyimpactevents.
284
ECSSEHB3220Part2A
20March2011
Debonds can also develop from poor adhesion between two elements during manufacturing,
initiatingdelaminationinadjacentlaminatelayers.Delaminationsordebondscangrowundercertain
conditions when subjected to cyclic loading and can subsequently cause catastrophic failure if the
laminateisloadedincompression.
Theseverityofadelaminationordebonddependson:
Dimensionsofthedelaminationordebond.
Numberofdelaminationsatagivenlocation.
Location,including:
inthethicknessoflaminate,
inthestructure,
proximitytofreeedges,
stressconcentration,
region,
geometricaldiscontinuities.
Loads thebehaviour ofdelaminationsand debonds depend on the loading type. They have
littleaffectontheresponseoflaminatesloadedintension.Undercompressionorshearloading
however, the sublaminates adjacent to delaminations or debonded elements can buckle
causingloadredistributionandleadingtostructuralfailure.
19.4.2.5 Combinations of damage

Inmanyinstancesimpacteventscausecombinationsofdamage.Forexample,highenergyimpactsby
large objects can lead to broken elements and failed attachments.The resulting damage can include
significant fibre failure, matrix cracking, delamination, broken fasteners and debonded elements.
Damage caused by lowenergy impact is more contained, but can also include a combination of
brokenfibres,matrixcracksandmultipledelaminations.
Most types of impact damage affect mechanical properties. This generally includes significant
reductions in static strength as well as detrimental effects on fatigue properties and environmental
resistance. For load bearing applications, the design requirements are related to the possibility of
impact damage and the potential effects on performance. Significant improvements in impact
performance are being made through the use of hybrid reinforcements and by modifications to
structuraldesigns,amongstotherthings.
[See:19.6fordescriptionsoftheimpactdamagemechanismsandsomeguidelinesondesignrelevant
features]
At first glance the major damage mechanisms (matrix cracking, debonding and fibre failure) can
appear relatively simple when considered in isolation. However, these individual failure modes
combinewithotherfeatures,suchas:
fibretype,
matrixtype,
laminatelayup,
environmentaldegradation,
285
ECSSEHB3220Part2A
20March2011
theconditionofthefibretoresinbond,
loadconditions.
Thiscreatesacomplexsetofparameterstobetakenintoaccountwhenconsideringresidualstrength
afterimpactandpotentialfailure.
19.4.3 Sandwich panels

Impact damage to composite sandwich structures caused by foreign objects can result in significant
reductions in strength, modulus and durability. Hence the damage tolerance of the system can be
compromisedbylocaldamagetothefacings,core,andcorefacinginterface.
In general, the material in the core of sandwich constructions largely dominates the impact
characteristics. The damage size depends on the properties of the core material and the relationship
betweenthepropertiesofthisandthoseofthefacings.
The severity of foreign object impact damage on sandwich composites depends on many factors,
including:
manufacturingprocesses,
facinglaminatelayupandthickness,
corematerialandthickness,
strengthofthebondbetweenfacingandcore,
impactvelocityandenergy,
indentor/foreignobjectshape,
temperatureattimeofimpact,
structuralloadingconditions,
environmentalfactors,e.g.thermalhistory,humidity.
Following impact on sandwich panels with foamed cores, coretofacing debonds can occur around
theimpactsiteandthecorecanbepermanentlydeformed.Cracksinthefoamcorearelikelytoresult
fromlowenergyimpacts.Inhighenergyimpactsthecoreisgenerallycompressedundertheimpact
site.Anunderstandingofthecoretofacingmechanicalpropertiesisimportantinordertounderstand
theeffectofimpactdamageontheresidualpropertyresponsesofsandwichcomposites.
19.5 Impact behaviour
19.5.1 General
Sincestructuresfunctioninenvironmentswheretheycanpotentiallybedamagedbyimpactevents,
regularinspectionsordetectionmethodsshouldbeimplementedsothatthestructurecanberepaired
before catastrophic structural degradation or failure occurs. The design should where possible use
286
ECSSEHB3220Part2A
20March2011
multiplepathsforloadbearingpathssothatotherstructuralmemberscarrysomeoralloftheload
from damaged members. Owing to mass limitations this should be achieved without unnecessary
duplication or overdesign of the structure. This subject is discussed in more detail under design
allowables[See:3.9,Chapter37].
Ifanimpactresultsinvisibledamage,thiscanbeseenduringinspection.Ifitoccursonacomposite
structure,thentheappropriateremedialactioncanbetaken.
[Seealso:Chapter41forrepairs]
19.5.2 BVID
Of significant concern is the damage that could go undetected or is barely visible. This is often
referredtoasbarelyvisibleimpactdamageorBVID.
BVIDcangrowsignificantlyinservicethroughfatigueloadingandenvironmentalconditionsandcan
eventuallyleadtostructuralfailure.Inthistypeofdamage,delaminationsareoftennotvisibleeven
thoughtherecanbesignificantinternaldamagetothelaminate,asshowninFigure19.5.1,Ref.[193].
The damage is often more severe towards the back face of the laminate, making detection more
difficult.
Figure19.51Multipledelaminationsina[(02,45)2]sCFRPlaminatecausedby
dropweightimpact
Thickerlaminatescannotalwaysrespondbyflexingandimpactcausesfrontsurfacedamage,rather
like erosion damage by small particles. If the particle has sufficient energy to penetrate a thick
laminate,thefeaturesareusually:
Frontfacedamagetoadepthofapproximatelyhalfthethickness,and
Delamination in the rear half of the laminate, due to flexural response, as shown in Figure
19.6.2,Ref.[19-3].
287
ECSSEHB3220Part2A
20March2011
Figure19.52Impactdamageina64plyCFRPlaminatecausedbyhighvelocity
impact
Impactstudiesoncompositestructuresincorporatingfeaturessuchasstiffenersshowedthat:
Impact on the middle of the bay needed between 30 Joules and 50 Joules of energy to cause
BVIDbecauseoftherelativelylargeareaofmaterialbeingdeformed.
Near a stiffener, where a flexural response was limited, only a few Joules were needed.
Stiffeners also altered the local response of the panel so that the observable damage was
sometimescentredtoonesideofthepointofimpactorevenawayfromtheimpactsite.
Tominimisecostsinservice,inspectiontechniques,[See:Chapter34]analysisshouldbecarriedoutto
calculateinspectionintervalsforeachpartofastructure.Thisneedsaclearunderstandingof:
damagethreats,
howrapidlydamagegrows,
thelikelihoodofdetectionofthedamage,
theallowabledamagesizesthatdonotthreatenthestructure.
In order to avoid the costs associated with unnecessary repairs, inspection methods also need to
quantifystructuraldegradationtosupportstrengthassessments.
Damagetoleranceistheconceptofcombining:
inspectionplans,
knowledgeofdamagethreats,
damagegrowthrates,
residualstrength.
[See:Chapter33]
Cyclic mechanical testing and analytical methods are used in damage tolerance calculations to
determineinspectionschedulesthatdiscoverdamagebeforeitcausesstructuralfailure.
288
ECSSEHB3220Part2A
20March2011
19.5.3 Impact tests

Traditionally,impacttestssuchIzodandCharpyhavebeenusedtoindicatetheimpacttoughnessof
quasiisotropic materials. They have yielded useful information on notch effects and the ductileto
brittletransitiontemperatures.However,owingtothecomplexityofcompositefailureprocesses,such
testsareonlyoflimitedvalueforcompositematerials.Theygivenoindicationofresidualproperties
afterimpact.
TheIzodandCharpyimpacttestsaretypicallyperformedonplasticandmetallicmaterials.TheIzod
test uses a notched rectangular sample. One end of the specimen is held in a clamp as a vertical
cantilever,impactedbyaweightedpendulumonthesamefaceasthenotchatafixeddistanceabove
the notch and clamp. The energy lost by the pendulum during impact is measured and the Izod
impactstrengthisthencalculated.
The Charpy test also uses a notched rectangular bar and the specimen is supported as a horizontal
simple beam and is broken by the pendulum in threepoint loading. The impact site being midway
betweenthesupportsanddirectlyoppositethenotch.Thetestprocedurecanbeusedontestsamples
with differing notch radii to study the notch sensitivity of the material. The unnotched impact
strengthofthematerialcanalsobeassessed.
However, neither the Izod or Charpy tests are particularly appropriate for use with continuously
reinforcedcompositematerials.
Dropweighttests(simulatingdroppedtools)andballistictests(simulatingforeignobjectimpact)are
morecommonforcomposites.ImpactusingfallingweighttestingispartoftheCAIcompressionafter
impacttestwhereaflatpanelisimpactedat90toitssurface.Typicallya12.7mmto25.4mmdiameter
hemisphericalindentorisused.ThefallingweightusedforCAItestingisusuallydroppedfromafew
metres giving a relatively low velocity impact and the mass used is typically 5kg to 10kg. Other
systemsforlow velocity tests includependulum systems and servohydraulic actuators. Ifvery low
velocityimpactsaretobestudied,alongfulcrumpendulumcanbeusedtoimpactthetestsamples.
Thevelocitiesassociatedwithhighspeedmeteoroidsandspacedebrisarenotsimulatedinthistype
oftest.However,testrigsareavailablethatincludesecondaryacceleratingmechanismssuchaselastic
cords, springs or pneumatic systems to obtain higher velocities as well as highpressure pneumatic
systemstoobtainballisticvelocities.
Quasistaticindentationtestscanalsobeperformed.Thesetypicallyinvolvemountingatestsample
(usuallyaflatplate)inasupportframeandtestinginauniversaltestingmachine.Anexampleofsuch
atestisgiveninASTMD6264.Theloadandcrossheaddisplacementistypicallymeasuredduringthe
test.Thetestisstoppedwhenapredefinedlevelofdamageorcrossheaddisplacementisreached.The
depthoftheindentationandtheextentofthedamagearethenevaluated.
Other programspecific tests can be included in a test schedule to more closely simulate the actual
eventsexpectedinthelifetimeofthestructure.
Whichever impact test method is used, an assessment of the damage is performed after impact and
beforefurthertestingsuchastheCAItest.
Itisimportanttounderstandthenatureandsizeoftheimpactdamageinordertobeabletoobtain
accuratedatafordamagetoleranceevaluation.Damageassessmentcaninclude:
measurementofthedepthoftheindentation.
measurementofthesizeofthevisibledamage.
nondestructiveevaluationforinternaldamageorBVID.
289
ECSSEHB3220Part2A
20March2011
Oncethishasbeenevaluated,additionalmechanicaltestscanbeperformed.
19.5.4 Compression after impact (CAI)

The CAI compression after impact test is an evaluation of the reduction in laminate compressive
strengthafteranimpactevent.Itisoftenreferredtoasadamagetolerancetest.
The CAI tests were developed to enable candidate composite materials to be assessed for their the
damage tolerance [See also: Chapter 32]. The impact event before CAI testing is generally at a
relativelylowvelocityanditisnotcommonlyusedtoinvestigatedamagetolerancetoballisticevents
Ref.[195].
19.6 Detection of defects

[Forrepairofdefects,See:Chapter41]
19.6.1 Damage detection techniques

Delaminations can be detected by ultrasonic techniques. Cscans give a silhouette of the damaged
areas.AscansorBscansprovideinformationonthedepthofthedamage.
Complexultrasonicbasedinspectionsystemsareusedtoinspect aerospacestructures,suchaslarge
subassembliesoftheJointStrikeFighteraircraft,Ref.[195].
Useful information on impact damage can also be obtained with Xrays, but the technique uses a
radioopaquedyetopenetratethedamagedregion,whichisnotalwaysdesirable.
Therearenowawiderangeofinspectionsystemsbasedonavarietyofprinciples;eachperformsa
differentfunction.Inbroadterms,thetechniquescanbegroupedasbeingapplicableto:
laboratory (L) and production (P) environments for the examination of components and
assembledstructures.
inservice(S)environmentsfortheexaminationoffullyassembledstructures.
[Seealso:Chapter34]
19.6.2 Laboratory and production based NDT

Forcomponentsandassembledstructures,thetechniquesavailableinclude:
Portableconventionalandaircoupledultrasonics,[See:34.10].
Eddycurrents,[See:34.18].
Lasershearography,[See:34.14].
Thermography,[See:34.15].
19.6.3 Other techniques

Techniquesthatprovideasupportingroleinclude:
290
ECSSEHB3220Part2A
20March2011
Visualinspection,[See:34.9].
Cointapping,[See:34.9].
Dyepenetrants,[See:34.9].
Resonancebondtesters,[See:34.9].
Acousticflawdetectors,[Seealso:34.20]
Acousticemission,[See:34.11].
Increasedaccesstocompactandpowerfulcomputinghasencouragedthedevelopmentofinspection
systems that are integrated into structures to provide continual monitoring of thermal and strain
responses. These systems are described as structural health monitoring, [See: Section XX], and can
alsobecapableofinservicedamagedetection.
19.7 References
19.7.1 General
[191] FulmerResearchLaboratories(UK)
Guidelinesfornondestructiveexamination(NDE)ofadvancedfibre
compositematerialsforspaceapplications
ESAContract4389/80/NL/AU(SC)
[192] R.J.Lee&D.C.Phillips
Thedamagetoleranceofhighperformancecomposites
CompositeStructures1,AppliedSciencePublishers,1981
[193] G.Dorey:RAE(UK)
Impactdamageincompositesdevelopment,consequencesand
prevention
[194] ESAPSS03207:Guidelinesforcarbonandotheradvancedfibre
prepregprocurementspecifications.
NotcurrentlyundertheECSSdocumentsystem.
[195] Privatecommunication(2003)
USLUltrasonic Sciences Ltd.(UK)
[196] MILHDBK17
19.7.2 ASTM standards

[See:ASTMwebsite]

ASTMD6264 Standardtestmethodformeasuringthedamage
resistanceofafiberreinforcedpolymermatrixcompositetoa
concentratedquasistaticindentationforce
291
ECSSEHB3220Part2A
20March2011
20
Environmental aspects of design
20.1 Introduction
The Earth and space environments to be experienced by a spacecraft have to be considered fully
duringthedesignprocess.
Space structures are manufactured, stored and transported on Earth prior to launch, hence a
controlledenvironmentisnecessarytominimisecontamination,e.g.bymoisture,whichcandegrade
materialsorpromotegalvaniccorrosionbetweenCFRPandcontactingmetals.
Inspace,theconditionsexperiencedcandegradethepropertiesofpolymercomposites.Theextentof
property loss depends on the duration of exposure, as well as the severity of the environment.
Problems of this type are more significant for longterm deployed structures; especially when the
designlifeextendstowards30years.
In addition to their widespread use in GEO geostationary earth orbit environments, composite
structures are increasingly destined for use in LEO low earth orbit. Here they are subject to
aggressive attack by atomic oxygen (ATOX). Longlife composite structures can need specific
protection against the more aggressive occurrences, e.g. from ATOX, impacts from debris and
meteoroids.
Theeffectsofenvironmentsonthepropertiesofcompositematerialsaredescribed,[Seealso:20.2].
Design guidelines are provided which aim to avoid common environmentalrelated problems and
appropriatematerialsrequirementsstandardsarecited,e.g.ECSS.
For information on materials and process selection for space applications, [See also: ECSSQST70;
ECSSQ7071].
20.2 Description of environments
20.2.1 Earth environment

In general, space structures are manufactured and then spend a certain amount of time on Earth
beforebeinglaunchedintospacetoencounterdifferentspaceenvironments.
The effects of temperature, moisture and various forms of organic attack can be controlled by
appropriate materials selection, adequate design and by ensuring that transportation and storage
takesplaceunderwelldefinedconditions,Ref.[2043].
Exposuretomoistureispossibleduringstorageandintegrationofthelauncherandpayload.Moisture
contaminationcanalsooccurifalaunchisaborted.
292
ECSSEHB3220Part2A
20March2011
Materialsselectionisthereforeinfluencedbytheneedtocontrol:
Moistureabsorptionfordimensionallystablestructures,
Galvaniccorrosioninthepresenceofmoistureorfuels,
Stresscorrosioncracking,SCC,oflight alloys.
20.2.2 Space environment

Afterlaunch,astructureissubjectedtoavarietyofenvironmentaleffects,suchas,Ref.[201]:
temperature,
vacuum,
radiation,
micrometeoroids,and
debris.
ThetwomostcommonEarthorbitsare:
LEOlowEarthorbit,and
GEOgeostationaryEarthorbit.
It is important to differentiate between LEO and GEO orbits, because the environments differ, [See:
20.3;20.4].
Someoftheeffectsofspaceenvironmentalfactorsaresummarisedin:
Figure 20.2.2forradiation,Ref.[20-1].
Figure 20.2.3fortemperature,Ref.[20-1].
Figure 20.2.4formicrometeoroidsanddebris,Ref.[20-1].
Figure 20.2.5foratomicoxygen(ATOX),Ref.[20-1].
Figure 20.2.6forreentry,Ref.[20-1].
NOTE Not all of these environmental factors relate to composites.
The space environments are not overly aggressive, but materials are evaluated to ensure adequate
performanceunderavarietyofcombinedconditions.
Such data is sparse, so specific testing is necessary. Despite this, reliable configurations for long
durationmissionsarefeasible.
293
ECSSEHB3220Part2A
20March2011
Figure 20.2-1 - Space environment effects: Vacuum
Figure 20.2-2 - Space environment effects: Radiation
294
ECSSEHB3220Part2A
20March2011
Figure 20.2-3 - Space environment effects: Temperature
Figure 20.2-4 - Space environment effects: Micrometeoroids and debris
295
ECSSEHB3220Part2A
20March2011
Figure 20.2-5 - Space environment effects: Atomic oxygen (ATOX)
Figure 20.2-6 - Space environment effects: Re-entry
296
ECSSEHB3220Part2A
20March2011
20.2.3 Composite structures
20.2.3.1 General
Thesignificanceoftheeffectsofoperatingenvironmentsvarieswiththefunction,orbitanddesignlife
ofdifferentstructures.Thebroadclassificationsofstructuresandenvironmentalconcernsare:
Structuralloadbearing;guaranteeingthemechanicalintegrityofthecomposite,particularlyfor
manratedstructures.
Dimensionallystable,includingthestabilityofphysicalproperties,e.g.CTE.
[Seealso:ECSSQST70andECSSQ7071]
20.2.3.2 Temperature range

Temperature changes have significant effects on structures because of the wide range typically
experienced. This temperature range is more severe in GEO, compared with LEO, although in LEO
the cycling is more rapid giving a greater number of cycles in a given period. Additional cooling
effectscanarisefromstoredcryogenicfuelsorhelium.
Although there are ways of controlling the temperature range experienced by satellites, thermal
cyclingremainsasignificantproblem,[Seealso:20.9].
At low temperatures, the resin matrix can suffer from embrittlement, and at high temperatures the
resin can soften and creep. In both cases, strength is reduced and the elastic properties of the
compositearechanged.
Microcracking damage, caused by thermal cycling, remains a major problem, as this modifies the
thermalexpansion(CTE)characteristicsofthecomposite.
20.2.3.3 Elevated temperature and vacuum

Underthesecombinedconditions,thematrixresinoutgasestosomeextent,causing:
condensation of volatile organic species on optical equipment, which can seriously degrade
theirperformance,
degradationofthemechanicalpropertiesofthecomposite.
Thesecanbecontrolledbythecorrectselectionofmaterials,[Seealso:20.10;ECSSQST70andECSS
Q7071].
All polymeric composites have a maximum permissible operating temperature, which takes into
accountthedegradationmechanismsthatshouldbeavoidedorminimisedinservice.
20.2.3.4 Ultraviolet radiation

UV radiation affects the physical properties of resins by causing increased crosslinking in the
polymer. This produces a higher strength and stiffness, but also makes the resins more brittle. UV
effectsareusuallyinsignificantforcompositesastheyaresurfaceeffects;thefibresscreenthebulkof
297
ECSSEHB3220Part2A
20March2011
the resin matrix. Vacuum ultraviolet (VUV) degradation is a problem for thin polymeric films and
ultrathincompositesections,[Seealso:20.11].
20.2.3.5 Penetrating radiation

Theeffectsofpenetratingradiation,i.e.highenergyprotonsandelectrons(and particles),are
considered to be largely insignificant for mission durations less than 7 years. It is recognised that a
thresholddosageisreachedbeforedegradation,intheformofbreakingofmolecularchains(scission),
occurs in a composite matrix. For GEO, in particular, this threshold is reached before 30 years
exposure,[Seealso:20.11].
20.2.3.6 Meteoroids and debris

Generalstatementsontheeffectsofmeteoroidsordebrisimpactareimpossible,becausetheydepend
onmanyparameters,including:
Mission specific factors, e.g. orbit, duration, required safety and orbit height, inclination and
eccentricity.Therequiredsafetydependsonthemannedorunmannedrating.
Impacting particle, e.g. velocity, size, density, impact angle and shape. Sharp, angular
projectiles are more damaging than spherical particles. Shape is difficult to account for in
impactmodellinganddesignequations.
Protectionprovidedbytherestofthestructure.
Assessmentofthedamagecausedbysuchimpactsisunderreview.Theproblemisparticularlyacute
inLEOandforlarge,longduration,highsafety(e.g.manned)structures,[See:20.16].
20.3 Low earth orbit (LEO)

The environmental conditions inLEOareimportant to longlifestructures, e.g. the ISSinternational
spacestationandColumbus.
Vehiclesoperatinginorbitsbetween400kmand800kmaltitudeencounter:
Temperaturechanges:
typicallywithintherange90Cto+90C,
~16cycles/day.
Vacuum:intherangeof104torrto109torr.
Radiation:
principallyvacuumultraviolet(VUV),
someparticleradiation(98%electron,2%proton,fromVanAllenbelts).
Atomicoxygen(ATOX).
Meteoroidordebrisimpact.
Ofthese,atomicoxygenanddebrisimpactarepotentiallythemostdestructivemechanisms.Thiswas
confirmedbyexperiencefromtheLDEFprogramme,[See:20.8].
298
ECSSEHB3220Part2A
20March2011
For CFRP composite space structure elements operating in LEO, the polymeric matrix is the
degradationvulnerablephase.
Degradationcanleadtosignificantchangesinmechanicalproperties,CTEand,possibly,thethermal
conductivityofthecomposite.
20.4 Geostationary orbit (GEO)

The environmental conditions in GEO are important to longlife telecommunication and Earth
resourcessatellites.
Vehiclesoperatinginorbitsaround36000kmaltitudeencounter:
Temperaturechanges:
typically150Cto+120C.
1cycle/day.
Vacuum:intherangeof109torrto1010torr
Radiation:
Trapped Van Allen belts, which contribute to the total radiation spectrum and are
influencedbythemagneticfieldoftheEarth.
Additional radiation sources include: galactic radiation, particle radiation from solar
flares,butverylittleultraviolet.
Meteoroidordebrisimpact.
ThepreciseorbitconditionsdependonwhetherastructureisinGEO,GTOorpolarorbit.
Thermal cycling and radiation damage are potentially the most damaging mechanisms for long
durationmissions.
20.5 Deep space exploration

Deepspacemissionshavedifferentlevelsofenvironmentalinteraction,althoughthebasicphenomena
oftemperaturechange,radiation,vacuumandmicrometeoroidsremain.
Othermissionspecificconditionswhichcanbeencounteredinclude:
Cometdebris,e.g.GiottomissiontoHalleyscomet.
Planetaryatmospheres:e.g.Mars,Saturn,Jupiter.
hightemperatures,with
gaseousreactivechemicalspecies.
Cryogenictemperaturesintherangeof180Cto260C
Vacuumto1014torr:
299
ECSSEHB3220Part2A
20March2011
Wherepolymercompositesarenotexposeddirectlytotheseenvironments,i.e.theyformpartofan
underlyingstructure,aprotectiveshieldingcanstillbenecessary.
20.6 Galvanic corrosion
20.6.1 General
Galvaniccorrosioncanleadtothecatastrophicfailureofastructureandcanoccurwherecarbonfibres
areusedinmixedmaterialconstructions.
Someofthepossiblelocationsforgalvaniccorrosiontooccurinclude:
CFRPtometalboltedjoints,e.g.aluminium
Sandwichstructures:
directcouplingofCFRPtoaluminium,e.g.honeycombcore.
contactbetweentheface-skinandcoreorfaceskintoinsert.
PoorlybondedjointsbetweenCFRPandmetal,e.g.aluminium
Contactbetweendissimilar metals.
[Seealso:Chapter44formagnesium;Chapter46foraluminium]
Most of the research into this problem has been carried out under typical aircraft structure
environments.Thereforetheextentandconsequencesofgalvaniccorrosioninthespaceenvironment
needdefinition,Ref.[202].
[Seealso:ECSSQST70andECSSQ7071]
20.6.2 Physical basis of galvanic corrosion

Galvanic corrosion occurs when two or more dissimilar materials are electrically connected, in the
presenceofacorrosiveenvironment.Agalvaniccellisformed,andthemoreanodicofthematerials
sacrificiallycorrodesaway,leavingthecathodicmaterialunaffected.Therateofattackcanbemuch
higherthanthatofthesinglematerial.
Foragivenenvironmenttheseverityofthecorrosioniscontrolledbythegalvaniccellstrength,which
isgivenbytheelectrochemicaldifferencebetweenthematerialsinthegalvanicseries.
Figure 20.6.1 shows an example of a galvanic series for some common materials in a 3.5% NaCl
solution,Ref.[202].Theactualelectrochemicalpotentialvaluesdependonthedetailsofthecorrosive
environment.
300
ECSSEHB3220Part2A
20March2011
Figure 20.6-1 - Galvanic corrosion: Potential in 3.5 % NaCl solution
EXAMPLE,fromFigure20.6.1:
Potentialdifference:CarbonfibreAluminiumU=0.7V
Inpractice,agreatmanyphysical,chemicalandenvironmentalvariablesplayapartindetermining
therateofgalvaniccorrosionforanyparticularsituation.
20.6.2.1 Carbon fibre to metal connections

Whencarbonfibrecompositesareinelectricalcontactwithmetallicstructures,inthepresenceofan
electrolyte,agalvaniccelliscreated.Thehighlycathodicnatureofcarbon,withrespecttostructural
metals,resultsincarbonbecomingthecathodeand,forexample,aluminiumtheanode.Underthese
conditions,thealuminiumcorrodes.
301
ECSSEHB3220Part2A
20March2011
20.6.3 Prevention of galvanic corrosion in space structures
20.6.3.1 General
No specific studies have been conducted for the space environment, consequently, only general
guidelinesaregivenforunmannedstructures,[See:ECSSQST70andECSSQ7071].
20.6.3.2 Guidelines
Somegeneralguidelinesforavoidinggalvaniccorrosioninunmannedspacestructuresinclude:
Maintain a controlled environment (especially with respect to humidity) during manufacture,
transportationandstorageonEarth.Preventionoftheinitiationofanelectrolyticcellunderthe
specialconditionsofthespaceenvironmentensuresthatgalvaniccorrosiondoesnotbecomea
problem.
After launch, outgassing as a result of UV radiation and contamination by organic materials
(such as hydrazine) can provide the electrolyte for a galvanic cell between carbon fibre and
metals.
Corrosioncontrolfactorsinclude:
insulation,e.g.specialcoatingstointerruptthegalvaniccell,
reductionincathodearea,thusreducingthedrivingcorrosioncurrent,
useofcorrosioninhibitors,
eliminationofmoisture,
minimisingdifferencesingalvanicpotentialbetweenjoinedelements.Usematerialswith
aminimumpotentialdifferencetocarbon(lessthan0.5V,ifpossible).
[Seealso:ECSSEHB3223foradviceonavoidinggalvaniccorrosioninmechanicallyfastenedjoints]
20.7 Effects of moisture on composites
20.7.1 General
Themainfactorsthatinfluenceaspectsofdesignare:
Changes to coefficients of thermal expansion(CTE)causedbyoutgassing.
Low coefficient of moisture expansion(CME)resins.
Combined hot/wet performance.
[Seealso:Chapter13fordetailedinformationonmoistureeffectsincomposites]
20.7.2 Modification of CTE by outgassing

Absorbedmoisturereadilyoutgasesinorbit.Itslossreducesthevolumeofthematrixandlowersthe
inorbitCTEcomparedwiththatundercontrolledstorageonEarth.
302
ECSSEHB3220Part2A
20March2011
Formanyapplications,dimensionalchangesaresosmallastobeinsignificant.However,dimensional
changes resulting from moisture content variations can cause problems where a configuration is
calibrated on Earth and then placed in orbit. This problem is more severe with aramid composites
thanCFRP,becausearamidfibresabsorbmoisturewhereascarbonfibresdonot.
20.7.3 Low coefficient of moisture expansion (CME) resins

To counter the changes in physical properties caused by moisture absorption, low absorbency resin
matricesarebeingconsideredfordimensionallystablestructures,Ref.[203].
Theoldergenerationofresins,suchasFiberite934epoxy,arebeingsupersededbyresinswhichhave
alowerCMEvalue.
ThechemistrytypesoflowCMEmatricesare:
Thermosetting:
Epoxy:e.g.Ciba/BrochierM18,Cytec9501,CytecFiberiteCycom5555.
Cyanateester,e.g.YLARS3andFiberite9543.
Thermoplastics,e.g.PEEK,PEI.
Thermoplasticsarenotwidelyappliedinspacestructures,largelybecauseofexistinginvestmentsin
thermosetprocessingtechnologiesandthelowvolumesofmaterialusedinthespaceindustry,[See
also:40.7].
[Seealso:2.4;6.35forlowCMEresins;Chapter13formoistureeffectsonpolymercomposites]
20.7.4 Hot/wet performance
20.7.4.1 Epoxy resins

Inaircraftconstruction,earlycompositeshadacceptablethermalperformancebuttheepoxymatrices
were brittle. In prepreg systems (from the mid1980s), toughening additives were introduced to
provideimprovedimpactresistance.Theresinformulationsalsohadtoretaintheuppertemperature
performancewhenthecuredresinwasinafullysaturatedstate(moistureabsorption).Examplesof
theseimprovedhot/wetepoxyresinsystemsincludeFiberite9772andFibredux924.
20.7.4.2 Cyanate ester resins

In addition to epoxies, cyanate esters are promoted to perform the same hot/wet role, e.g. Fiberite
9542andCycom785.Thesesystemstendnottobeappropriateforusewithindimensionallystable
structures because they are not available in a prepreg form with UHM carbon fibres. They can be
appropriate for launchersand reusablestructures where thermalexcursions are within the 130C to
200Crange.
20.7.4.3 Others
Goodhot/wetperformancecanalsobeobtainedfrombismaleimidesandsomethermoplastics.
303
ECSSEHB3220Part2A
20March2011
20.8 LDEF in LEO
20.8.1 Mission
Between 1984 and 1990, the long duration exposure facility (LDEF), a twelvesided cylindrically
shapedspacecraft,wasplacedinLEOfor69months(5.75years)ThecraftwasbuiltintheUSAand
wasabout9mlong,4.3mindiameterandhadatotalmassof10tonnes.Themainstructurewasmade
of6061T6aluminiumalloywithamassof3.6tonnes.ItwasplacedinorbitandretrievedbySpace
Shuttlemissions.Over10,000materialsampleswerecarriedonthe130m2spaceexposedsurface.The
originalmissionwasforoneyearbutwasextendedafterthelossofChallenger.Duringthe5.75years,
theorbitdecayedfrom480kmto340km.
The onboard experiments experienced harsh LEO conditions of combined vacuum, atomic oxygen
(ATOX), ultraviolet radiation (UV) and thermal cycling. Significantly higher ATOX concentrations
were experienced in the last year as the craft neared the Earths atmosphere; some 80% of the total
exposure.Thermalcycles(32,422at~16/day)wereenduredbyallmaterialsandexperiments.
The LDEF programme used the abbreviation AO to indicate experiments related to atomic oxygen,
whereasATOXistheacceptedabbreviationwithintheEuropeanspaceindustry.
LDEF was retrieved a few months before its decayed orbit resulted in reentry and burnup in the
Earthsatmosphere.
20.8.2 Materials and experiments
20.8.2.1 General
The prolonged exposure to the LEO environment has provided extensive data on the ability of
materialstowithstandtheharshconditions.Thematerialsexposedincluded:
Structuralmaterials,i.e.lightalloysandcomposites,
MMCmaterials,
Polymerfilms,
Adhesives,
Lubricants,
Seals.
The selection of the materials flown on LDEF was made in the late 1970s and early 1980s. The
products chosen were those favoured by NASA and US aerospace companies at that time. Whilst
someofthesearenolongercommerciallyavailable,theresultsreportedindicatetheperformanceof
genericmaterials,inparticularpolymercomposites,suchasCFRP.
20.8.2.2 Material performance

The detailed evaluation of results from LDEF continued some 5 years after retrieval. However, the
overall conclusions indicate that considerable care is necessary when deploying polymers and
polymercompositesdirectlyinLEOenvironmentsbecauseofseveredegradation.
Thedamagemechanismsrelatetospecificoraccumulateddegradationarisingfrom:
304
ECSSEHB3220Part2A
20March2011
Surfaceerosionbyatomicoxygen(ATOX).
Prolongedthermalcycling.
Impactdamagefrommicrometeoroidsanddebris.
Embrittlementbyradiation.
The material performances observed are significant for intended space station operation. For such a
construction, a 30year design life is typical, i.e. equivalent to 175000 thermal cycles, some 5 times
longerthanLDEF.
20.8.3 Variations in exposure conditions

ThepreciselocationofexperimentsonLDEFdeterminedthelevelofexposuretotheprinciplemodes
ofdegradation,i.e.:
ATOXflux,
VUVdosage,and
Impactparticleflux.
The orientation of the sides of LDEF with respect to forward motion and the angle to incoming
radiation means that comparison between different experiments included compensation for the
differentexposurelevels,i.e.positiononthecraft.Thatsaid,theinformationgainedwasconsiderable.
The different participating organisations assembled experiments covering a wide range of carbon,
glass and boron fibrereinforced composites with thermoset and thermoplastic matrices; as
summarisedinTable20.8.1,Ref.[205].
305
ECSSEHB3220Part2A
20March2011
Table 20.8-1 LDEF: Polymer composites
Row Angle off ATOX flux VUV ESH Experiment

Organisation Materials
No. RAM x1021 a/cm2 x103 No.
AO 134 NASA Langley 5208/T300, 934/T300, P1700/C6000, P1700/C6000, 930/GY70
CE-339/T300, F263/T50, 934/T50, X904B/T50, E788/T50, 3501-
9 8 8.32 11.1 M0003-9 Lockheed 5A/HMS, E788/C6000, 934/HMF176, CE-339/E-glass, F593/P75,
934/P75
M0003-8 Boeing 934/T300, P1700/T300, PMR15/C6000
10 22 7.78 10.7 AO 054 TRW Epoxy/E-glass
AO 171 NASA-MFSC 934/HMS, 934/P75S, P1700/HMF, Epoxy/S-glass
M0003-10 Lockheed X904B/GY70, 3501-A/HMS, X904B/E-glass
X-30/GY70, CE-339/GY70, CE-339/P75S, 934/P75S, 934/GY70,
8 -38 6.63 9.4 M0003-10 General Dynamics
P1700/W722, V378A/T300
M0003-10 McDonnell Douglas 5208/T300, P1700/T300, PES/T300, PI/C6000
M0003-10 Boeing 934/T300, 3501-6/AS, P1700/T300, PMR15/C6000, LaRC 160/Gr
7 -68 3.16 7.2 AO 175 Rockwell F178/T300, PMR15/C6000
University of
AO 180 934/T300, 5208/T300, SP288/T300, SP328/K49, SP290/Boron
Toronto
12 82 1.2 6.9
University of
AO 019 5208/T300 interleaved with Kapton
Michigan
1 112 0.06 7.5 AO 175 Rockwell 934/T300, LaRC 160/C6000
5 -128 0 8.2 P0005 Morton Thiokol C/C and sundry polymer composites
4 -158 0 10.4 AO 054 TRW Epoxy/E-glass
3 172 0 11.1 AO 138 Matra Espace 934/GY70, V-108/K49, V108/T300, V108/GY70, V108/G837
Key: AO and ATOX - Atomic Oxygen VUV - Vacuum Ultra Violet ESH - Equivalent Sun Hours
306
ECSSEHB3220Part2A
20March2011
Figure20.8.1showshowtheexposureconditionsvarieddependingontheexperimentlocation,Ref.
[204].Table20.8.2showshowthistranslatedintoexposurelevelsforselectedexperiments,Ref.[204].
Thedirectionofcraftmotion(ram)isusedasareferencedirection.Theleadingedgeexperiencedthe
highest atomic oxygen ATOX exposure (impacts/cm). The ATOX exposure diminished with the
relative angle of incidence to ram and the (12sided) angle of craft surface. The trailing edge
experiencednoATOXexposure.
Figure 20.8-1 LDEF: Experiment locations (M0003-10)
Table 20.8-2 - LDEF: Exposure levels
Leading Edge Trailing Edge Leading and

A-Deck A-Deck Trailing Edge
Row 8 Row 4 B-Deck
ATOX Exposure
6.93 x 1021 9.32 x 104 0
(Impacts/cm2)
Incident Solar and
Earth Reflected 9300 10500 0
Radiation (ESH)
Thermal cycling (as -47 to +84C (32422 -33 to +77C (32422
Not known
measured) cycles) cycles)
20.8.4 Composite materials aboard LDEF

A rsum of all of the polymer composites flown on LDEF is given in Table 20.8.3, Ref. [205]. The
compositeswerepreparedinmanydifferentlaminateconfigurations.CompositesfromMatraEspace
(France)weretheonlyEuropeansourcedmaterialsincluded,butwerenotexposedtoATOXbecause
theywerelocatedonthetrailingedge.
307
ECSSEHB3220Part2A
20March2011
Table20.83LDEF:Compositematerialclassifications
Composite Class Matrix/Fibre
5208/T300,
934/T300,
First generation epoxy CFRP/high strength fibre 3501-6/AS,
SP288/T300,
V108/T300.
934/T50,
First generation epoxy CFRP/high modulus fibre 3501-5A/HMS,
934/HMS,
F263/T50.
930/GY70,
CE-339/GY70,
First generation epoxy CFRP/ultra high modulus fibre CE339/P75S,
934/P75S,
934/GY70,
V108/GY70.
P1700/C6000,
P1700/C3000,
Polysulphone (PS) thermoplastic CFRP P1700/T300,
P1700/HMF,
P1700/W722.
Polyethersulphone (PES) thermoplastic CFRP PES/T300.
V378A/T300,
Bismaleimide CFRP
F178/T300.
PMR15/C6000,
Polyimide CFRP
LaRC 160/C6000.
CE-339/E-Glass,
Glass fibre epoxy
Epoxy/S-Glass.
SP328/Kevlar,
Aramid fibre epoxy
V108/Kevlar.
Boron fibre epoxy SP290/Boron.
F593/P75,
X904B/T50 (Ferro),
X904B/GY70,
Uncertain X904B/E-Glass,
X-30/GY70,
V108/G837,
E788/T50.
Resin/fibre combinations shown in bold, underlined text were flown by more
than one organisation and at different positions on LDEF.
[See also: Table 20.8.1 for LDEF experiment number, exposure conditions and originating
organisation]
308
ECSSEHB3220Part2A
20March2011
ThegeneralfindingsfromselectedLDEFexperimentswithrespecttoATOXarediscussed:
experiments M0003-9/10.
experiment AO 171.
experiment AO 180.
[Seealso:20.9forthermalcycling;20.11forradiation;20.13forgeneralcommentsonatomicoxygen;
20.16formicrometeoroidanddebrisimpact]
20.8.5 LDEF experiments M0003-9/10
20.8.5.1 General
Theexperimentsincludedover250polymercompositesamples.Theconclusionsfromtherespective
contributorsarepresented,Ref.[206].
20.8.5.2 The Aerospace Corporation

Theconclusionscanbesummarisedas:
ATOXerosiondepthsvariedbetween38mand89m;baseduponmasslossmeasurements,as
showninFigure20.08.2,Ref.[20-6].
ATOXerosiondepthisaninversefunctionofthefibrecontentoftheCFRP.
TwotypesofATOXerosionmorphologywereobserved,probablyrelatedtofibremodulusor
microstructure.
Figure 20.8-2 - LDEF: Estimated AO erosion depth versus fibre content for various
epoxy CFRP materials
309
ECSSEHB3220Part2A
20March2011
20.8.5.3 General Dynamics Space Systems Division
ATOX erosion on the leading edge of LDEF induced a 20% to 30% reduction in the flexural
strengthandmodulusofuncoated(unprotected)epoxyCFRPcomposites.
AnuncoatedT300/V378Abismaleimidesuffereda60%reductioninflexuralstrength.
In general, uncoated CFRP composites suffereda 10% reduction in short beam shear strength
(SBSS).
Uncoated CFRP composites on the trailing edge of LDEF experienced neither significant
changesinflexuralstrengthormodulusnorinshortbeamshearstrength.
TinIndium(SnIn)coatingsappliedtosomecompositesprovidedprotectionagainsttheeffects
ofATOX;withnodiscerniblelossinmechanicalproperties.
The lap shear strength of spotwelded W722/P1700 polysulphone thermoplastic composites
havingZnOorTiO2coatingswereunaffectedbyexposureontheleadingandtrailingedgesof
LDEF.
20.8.5.4 Lockheed Missiles and Space Company

Nolossofmechanicalpropertieswasreportedforepoxymatrixcompositesfromtrailingedge
locations.
The lap shear strength for an HMF330/934 composite bonded to 2024 aluminium with Hysol
9628epoxyfilmadhesivewasreducedby~25%onthetrailingedge.
Nolossofshearstrengthwasnotedforsamplesexposedontheleadingedgeorflightcontrol
samples.
No significant effect on CTE was seen for UD GY70/CE339, T50/F263, T50/934, or T50/X904B
epoxyCFRP.
Epoxy CFRP samples on leading edge locations eroded by 100m to 150m, compared with
13mto30mforepoxyglass,Ref.[20-7].
20.8.5.5 Boeing Defence and Space Group

TheBoeingportionofM000310exposedmaterialsfrombothleadingandtrailingedgelocations,[See:
Figure20.8.1].
Thematerialandlaminateconstructionsevaluatedwere,Ref.[204]:
934/T300(0)16ply,
35106/AS(0)16ply,
PMR15/C6000(0,45,0,45)s,
LaRC160/Graphite(0),
P1700/T300(0,90)fabric8ply.
Theobservationsmadeonmaterialperformanceincluded:
310
ECSSEHB3220Part2A
20March2011
934/T300,35016/ASandLaRC160/Grshowednosignificantlossinflexuralpropertiesperunit
crosssectional area between the different positions on LDEF and the ground control. Any
mechanical performance reduction was directly attributable to the amount of eroded material
allofwhichwasof0orientation.
Ply orientation played a significant role in flexural property behaviour for CFRP with ATOX
exposure.
PMR15/C6000,withangleplystacking,showedasignificantreductioninflexuralstrengthand
modulusduetothenearcompletelossoftheexposedouter0ply.Thisouterplyrepresented
around25%ofprimaryloadbearingcapacity;the45orientationscontributingverylittle.
NomodificationstoCTEbehaviourweredetectedafterLDEFexposure;thetestwasaccurateto
approximately0.001%ofrelativeexpansion.
ExaminationformicrocrackingshowedthatPMR15/C6000andP1700/T300exhibitedsignificant
levels ofsurfaceintraply cracking(up to 14 cracks/cm). No cracking wasseen on any ground
controlspecimensorinanyoftheothercomposites.
CalculationsshowedthattheATOXreactivitiesforthematerials,basedonmassloss,differed
byuptoafactorof4.Thisdidnotcorrespondwellwithmateriallossobservedbymicroscopy.
Itsuggestedthatotherfactors,e.g.presenceofcontaminants,wereinfluencingerosionrates.
Eroded CFRP surfaces showed a jagged peak like structure with stringy, ashlike material
concentrated in clumps around peaks. The level and texture of these features varied between
compositetypes.
Sodiumsulphatewasfoundintheashfrom934andP1700composites.Itwassuggestedthat
thesodiumwasaresiduefromthefibresandsulphurfromthematrix.
Silicatecontaminationwasfoundonallerodedsurfaces.
20.8.6 LDEF experiment AO 171
20.8.6.1 NASA Marshall Space Flight Center

TheAO171exposureconditionsaresummarisedas,Ref.[208]:
Highvacuum:106to107torr,
UVradiation:10471ESH(equivalentsunhours),
Protonflux:109p+/cm(0.5MeVto200MeV),
Electronflux:1012to1018e/cm(0.05MeVto3.0MeV),
Atomicoxygen:6.93x1021atoms/cm,
Micrometeoroidsorspacedebris:2to5impactsoflessthan1mmdiameterper25cm.
Theconclusionsmadewere:
MasslossonCFRPspecimenswasconsistentwithlosingapproximately1plyofthelaminate.
TheP1700polysulphonematrixshowedgreatererosionthanepoxies.
Organic materials, such as polyimides, silicones and polyurethanes, were found to be
luminescentuponexposuretofarUVirradiationafterflight.Luminescenceofcompositeswas
notseen.
311
ECSSEHB3220Part2A
20March2011
FibrousashmaterialwasobservedonerodedCFRPmaterials.
Composite surfaces become more diffuse and darker, accompanied by small increases in
emissivityandabsorption.
A high density of small (<1 mm) micrometeoroid and space debris impacts was seen on all
materials.
Compositetensilestrengthdecreasedapproximatelyinlinewiththicknessloss;assummarised
inTable 20.8.4,Ref.[20-8].
Epoxymatricesbecameslightlyembrittled;probablyaresultofcontinuedcuringunderUVor
electronbombardment.
ChangesintheultimateyieldstressofepoxyCFRPdidnotcorrelatewitheitherweavedirection
orfibretype.
Synergisticeffectswerenoted,whereATOXandcopiousamountofcontaminationinteracted.
S-glassreinforcedepoxycompositesbecameselfprotectedwhentheimmediatesurfacematrix
wasremoved.
Surfacecontaminationwaslocationspecific.Thiswasattributedtodischargedspacecraftfluids
andinteriorunbakedpolyurethanepaintsandprimersystems.
Table20.84LDEFexperimentAO171:Tensilepropertiesofcomposites
UTS E Poissons
Sample
[MPa] [%] [GPa] Ratio
P1700/HMF 322 (45)
Control 1609.7% - 13.65.3% -
Flight 1393.6% - 11.80.8% -
Change -13.0% - -13.2% -
934/HMS (0)
Control 11117.2% 0.528.5% 1964.1% 0.244.9%
Flight 88317.2% 0.3322.3% 1932.3% 0.265.9%
Change -20.8% -35.4% -1.75% +8.1%
934/HMS (90)
Control 22.79.7% 0.298.5% 7.84.7% -
Flight 22.817.3% 0.298.5% 8.03.7% -
Change +1.3% -1.7% +2.7% -
934/P75S (90)
Control 18.67.8% 0.288.9% 6.52.8% -
Flight 12.414.8% 0.1814.1% 6.80.9% -
Change -34.4% -36.2% +3.3% -
934/P75S (0)
Control 65613.6% 0.31NA% 2255.6% 0.30NA
Flight 66930.6% NA 2478.1% 0.336.9%
Change +1.9% NA +9.8% +11.1%
312
ECSSEHB3220Part2A
20March2011
20.8.7 LDEF experiment AO 180
20.8.7.1 University of Toronto Institute for Aerospace Studies

The experiment was located at 82 relative to ram. The total ATOX flux was estimated at ~1.2x1021
atoms/cmandthetotalequivalentsunhoursofVUVradiationwas~6900ESH,Ref.[209].
Theconclusionswere:
MaterialerosionwasattributedtocombinedATOXandVUVexposure.
Boron fibre/epoxy composites show far less erosion than carbon or aramid fibre composites.
Whilst surface epoxy was removed, the boron was effective in inhibiting material loss by
oxidation
Kevlar/epoxy composites were significantly eroded, with evidence that material adjacent to
aluminiumfixturesexperiencedhigherlevelsoferosionbecauseofreflectedATOX.
Ona4ply43934/T300CFRPtube,theexternallayerofepoxywasremovedanderosionof
the carbon fibres in the outer layer clearly evident. Surface measurement of the eroded CFRP
revealed a reduction in CO content and a large increase in both O and Si content, the latter
probablyduetocontamination.
Impactdamagelevelsfrommicrometeoroidsordebriswerenotedonthepolymercomposites.
Somefullpenetrationof4plylaminateswasseen,withsurfacedamagedareasupto1.45mm.
20.8.8 Surface characterisation of eroded composites

Thepresenceofchemicalspeciesonerodedcompositesurfaceshasanimportantroleindefiningthe
rateoferosionunderATOXflux,Ref.[2010].
Ascarbonandhydrogenarestrippedaway,thesurfaceconcentrationofnonvolatilechemicalspecies
increases,oftenintheformofstableoxides.Thesespeciesaccumulatefrom:
residualchemicalsfromerodedmatrixandfibres,
vapourphase deposition of contaminants from other sources (excluding physical debris and
micrometeoroids).
CharacterisingthesurfaceofLDEFmaterialsshowedthat:
Acommontrendforpolymersystems,excludingFEP,wasanincreaseintheoxygenandsilicon
contentswithasubsequentdecreaseinthecarboncontent.
Aluminium specimen tray clamps (6061T6) with thermal control paints, Chemglaze A276
(white)andZ306(black),wereasourceofsiliconcontamination.
FEPwasidentifiedasasourceoffluorinecontamination.Despitethis,thefluorinecontaining
polymershowedsignificanceresistancetoATOXerosion.
Levelsofcontaminationwerelocationdependent.
Higher levels of contamination were found on nearleading or leading edges than on near
trailingortrailingedgesofLDEF.
Thepassivatingeffectofsurfacesiliconoxidesexplainsthereasoningforsiloxanebasedcoatingsand
matricesasmeansofcontrollingATOXerosioneffects,[Seealso:20.14].
313
ECSSEHB3220Part2A
20March2011
20.8.9 Overall conclusions on LDEF

AsLDEFnearedreentryinthelaterstagesofthemission,ATOXerosionofpolymermaterialswas
verysevere,Ref.[2011].Inextremecases,awholeplywaslostfromthinCFRPlaminates.
LDEFwasinorbitforapproaching6yearscomparedwiththe25to30yeardesignlivesofsomeLEO
projects, such as space stations. An important conclusion is that polymers and polymer composites
need surface protection to slow or eliminate surface erosion; particularly in leadingedge ram
locations.Also,theeffectofaccumulateddebrisdamageshouldbeincludedinthedesign.
The combined effect of all of the degradation mechanisms depends on the location on the orbiting
structure.Tobeefficient,solutionstocombatdegradationneedtoberelatedtothelocationandnot
globaltotheentirestructuraldesign.
20.8.10 Non-polymeric composites on LDEF

Therangeofmaterialsflownincluded:
MMCmetalmatrixcomposites,i.e.Al/C,Mg/C,Al/SiC.
GMCglassmatrixcomposites.
ThesewereunaffectedbyATOX,butweresusceptibletoimpactdamagefromdebris.
20.8.11 Polymer films on LDEF
20.8.11.1 Silvered Teflon (Ag/FEP)

Ag/FEPblanketsremainedfunctionalasathermalcontrolsystemoverthelifetimeoftheLDEF,Ref.
[2011].However,thedegradationmechanismsobservedwere:
ATOXattackreducedthethicknessofthefilmsuchthatitsemissivitydecreased.
SolarUVandVUVdegradedthemechanicalproperties.
Delaminationzoneswereobservedarounddebrisimpactsites.
The silver layer in the adhesive backed Ag/FEP was cracked during application onto the
aluminiumsubstrate,causingbleedthroughandsubsequentdarkeningoftheadhesiveunder
solarexposure.
Roughening of the surface texture of the FEP layer dramatically increased the diffuse
componentofreflectance.
20.8.11.2 Kapton
Upto250mofthethicknessofpolyimideKaptonwasremovedfromnearleadingedgelocations.
20.8.12 Lubricants, adhesives and seals on LDEF
20.8.12.1 General
ThesewereusedwidelythroughoutLDEF,butnotonexternalsurfaces.Somegeneralcommentsare
providedontheirdurability,Ref.[2011],[2012].
314
ECSSEHB3220Part2A
20March2011
20.8.12.2 Lubricants
Lubricantsunderwentviscositychangesasorganicbinderswereremovedundervacuum.Otherthan
one occasion, where loss of cetyl alcohol lubricant caused severe galling (seizure) of fasteners, the
overallbehaviourwassatisfactory.Otherdryfilmlubricants,suchasmolybdenumsulphide,behaved
well,asdidheatsinkgrease(ApiezonH)andgraphitefilledpolyimide.
The volatiles and contaminants released by lubricants need closer examination as they can affect
adjacentmaterials.
20.8.12.3 Adhesives
Aralditeadhesiveswereextensivelyused,including:
AV100/HV100.
AV138/HV998.
AV138/HW2951.
AW136/HY994.
AW2101/HW2951.
MY750/HY956.
DiscolourationfromUVexposurewasnoted.
ExperimentM0003testedtwo3Madhesives;AF143filmadhesiveandEC2216RTepoxy,inlapshear
specimens with epoxy CFRPtoCFRP or TitoCFRP adherends. In all cases the LDEF trailing edge
exposedjointswerestrongerthanpreflightcontrolsby10%to15%,typically.
Some adhesive bond failures were noted where four solar cells detached. These were bonded to
aluminiumusinganepoxy.Failureoccurredatthesolarcellinterface;sucheventswererandomand
unexplained.Thedarkeningofadhesiveswasgreatestwithsolarcells.
Silicone adhesives were successfully used for bonding FEP and Kapton films. Tapes, conformal
coatingsandpottingcompoundswerealsosuccessfullydeployed.
20.8.12.4 Seals
The materials included butyl, nitrile, ethylenepropylene and Viton for Orings, with EPDM, NBR,
neopreneandsiliconesforgaskets.
All seals, both Oring and compressed sheet, were protected from direct exposure to the space
environment.Nosignificantfailureswereobserved.
20.9 Thermal cycling
20.9.1 Conditions
Thermal cycling of a structure occurs as the spacecraft moves in and out of the Earths shadow.
Typicalthermalcyclingrangesforthecommonorbitsare:
LEO(spacestation):within90Cto+90C,e.g.between57Cand+38CformetalcoatedCFRP
trusstubes,Ref.[20-13].
315
ECSSEHB3220Part2A
20March2011
GEO:
communicationssatellitebetween150Cand+120C.
exceptionalconditionsfrom180Cto+130C,e.g.PolarPlatform.
Thenumberofcyclesexperiencedineachoftheseorbitsistypically:
LEO:15.4cycles/dayfor30years=169000cycles,
GEO(longduration):canexperience10950cycles.
The exact conditions experienced and the number of thermal cycles depends on the details of the
mission.
20.9.2 Damage mechanisms
20.9.2.1 General
The relationships identified between thermal cycling, damage mechanisms and thermomechanical
propertiesare:
Microcrackingoccursmainlyduringthefirst103thermalcycles.Ifcyclingcontinuesbeyondthis,
thegrowthrateisreduced.
Strengthdiminishesasafunctionofthenumberofcracks.Areductionofbetween3%and9%
wasseeninthecaseofepoxybasedCFRPcomposites;asshowninFigure 20.9.1,Ref.[20-14].
Thereductioninstiffnessappearstobesimilartothatforstrength,althoughthetendencyfor
changestodiminishbeyondacertainnumberofcyclescannotgenerallybeexpected;asshown
inFigure 20.9.2,Ref.[20-14].
Microcracking causes a reduction in the matrixdominated properties of composites. These
combinetocausechangesinthecoefficientofthermalexpansion.TheextentofchangeofCTE
depends strongly on laminate layup and also on the particular fibre/resin combination, as
shown in Table 20.9.1, Ref. [20-14]. CTE can decrease and show less scatter as the number of
cyclesincreases.
Table20.91Coefficientofthermalexpansionfor45
Laminate Type 45 Cross Ply
Material Carbon/Epoxy Carbon/Polyimide
914C-TS-5 HY-E 1548A/B LY556/T300 T3TF178
Number of -6 -1
thermal cycles CTE x (10 K )
0 1.9 0.6 1.5 -0.9
1170 - - 1.7 -
2295 1.9 0.6 2.0 -0.9
3480 2.0 0.6 2.4 -0.8
316
ECSSEHB3220Part2A
20March2011
Figure20.91Ratioofresidualtensilestrengthtotheinitialstrengthversus
numberofthermalcycles(160to+95C):[45]2SspecimentensiletestedatRTand
100C
317
ECSSEHB3220Part2A
20March2011
Referenceload=Loadresultinginelongationof2mm/minavirginspecimenatRoomTemperature.
Figure20.92Degradationofstiffnessofvariousmaterialsduetothermalcycling
(160Cto+95C):[45]2SspecimentensiletestedatRTand100C
318
ECSSEHB3220Part2A
20March2011
20.9.2.2 Microcracking
An ESTECfunded study in 1994 showed that microcracking can occur at different levels in
composites with nominally the same prepregs, Ref. [2015]. This was attributed to space contractors
havingdifferentpreferencesastohowtheprepregswereconsolidated.
Theparameterswhichvariedincluded:
Useornotofpeel plies,e.g.bleedornobleed.
Dwellperiods.
Mouldtoolmaterials.
Consolidationpressures.
It cannot therefore be automatically assumed that microcracking behaviour is dependent only on

materials selection. Laminate layup and thickness are additional parameters which affect
microcrackingbehaviour.
20.9.2.3 ECSS requirements

ECSSQST7004detailsathermalcyclingtestforthescreeningofspacematerialsandprocesses.
20.10 Vacuum
20.10.1 Effects of vacuum
20.10.1.1 General
The vacuumof space varies fromabout 104 torr (at100km) to less than1010 torr (in excess of1000
km).
20.10.1.2 Outgassing
The combination of temperature and vacuum encountered by a space structure causes matrix
materials to outgas. This, apart from reducing the mechanical properties of the composite, liberates
materialthatcancondenseonopticalequipmentandaffecttheperformance.
20.10.1.3 Resin matrix characteristics

For dimensionally stable, and optical structures in particular, there is a progressive move towards
matriceswithlowoutgassingcharacteristicsandlowmoistureabsorption.Theseinclude:
thermosets,[Seealso:2.4;6.35]:
cyanateesters,
newlowmoistureepoxies.
thermoplastics,[Seealso:6.17].
metalsandceramics,[Seealso:Chapter42].
319
ECSSEHB3220Part2A
20March2011
The close proximity of materials to optical devices can be sufficient to dictate the use of very low
outgassingmatrices.
[Seealso:ESMATwebsiteforoutgassingdatafromtheESTECMaterialsandProcessesdatabase]
20.10.1.4 Offgassing
Formannedvehicles,thetoxicityofoffgassedproductsneedscarefullyconsideration.

In order to minimise the effects of outgassing, the normal ECSS requirements (see ECSS for latest
standards),formaterialsat106torrandspacestructuretemperatures,are:
aprovenmaximumrecoveredmassloss(RML)<1%,and
collectedvolatilecondensedmaterials(CVCM)<0.1%.
[Seealso:5.2forcommentsonoutgassingcharacteristics]
ECSSQST7002detailsathermalvacuumtestforthescreeningofspacematerials;previouslyESA
PSS01702.
TheoutgassingvaluesstatedinECSSQST7002representthemaximumallowableforspaceuse.
Someapplicationsstipulatematerialswithverylowoutgassinglevels,e.g.foruseincloseproximity
toopticaldevices.
ECSSQST7029 describes the determination of offgassing products from materials and assembled
articlestobeusedinamannedspacevehiclecrewcompartment;previouslyESAPSS01729.
20.11 Radiation
20.11.1 Radiation spectra
20.11.1.1 General
Inthespaceenvironment,polymericmatrixmaterialscanbesubjecttoawiderangeofradiation.The
effects depend on the molecular structure of the material and the amount of radiation energy
absorbed. Initially the production of free radicals and ions within the matrix material gives rise to
crosslinking,chaincrackingorchainpolymerisation.Thisisaccompaniedbyaslightincreaseinthe
glass transition temperature, Tg. With a very heavy radiation dose, breaking of molecular chains is
widespread and the degradation can become acute. This is accompanied by a decrease in Tg. The
typesofradiationthatcanbeencounteredinclude,Ref.[2016],[2044]:
electrons,
protons,
ultraviolet,
solarflares,
cosmicrays.
320
ECSSEHB3220Part2A
20March2011
TheoveralllevelsofradiationarehigherinGEOcomparedwithLEOandspectradiffer.30yearsin
GEOisequivalentto~100MGray(10,000Mrad).
Carbonandglassfibresarenotaffectedbyradiationintermsofchangestomechanicalpropertiesor
degradation.
Aramid fibres degrade, notably by UV, unless shielded by the matrix. Very high doses degrade the
polymericaramidfibreswithinacompositeatthesametimeasdegradingthematrix.
20.11.1.2 UV radiation
UVradiationdoesnotappeartohavemucheffectonepoxybasedCFRPcomposites.
Figure20.11.1showsthatthestiffnessdoesnotchangesignificantly,becauseUVradiationaffectsonly
the surface of the structure, Ref. [2018]. Likewise, CFRP shows no significant change in CTE when
exposedtosunandvacuum.
Figure20.111UVradiation:Effectsontensilemoduliofcarbon/epoxy
composites
20.11.1.3 Particle radiation

ParticleradiationpenetratesCFRPmaterialandcauseschangesinmechanicalproperties.Depending
on the fibre orientation and dosage (up to 10000 Mrad), a reduction in tensile strength of ~ 25% is
notedinsomepublications;somealsoreportincreasesinmodulusupto13%,Ref.[2017].
300Mradisequivalentto1yearinorbit;10,000Mrad=30years.
321
ECSSEHB3220Part2A
20March2011
The changes in mechanical properties are due to molecular crosslinking and chain scission
mechanisms. These embrittle the matrix and lead to microcracking during thermal cycling. As
radiation doses increase, chain scission becomes the dominant effect. The coefficient of thermal
expansion,CTE,alsochanges.Theeffectsarematerialdependent,but,asshowninFigure20.11.2,are
lesspronouncedathighdosages;10000Mrad,Ref.[2019].
Figure20.112Effectsoftotalradiationdoseonmicrocrackformationin
compositespecimenssubjectedto100thermalcycles(150Cto+80C)
Figure20.11.3showssmallchangesinCTEoccuratlowdosesandanapparentthresholdof108rad,
whereadecreaseinCTEoccurs,Ref.[2019].
322
ECSSEHB3220Part2A
20March2011
Figure20.113Effectsoftotalradiationdoseonthecoefficientofthermal
expansionofatoughenedcompositesystem(150Cto73C)
20.11.2 T300/934 in GEO
20.11.2.1 General
NASAcarriedoutastudyonthewidelyusedFiberite934combinedwithacommonfirstgeneration,
highstrengthfibreT300,Ref.[2020].
Theprincipaltestparameterswere:
BulkresinspecimensandUDcompositesorientedat0,10,45and90.
Creepbehaviourat20%,40%,60%and80%ofultimatetensilestrength,UTS.
10000Mradexposureto1MeVelectronsat35C.
MechanicaltestsatRT(23C)and121C.
ChangesinTg.
The aim of the work was to investigate the creep behaviour of composites exposed to very high
radiationdoses,wheremolecularchainmodificationsareknowntooccur.
Theconclusionsmadewere:
Changesincreepbehaviourwereonlyobservedinoffaxisspecimens,i.e.non0.
Radiationmodificationsignificantlyincreasedcreepat121CbutnotatRT.
The matrix Tg was lowered from 240C to 180C by the total radiation dosage, as a result of
molecularchainbreaking.Thisenablesincreasedchainmobility,assistingcreepdeformation.
323
ECSSEHB3220Part2A
20March2011
Prolonged exposure at 121C for ~120 hours promoted a partial recovery in the Tg to 210C
owingtorecombinationoffreeradicals,increasingthelevelofcrosslinking.
Greater residual strain was noted in 10 and 45 irradiated specimens at 121C than the
equivalent90specimens.
20.11.2.2 ST- requirements

ECSSQST7006 describes the particle and UV ultraviolet radiation testing of space materials;
previouslyESAPSS01706.
20.12 Damage by combined environmental factors
20.12.1 General
FororbitingstructuresinGEO,thetypicaldesignlifeisbetween5yearsand7years.Studiesonthe
effects of thermal cycling and radiation on composites over these periods indicate small changes in
mechanical and physical properties. These changes were considered to be largely insignificant for
suchshortlifeapplications.
Withthecurrentgoalsofextendingmissionsto30yearsinGEOandinLEO,materialdegradationis
beingreevaluated,particularlythepossibilityofdamagefromcombinedspaceenvironmentfactors.
Eachorbithasdifferentprofilesofdegradationmechanismseverities.Whichneedtobeconsideredin
studiesofprolongedspaceexposureonmaterials.
Forcomposites,thedominantmechanismscanbesummarisedas:
GEO:
highnumberofseverethermalcycles,
highradiationdosages.
LEO:
highatomicoxygen(ATOX)fluxes,
highnumberofimpactsbydebrisandmicrometeoroids.
However, the lesser effects cannot be ignored. Structures in GEO also experience outgassing and
debris impacts. Likewise, in LEO, a high number of thermal cycles occur (but with a less severe
thermalrangethanGEO),andalsoVUVexposure.
A number of test programmes have been carried out to consider the response of CFRP materials to
GEOandLEOtypeenvironments.Theresultsofsomeofthesestudiesaresummarised.
20.12.2 P75/930 in GEO

NASAundertookanevaluationofacarbon/epoxyP75/930compositesystemtostudytherelationship
between radiation modification and thermal cycling, and their effect on microcracking density, Ref.
[2021].
Modificationstothecureschedulewereproposedasameansofreducingcrackdensities.
324
ECSSEHB3220Part2A
20March2011
Fiberite930epoxyiscuredbetween135Cand176C.
ReferencematerialsforcomparisonwereT300/934andP75/934(177Ccure).
Thermalcyclingbetween+65Cand157Cforupto500cycles.
Specimenconfigurations:
(10)4forshearproperties,
(90)4and(0,902,0)formicrocrackingdensity.
10,000Mradof1MeVelectronsinvacuum.
Theresultsandobservationsthatweremadeincluded:
Figure 20.12.1showsTgvariationwithcuretemperatureandradiationdose,Ref.[20-21].
For 135C cure specimens, radiation exposure initially increased the Tg from 110C to 124C,
followedbyareductionto93Cafter10,000Mrad.
Initial radiation exposureincreases cross-link density, before chain scission becomes significant
atathresholdof100Mradto1000Mrad.
The shear strength of 10 specimens was not significantly dependent on cure temperature; as
illustratedinFigure 20.12.2,Ref.[20-21].
Shearstrengthwasdramaticallyloweredbyradiationdosage,i.e.80%ofunexposedafter100
Mrad,butonly10%after10,000Mrad(90%loss).
AsmadeP75/930compositesdidnothavemicrocracks:
After 50 cycles, all nonirradiated materials, irrespective of cure temperature, had
significantmicrocrackdensitiesof8to12cracks/cm.
After500cycles,thecrackdensityincreasedtobetween16and18cracks/cm.
Forirradiatedsamplesthemicrocrackdensitiespeakedat20to22cracks/cmafter10,000Mrad
and500cycles.
325
ECSSEHB3220Part2A
20March2011
Figure 20.12-1 - Effect of cure temperature and electron radiation exposure on the Tg of
P75/930 in GEO
326
ECSSEHB3220Part2A
20March2011
Figure20.122Effectsofcuretemperatureandelectronradiationexposureonthe
shearstrengthofP75/930inGEO
The results from the NASA work when compared with those of previous findings for P75/934
showed:
BothFiberite930and934undergochainscissionatveryhighradiationdosages.930wasmore
susceptiblethan934tochangesinTg.
Nonirradiatedmaterials:
P75/930startedwithnomicrocracks,butattained18cracks/cmafter500cycles.
P75/934startedat5cracks/cmasmadeandstabilisedat11cracks/cmafter10cycles.
Fiberite 930, cured at 135C, offered no advantages over 934 (the higher temperature curing
alternative),whenusedwithP75Sfibre.
327
ECSSEHB3220Part2A
20March2011
P75 is a first generation pitch carbon fibre of USA origin. Many organisations consider it to be
inadequate for composites experiencing severe thermal cycling because of its poor composite
transverseproperties,i.e.lowfibre/matrixinterfacestrength;hencelow90plystrength.
20.12.3 UHM CFRP in GEO and LEO

CERTONERA(France)examinedthebehaviourof4epoxybasedCFRPmaterials,Ref.[2022].
NOTE GY70 fibre and Amoco ERL1939 resin are no longer commercially
available. An HS carbon fibre was used rather than an UHM.
CFRPmaterialsevaluated:
GY70/M18,
TenaxHTA/M18,
GY70/Fiberite954,and
P75/ERL1939.
Crossply[02,902]slaminates.
Inconjunctionwiththermalcycles(between100Cand+100C),astepsimulationprovided,as
showninFigure 20.12.3,Ref.[20-22],either:
asymmetricirradiation,or
uniformirradiation.
Radiationdosagescorrespondedto21yearsinGEO.
Asymmetric irradiation reproduces the GEO environment where absorbed energy at 100m
subsurfaceis40timesgreaterthanat1000m.
Electronirradiationwithsuccessiveelectronenergies(150keV,250keVand500keV)wasused
togiveadosegradientthroughthethickness.
NOTE Applied to M18 and ERL1939 laminates.
Uniformradiationwasachievedwithmoreenergetic2MeVelectrons.
NOTE Applied to GY70/M18 laminates only.
328
ECSSEHB3220Part2A
20March2011
Figure20.123Stepsimulationof7yearsinGEO
Propertiesstudiedwere:
Flexuralmodulus,
GlasstransitiontemperatureTg,
Microcrackdensities,and
ResponsetoatomicoxygenATOXerosion.
Theobservationsmadewere:
329
ECSSEHB3220Part2A
20March2011
Flexuralmodulus:Thecombinedeffectsofradiationandthermalcyclingreducedtheflexural
modulus.Themaximumchangewasa5%reductionforasymmetricallyirradiatedGY70/M18.
Tg:Aftermaximumexposureandthermalcycling,allmaterialsexperiencedaTgreductionof
~10C,e.g.M18reducedfrom200Cto190C.
Microcracking:
GY70/M18:Anaverageof9cracks/cmwasnotedintheasmadestate.Thisincreasedto
13cracks/cmafterirradiationandthermalcycling.Irradiationalonedidnotincreasethe
crackdensity;asshowninTable 20.12.1,Ref.[20-22].
GY70/954:7.5cracks/cmafter63thermalcycles.Theeffectofradiationwasnotstudied.
ATOXerosion:
The erosion rate of the composites differed under ATOX erosion up to 1.2 x 1021
atoms/cm;asshowninFigure 20.12.4,Ref.[20-22].
Therateoferosionisprimarilyafunctionoftheresin,butthefibretypewasalsoseento
haveaneffectwhenthedifferentUHMfibreswerecompared.
NOTE The maximum dose was approximately 15% of that for ram facing
samples on LDEF, [See: 20.8].
For the UHM composites, the ATOX erosion accounted for a 4% to 12% reduction in
dynamicmodulusinbending.
Table20.121Microcrackdensitiesin[02/902]SCFRPlaminates
A) Effect of thermal cycling on microcracking (Cracks/cm)
Number of Thermal Cycles
Composite [-100C to +100C]
0 21 42 63
M18/GY70 9.7 11.9 12.3 13.8
M18/HTA <1 <1 <1 <1
954/GY70 1 1.5 2.3 7.5
ERL1939/P75 <1 <1 <1 <1
B) Effect of simulated GEO environment on microcracking (Cracks/cm)

Equivalent Years in Space
Composite
0 7 14 21
M18/GY70 6.8 10.3 9.1 13.3
M18/GY70 8.2 8.0 12.2 13.1
HTA/M18 <1 <1 <1 <1
ERL1939/P75 <1 2.8 7.7 7.5
Key: : Uniform irradiation : Asymmetrical radiation
330
ECSSEHB3220Part2A
20March2011
Figure20.124EffectofATOXerosionondifferentCFRPcomposites
20.12.4 PEI CFRP in LEO and GEO

NASA investigated the use of thermoplastic polyetherimide, PEI, as a matrix for composite space
structures. The effect of high dosage radiation and thermal cycling was studied in conjunction with
ATOXerosion,Ref.[2023].
Theprincipalparametersofthestudywere:
Ultem1000PEI/CelaneseC6000composite.
Nonsizedand2%epoxysizedlaminates.
10,000Mradby1MeVelectrons.
Thermalcyclingfrom101Cto+66C,upto1000cycles.
ATOXexposureequivalentto26daysinLEO,e.g.spacestationlevels.
Flexuralproperties,interlaminarshearstrengthILSSandTgweremeasured.
Themainconclusionsofthestudywere:
Thepresenceofanepoxysizehadadetrimentaleffectonmechanicalproperties;asshownin
Table 20.12.2,Ref.[20-23].
TheTgofPEIincreasedby15Coverthetotal10,000Mrad,butthischangedidnotstartuntil
after7500Mrad.ThechangeinTgwaspresumablyduetocross-linkingeffects.
The relative changes in flexural strength, flexural modulus and flexural straintofailure are
showninFigure 20.12.5fordifferenttestconditions,Ref.[20-23].
The most dramatic changes were seen for transverse properties and not those in the
longitudinaldirection.
331
ECSSEHB3220Part2A
20March2011
There was no evidence that thermal cycling alone caused any significant degradation in the
composite.Indeedtherewasevidencethatsomepropertiesimproved.
Thecombinedeffectsofthermalcyclingandirradiationproducedlittlechangeinthematerial.
ATOX erosion produced the anticipated loss of material. The effect was more noticeable on
transverseratherthanlongitudinalcomposites.
ImideandphenylgroupsshowedgreaterresistancetoATOXthanethersandalkanefunctional
groups,orhighlybranchedgroupssuchasisopolypropylene.
Table20.122GEOandLEOeffects:MechanicalpropertiesofC6000/PEI
composites
Non-Sized Fibres Sized Fibres
T L T L
Flexural strength (MPa) 11115 2420110 14514 2010170
Flexural modulus (GPa) 8.240.42 1297 7.180.60 10111
Strain to failure (%) 1.4 2.0 1.8 2.0
ILSS (MPa) - 795 - 714
Key: T: Transverse L: Longitudinal
332
ECSSEHB3220Part2A
20March2011
Figure20.125Irradiationandthermalcycling:Changesinmechanicalproperties
forC6000/PEI
20.13 Atomic oxygen
20.13.1 Effects of atomic oxygen
20.13.1.1 Typical exposure levels

SpacecraftinLEOareexposedtoafluxofearthambientatmosphericspecies,themostdominantand
aggressiveofwhichisatomicoxygen,ATOX.
Althoughtheaveragedensityofatomsisonly5x108atoms/cm3,thevelocityofspacecraftinorbit(7.6
km/s)leadstoatotalfluxof1015atoms/cm2/swithacollisionenergyofabout5eV.
Theleveloffluxexperiencedvarieswiththe:
Altitudeofthespacecraftorbit,
Attitudemaximumforram,minimumforwake,
333
ECSSEHB3220Part2A
20March2011
Solaractivity,and
Durationofexposure.
20.13.1.2 Material sensitivity

Thereactivityofcompositematerialsisdeterminedbythecombinationofthereactivitiesofthematrix
andthereinforcement.
Inorganicreinforcingfibres,suchascarbon,areinitiallyshieldedbythematrix.Eventuallythematrix
isremoved,exposingthefibres.Longtermexposurecanleadtodegradationofbothfibreandmatrix
system.
The reaction efficiency is the degradation per solar cycle, for a given flux. There are differences in
ATOX reaction efficiencies between hydrocarbonbased matrices depending on the reactive groups
and radicals present. Reactivity can be reduced, by orders of magnitude, if silicon or fluorine is
presentinthemolecularstructure.
Table20.13.1showsthereactionefficienciesforsomeselectedmaterials,Ref.[2024].Thedegradation
persolarcycleforthedifferentmaterialscanbepredicted,providedthatthefluxandsolarcycleare
known.
Table20.131Reactionefficienciesforsomeselectedcompositepolymersand
organicfilms:Determinedby~40hoursofLEOexposureinspace
ReactionE ciency
Material
x1024(cm3atom1)
Kapton 3.0
Mylar 3.4
Tedlar 3.2
Polyethylene 3.7
PMMA* 3.1
Polyimide 3.3
Polysulphone 2.4
1034C epoxy 2.1
5028/T300 epoxy 2.6
Teflon, TFE 0.05
Teflon, FEP 0.05
Key: *PMMA:polymethylmethacrylate
:Estimatedaccuracy:30%to50%
20.13.1.3 Damage to composites

Thedamagingeffectsofatomicoxygenoccurbydirectoxidationoftheorganicpolymerchains,e.g.
epoxythermosetresin,resultingintheformationandliberationofvolatileswithacorrespondingloss
inmass.Onthemicroscopiclevel,thesurfacesareroughenedandultimatelysignificantquantitiesof
materialcanbelost.
[See also: 20.8 for ATOX erosion studies conducted as part of the long duration exposure facility
programme,LDEF]
334
ECSSEHB3220Part2A
20March2011
20.13.1.4 Property loss
Theoxidationofbothmatrixandreinforcementproduceschangesinbothmechanicalpropertiesand
CTE.Table20.13.2showstheeffectofshortexposuretime(40hours)onCTE,Ref.[2025].
Significant and continuous changes inCTE can occur when such materials are exposedfor up to 30
years.
Table 20.13-2 - Effects of short-term LEO exposure on coefficient of thermal expansion

for T300/934 carbon/epoxy laminates (RT to +82C)
-6 -1
CTE x 10 C
Sample lay-up
Control Flight
45 + 4.61 + 4.28 to 4.43
0 + 1.76 to 1.84 +0.65 to 0.70
20.13.1.5 Protection against ATOX

ForsatellitesurfacesandcompositesexposedtoveryhighlevelsofATOXflux,protectionschemesare
essential,[See:20.14;20.15].Differentprotectionsystemsarenecessaryforcomposites,polymerfilms,
thermalcontrolcoatingsandsolarcells.

AnaccuratesimulationofLEOfluxisdifficulttoreproduceonEarth,Ref.[2026].Testfacilitiesare
availableatESA/ESTEC,Ref.[2027].
NoECSSstandardforATOXtestingexistscurrently.
20.14 Siloxanes and silicon polymers
20.14.1 Protection methods against ATOX

The presence of siliconat the surface of compositeshas demonstrated its value in formingan oxide
barriertoATOXerosion.Theinclusionofglassfibres(silica)alsohasaprotectiveeffect.However,the
possiblereleaseofglassfibres,throughlossofmatrixbyerosion,hasraisedsomeconcernsaboutthis
routebeinganacceptablewidescalesolution,Ref.[2028].GFRPisnotanefficientstructuralmaterial,
sothecladdingofCFRPasameansofprotectionhasfewattractions.
The use of siliconcontaining polymers does have merits for passivating the surface. Various
techniqueshavebeenproposedforincreasingthesurfacesiliconcontent:
Siliconecoatings(paintorfluids),Ref.[20-29],
Siloxanecoatings,
Siloxanematrixcomposites,
Siloxanemodifiedepoxies,Ref.[20-30],
Siloxanemodifiedcyanateesters,Ref.[20-31],
Anoxicpolymersandcomposites,Ref.[20-32],
335
ECSSEHB3220Part2A
20March2011
Siloxanemolecularcomposites,Ref.[20-32],
Sputteredsilicacoatings,Ref.[20-30].
The questions of particular significance when considering protection against ATOX flux by these
proposedmethodsare:
Is there an acceptable process for applying the surface layer to the intended component or
structure?
Doestheprotectivesurfacecoatingsurvivethermalcyclinganddebrisimpacts?
20.14.2 Protection of polymer films

PossiblesolutionsforKaptonandsilver/FEPfilmsare:
Sputtered30nmSiO2coating.
DowCorningDC976andDC705siliconefluidsapplieddilutedinsolvent.
Coatingsforpolymerfilmsneedtobemuchthinnerthanthoseappliedtostructuralcomposites.
20.14.3 Protection of composites
20.14.3.1 General
There is a range of possible protection methods because there are many ways of grafting silicon
containing chemical groups onto thermosetting or thermoplastic polymers. For example, if a
conventionalthermosettingCFRPcompositeisusedastheprimarystructuralmaterial:
Asiloxanebasedcoatingcanbeappliedtotheexternalsurface.Althougheasilyimplemented,
doubtsremainaboutitsadhesionunderthermalcyclinganddebrisimpact.
Elastomericsiloxanecoatingsareflexibleandhavereasonablyhighsiliconcontents,soactasa
sacrificiallayertoprotectthecompositesubstrate.
The direct incorporation of siloxane groups into the matrix, or even the formation of self
reinforcingmolecularcomposites,isalsobeingstudied.Asiloxanechain(SiOSi)with,for
example, methyl branches, can be incorporated into a crosslinked molecular arrangement
(thermoset) or terminated with a block copolymer (thermoplastic). Examples of the methods
underevaluationare:
siloxanemodifiedcyanateester,and
anoxicsiloxanemolecularcomposites.
20.14.3.2 Siloxane modified cyanate ester resin

FromtheearlySpaceShuttlefights,e.g.STS8,thebenefitsofcrosslinkedpolydimethylsiloxanesand
poly(siloxaneimide)copolymerswereestablishedbecausetheyformedstableoxidelayerstoinhibit
ATOX erosion. By themselves, these polymers are elastomeric in nature and so lack the mechanical
propertiesforstructuraluse.
Thecyanateestershavenotableadvantagesascompositematrices,i.e.lowmoistureabsorptionand
goodtoughness,adhesionandshearstrength,[Seealso:6.35].
336
ECSSEHB3220Part2A
20March2011
A combination of siloxanes and cyanate esters, retaining, to some extent, the advantages in a cross
linkedstructure,wasobtainedbyadding10%to20%byweightofsiloxane,Ref.[2019].Themodified
resinformulationsshowedagreaterresistancetomasslossbyoxygenplasma.Inpracticethisgavea
twofoldreductioninmasslossforaCFRPmaterialpreparedwith10%siloxaneresincomparedwith
equivalentlaminatesusingepoxyFiberite934andcyanateester9541matrices.
Theadditionofthesiloxanehadasofteningeffectonthematrices,loweringtheshearandflexural
moduli. Therefore higher siloxane additions cannot be made without seriously compromising the
matrix moduli,on which the shearand compressive integrity of the composite is based. It therefore
remains to be proven whether this route provides sufficient protection to prevent carbon fibre loss
under prolonged ATOX flux, given its modest surface silicon yield. Composites with silicon
containingmatriceshaveyettobeflowninspace.
20.14.3.3 Anoxic siloxane molecular composites

Thetermanoxicisusedtodescribematerialsresistanttothermooxidativeattackbyatomicoxygen,
Ref.[2032].
Amolecularcompositeisthecombinationoftwomoleculargroupswhereonerigidphaseprovides
areinforcingeffect.Thedevelopmentofthesematerials,withsupportfromESA/ESTEC,includes:
Molecularcompositefilms(10mto30m),and
Molecularcompositefilms+carbonfibres=structuralcomposite(workfrom1995onwards).
Molecular composites of polysiloxaneetherimide/polyamideimide (flexible/rigid) in thin film form

wereexposedinLEOonboardESAsBIOPANmission.
In addition, test exposures to an oxygen plasma flux of 2.8 x 1020 atoms/cm2 demonstrated good
erosion resistance. These films are being investigated as solar cell support film for LEO spacecraft
solararraystoreplaceKapton.
NOTE The usual abbreviation for polysiloxane-etherimide/ polyamide-imide is

PSI/PAI molecular composites.
The composites are formulated by dissolving both Siltem 1500 PSI and Amoco Torlon 4203 PAI
polymers (G E Plastics B.V.), in various proportions in a solvent. These composites are likely to be
particularlyresistanttoATOXbecauseoftheirinherentabilitytocreateaprotectivesurfacesilicafilm.
NOTE PSI/PAI molecular composites are likely to conform to the outgassing

requirements in ECSS-Q-ST-70-02.
337
ECSSEHB3220Part2A
20March2011
20.15 Protective coatings
20.15.1 Surface coatings
20.15.1.1 General
Theapplicationofcoatingstopolymercompositesisnormallyconsideredasalastresortbecauseof
thedifficultiesinobtainingreliableadhesion.
Thecircumstanceswhichdictatetheneedtoapplyprotectiveorsurfacemodifyingcoatingsinclude:
Inhibitmoistureabsorption,e.g.SnInmetallicbarrier,
InhibitATOXerosion,e.g.siliconcontainingpolymercoatings,
Providethermalcontrol,e.g.paints,
Provideareflectivecoating,e.g.sputteredmetal,
Providecontinuityofelectricalconductivity.
ECSSQST70andECSSQ7071provideinformationontheselectionofmaterialsandprocessesfor
spaceapplications.
20.15.1.2 Atomic oxygen protection methods for CFRP

Compositesandotherorganicmaterialsintendedforlongtermdeploymentinspaceneedprotective
coatings.
ThereareseveralviablewaysofprovidingprotectionforCFRPelements.Thesearemainlybasedon
metalsorperfluorinatedpolymers.
Candidateprotectivematerialsare,Ref.[2025]:
appliedmetalfoil,
sputteredmetals,
vapourdepositedmetals,
perfluorinatedfilms,e.g.Teflon,
Teflonbasedcoatings,
siliconeorsiloxanecoatings[See:20.14].
Metals which form stable oxides within space environments can be considered for coatings. The
exceptionswhichhavetobeavoidedare:
Osmium,whichformsavolatileoxide.
Titanium,whichcontinuestooxidisewithacorrespondingincreaseinmassandlossoffunction.
Silver,whichsuffersfromcrackingandspalling.
338
ECSSEHB3220Part2A
20March2011
Any applied coating needs to be free from scratches and other minor surface damage. Any
imperfectioninthecoatingprovidessitesforoxidationtostartandcanresultinlocalisedpenetration
ofthecomposite.
20.16 Debris
20.16.1 Classification of debris
20.16.1.1 General
Impactingparticlescomefromtwosources:
Micrometeoroids,and
Accumulatedmanmadedebristrappedinorbit.
The impact flux is dependent on the orbit height, craft orientation and velocity, coupled with the
impacttrajectories.
Fluxisnotsolelyafunctionofapointinspace,butdependsheavilyontheorientationofindividual
surfacesofthespacecraft.
TheLDEFprogrammeprovidedthemostcomprehensiveinsightintoimpacteventsoverprolonged
exposuresinLEO,[Seealso:20.8].
20.16.1.2 Micrometeoroids
Micrometeoroidscomefromnaturalsources,suchascomets,asteroids,themoonandplanets.Their
characteristicscanbesummarisedas:
VelocityrelativetoEarth,from12to72km/s,typically.
Impactvelocitiesrelativetoorbitingsatellites,from4km/sto79km/s,typically.
Typicalvelocityrange:10km/sto30km/s(average19km/s).
Cometmeteoroidsaremainlyiceparticles,withadensityof0.5g/cm,typically.
Asteroidparticlesaremainlymineralbased,withdensitiesupto8g/cm,typically.
Most micrometeoroids in the nearEarth environment are in the range of 50m to 1 mm in
diameter.
Beta meteoroids (interplanetary dust of less than 1 m diameter) do not cause significant
damage.
20.16.1.3 Orbital debris

Thisconsistsofaverywiderangeofmanmadedebrisresultingfromstandardlaunch,deployment
andotherspacecraftoperations.
Thegeneralcharacteristicsofdebrisare:
Materials,suchas:
solidrocketpropellantexhaustproducts,e.g.10maluminaparticles,typically.
339
ECSSEHB3220Part2A
20March2011
paintflakes,and
structuraldebrisfromenvironmentalexposure.
A wide distribution of particles sizes, with a greater proportion larger than 1 mm than are
presentinthemicrometeoroidflux.
At500kmaltitudeand28.5inclination,e.g.SpaceStation,therelativeimpactvelocityisabout
10.7km/s.
Forhighlyellipticalorbits,relativevelocitiesrangefrom0km/sto19km/s;averagingabout8
km/sto10km/srelativenormalincidencevelocityforthemajorityofimpactsatobliqueangles.
Theincreaseofinorbitmanmadedebrisisestimatedat6%perannum.InLEO,90%ofallimpacts
are caused by manmade debris. Skylab and early Shuttle flights gave the first indications of the
impactsthatcanbeexperiencedinLEO,Ref.[2033],[2042].
20.16.2 Damage to LDEF

LDEFwasalargestructureorbitinginLEOfornearlysixyears,[See:20.8].Itexperiencedhundredsof
thousands, if not millions, of impacts, Ref. [2029]. These included a single impact from a large
particle0.7mmindiameteronthe7mareaexposedintheramdirection.Suchaparticle,travellingat
6km/s,iscapableofpenetratingatypical2.5mmaluminiumalloysatellitewall.
LDEFalsoexperienced~1impact/montheramexposedsurfacescapableofpenetratingatypical1.5
mmthickaluminiumelectronicsboxwall.Whilsttheseimpactscanbeextremelydamagingtointernal
components,suchaselectronics,batteries,motorsandmechanisms,theyareextremelyrare.
Forallimpacts,thedamageareaisfarlargerthanthediameterofthedebrisparticle.Theshapeofthe
damagearea,orcrater,isdependentontheangleofincidencetotheimpactedsurface.Alowangle
obliqueimpactproducesasignificantareaofimmediatesurfacedamagebutlacksthepenetrationofa
highangleimpact.Differentstructuralmaterialshavedifferentresponsestoimpactinghighvelocity
particles. Whilst individual impacts can be of low structural significance, the accumulated effect of
manyimpactscannotbeunderestimated.
20.16.3 Damage to composites

Damagecanbecharacterisedas:
Apointofimpactwithasurroundingspallzone.
The point of entry shows no sign of melting, but material has been lost and fractured fibres
overlapthesurfacehole.
Asubsurfacedelaminationzone,largerthantheentryhole.
Ifthroughpenetrationhasoccurred,theexitholecanbelargerthantheentryhole.
ThelargestimpactonLDEFproduceddamagewhichconsistedof:
Surfaceholeofabout0.5mmdiameter,
Exitholeofabout1mmdiameter,
340
ECSSEHB3220Part2A
20March2011
Spall area of several millimetres in size, with cracks in the composite matrix extending for
severaltensofmillimetres.
20.16.4 Damage to aluminium alloys

Damagecanbecharacterisedas:
Formationofacrater,withmaterialdisplacedfromthepointofimpact.
Someofthedisplacedmetalismeltedbytheimpact.
Thecraterhasarimwhichstandsproudofthesurface.
Someofthemoltenmaterialisdisplacedfromthecraterintheformofdropletswhichsplashon
thesurroundingarea.
Paintatthepointofimpactisspalledfromthesurfacetobecomeacontaminationsourceinthe
environmentaroundthecraft.
Impactcraterswithanangleofincidencenormaltothesurfacearevirtuallyhemisphericalinshape.
Forobliqueangleimpacts,themainelongatedcratercanhaveseparatecraterswithinit,presumably
resultingfromtheimpactingparticlebreakingupintosmallerpieces.
Ingeneral,impactcraterdiametersaverageabout5timesthediameteroftheimpactingparticle.The
totalregionexperiencingsplashedmaterialaverages10timesthediameteroftheimpactor.
ThesinglelargeimpacttoLDEFstruckastructuralaluminiumZframe.Theframewascoveredwith
silveredTeflonadhesivelybondedthermalcontroltape,Ref.[2034].
Theimpactcharacteristicswere:
The Zframefaced theram direction with the particle, in an elliptical orbit, striking the space
facingendofLDEF.
Particlesizewasestimatedat0.7mmto0.9mmdiameter.
AverylargedamageareawasobservedwithdelaminationattheTeflontosilverandTeflonto
aluminiuminterfaces.
Teflonwasmeltedandejectedfromthecrater.
Thecraterdiameterwas5.25mmwithseverebulgingonthebackfaceoftheframe.
Penetrationwaspreventedbythepresenceofabackingblockofaluminiumbehindtheframe.
Ejectedmaterialwassprayedovera200mmwideswathacrossadjacentareas.
20.16.5 Damage to thermal control materials
20.16.5.1 General
Theareaofmaterialaffectedbyanimpactislargerthanthatseenforaluminium,mainlybecauseof
thedelaminationoflayersthatoccurs,Ref.[2034].
341
ECSSEHB3220Part2A
20March2011
20.16.5.2 Thermal blankets
Atypicalmultilayerthermalcontrolblanket(MLI)onLDEFconsistedof:
Outerlayerofwhitethermalcontrolpaint(ChemglazeA276),
Mylarlayer,
22layersofaluminiumfoilseparatedbyDacronmesh,
Kaptontape,
Lexansubstrate.
These types of constructions not only suffered impact damage, but the subsequent access of ATOX
andVUVtosublayerscausedfurtherdegradationandmaterialbreakup.
20.16.5.3 Thermal control painted materials

The thermal control paints suffered ATOX erosion resulting in the removal of the polymer binders,
leavinglooselyattachedpigmentparticles.
Debrisimpactcompoundedthedegradationbycausingaringphenomenon.Here,theimpactcrater,
with its surrounding spall zone, was ringed by a wrinkled and rolled back top layer of paint. The
raisedringswereoutsidethespallzoneandappearedtobedisplacedpaintorpigmentpowder.The
concernisthatapressurewaveeffecttravelsthroughthepaintandinducesdamage.
20.16.6 Significance of impact events

Highvelocitydebriscancausedamagetoavolumeofimpactedmaterialthatfarexceedstheparticle
volumeorimmediatesurfaceimpactcrater.
Thedistributionofimpactsoverthesurfaceofasatellitedependsonmanyparameters,Ref.[2035].
Thepartsofthestructurethatexperiencethehigherimpactfluxdensitiescanneeddirectshielding,
Ref.[2030].
Asanindicationofthelevelofimpactswhichcanbeanticipated,referencecanbemadetoLDEF:
ImmediatesurfacecrateringonLDEFamountedtoatotaldamageareaofabout0.1%.
The extended damage areas due to the combined surface effects of cratering, spalling,
delaminationandcrackingproducedatotaldamageareaofabout3%.
Whilst these do not appear to be particularly high figures, the mission duration was only
approximately6years.Fora30yearmission,thetotalsurfacedamageareaapproaches15%;ignoring
anyincreaseinmanmadedebrisduringtheyearsofdeployment.
Forcomposites,thedamageisnotjusttotheimmediatesurface,buthasavolumeeffect.
20.16.7 Protective shielding
20.16.7.1 General
Theincreaseinmanmadedebrisovertheyearshasincreasedtheprobabilityofcatastrophicimpact
by a large particle. This is a very serious possibility for a manned, pressurised spacecraft such as
ColumbusandISS.
342
ECSSEHB3220Part2A
20March2011
Thecriticalparticlesizeonwhichprobabilitytheorycalculatesthelikelihoodofacriticalimpactalso
increases. The implications for structural design are that theory should be used to define the
maximumsize of particlewhich islikely to be encountered during a30yearmission,andadequate
protection provided against it. If shielding is required, it has to be lightweight and effective in
containingthepredictedimpactenergies,Ref.[2036],[2037].
Thesimulationofimpactscanbeachievedintestfacilities,Ref.[2038],butwithavelocitylimitation
of8km/sandcentimetresizedparticles;whereas90%ofspacedebrishasavelocitygreaterthanthis
andaremuchsmallerparticles.
20.16.7.2 Columbus bumper shield concepts

To protect the manrated Columbus structure against large debris, an offset bumper shield
arrangement has been proposed, Ref. [2039], [2040], [2041]. Such a shield should be orientated to
meet a specific ballistic threat, butalso needs to belightweight.Constructions based on aluminium,
GLARElaminatesandKevlarcompositeshavebeenevaluated,[Seealso:46.17forGLARE].
The shield construction has several functions in containing the kinetic energy of the debris particle
andsubsequentspalledmaterial.Anoffsetarrangementofpanelsisnecessary,fixedatupto120mm
from the primary structure. The typical areal mass for these types of shield is 13.5 kg/m2, which
includes the pressure module wall as part of the shield system, Ref. [2041]. Hence the shield has a
structural,aswellasprotective,function.
20.17 References
20.17.1 General
[201] J.Dauphinetal
Materialsforspaceapplication
MetalsandMaterials,July1991,p422430
[202] B.D.Dunn
MetallurgicalAssessmentofSpacecraftParts&Materials
Published by Ellis Horwood Limited, ISBN-0-7458-0762-3, 1989
[203] B.Fornari&D.Dosio
CharacterisationofastateoftheartUHMCFRPprepregsystemfor
satelliteapplication
Proceedings of the International Symposium on Advanced Materials for
Lightweight Structures '94, ESTEC , Noordwijk
March 1994, ESA WPP-070, p569-576
[204] P.George
'Space environmental effects on LDEF low earth orbit exposed graphite
reinforced polymer matrix composites'
NASA Conference Publication 3162, Part 2. Proceedings of a Workshop held
at H.J.E Reid Conference Centre, NASA Langley Research Centre, 19-22
November 1991, p543-570
[205] R.C.Tennysonetal
'Proposed test program and data base for LDEF polymer matrix composites'
343
ECSSEHB3220Part2A
20March2011
NASA Conference Publication 3162, Part 2. Proceedings of a Workshop held
at H.J.E Reid Conference Centre, NASA Langley Research Centre, 19-22
November 1991, p593-600
[206] G.L.Steckeletal
PolymermatrixcompositesLDEFexperimentsM00039&10
NASAConferencePublication3162,Part2
ProceedingsofaWorkshopatH.J.EReidConferenceCentreNASA
LangleyResearchCentre,1922November1991p515542
[207] C.Blair&B.C.Petrie
Lowearthorbitenvironmentaleffectsoncompositematerials:results
fromthelongdurationexposurefacility(LDEF)
24thInternationalSAMPETechnicalConference,2022October992,
pT186T200
[208] Propertychangesinducedbythespaceenvironmentincomposite
materialsonLDEFexperimentAO171(Sample)
NASACR192572,April1993,18pages
[209] R.C.Tennyson
Additionalresultsonspaceenvironmentaleffectsonpolymermatrix
compositesexperimentalAO180
NASAConferencePublication3162,Part2.Proceedingsofa
WorkshopheldatH.J.EReidConferenceCentre,NASALangley
ResearchCentre,1922November1991,p571592
[2010] J.P.Wightman
SurfacecharacterisationofLDEFmaterials
NASACR194519,October1993
[2011] DiscussionSessions:NASAConferencePublication3162,Part2,
ProceedingsofaWorkshopheldatH.J.EReidConferenceCentre,
NASALangleyResearchCentre,1922November1991.
Polymersandfilms(IncludingAg/FEP),p707717
Polymermatrixcomposites,p727736
Lubricants,adhesives,seals,fasteners,solarcells&batteries,
p737747
[2012] B.Keough
Identificationandevaluationoflubricants,adhesives&sealsusedon
LDEF
NASAConferencePublication3162,Part2.Proceedingsofa
WorkshopheldatH.J.E.ReidConferenceCentre,NASALangley
ResearchCentre,1922November1991,p603618
[2013] D.F.Thompson&H.W.Babel
MaterialapplicationsonthespacestationKeyissuesandthe
approachtotheirsolution
SAMPEQuarterly,October1989,p2733&34thInternationalSAMPE
Symposium,811May1989,p759769
344
ECSSEHB3220Part2A
20March2011
[2014] W.Hartung
InfluenceofthermalcyclingonthebehaviourofCFRPmaterialfor
spacestructures
ESASP225,August1984
[2015] D.P.Bashford
MicrocrackingbehaviourofepoxyandcyanateesterUHMCFRP
composites
ERATechnologyReport940362,FinalReportonWorkOrderNo.34,
ESTECContract7090/87/NL/PP,April1994
[2016] ESAPSS01609Issue1(May1993):Radiationdesignhandbook
[2017] MatraEspace
ReportDM51.C/CC/FUO13882
[2018] R.C.Tennysonetal
TheeffectsofcombinedUVradiationandhighenergyelectronson
thebehaviourofpolymermatrixcompositesinhardvacuum
AGARDCP327,September1982
[2019] G.F.Sykes&D.E.Bowles
Spaceradiationeffectsonthedimensionalstabilityofatoughened
epoxygraphitecomposite
SAMPE Quarterly, July 1986, p39-53
[2020] R.N.Yancyetal
'Radiation and thermal effects on the time-dependent response of T300/934
graphite/epoxy'
34th International SAMPE Symposium
8-11 May 1989, p1174-1189
[2021] J.G.Funk
'Effects of cure temperature, electron radiation and thermal cycling on
P75/930 composites'
NASA Technical Memorandum 102712, Dec. 1990
[2022] A.Paillous&C.Pailler
'Degradation of multiply polymer-matrix composites induced by space
environment'
Composites, Vol. 25, No. 4, 1994, p287-295
[2023] K.T.Kernetal
'Simulated space environmental effects on a polyetherimide at its carbon
fibre-reinforced composites'
SAMPE Journal, Vol. 29, No. 3, May/June 1993, p29-44
[2024] L.J.Legeretal
'A consideration of atomic oxygen interactions with Space Station'
AIAA-85-0476, January 1985
[2025] D.P.Bashford
'A guideline on the response of composite materials to environmental
conditions encountered by space structures and launch vehicles'
Fulmer Research Laboratories. ESA Contract: 5497/83
345
ECSSEHB3220Part2A
20March2011
[2026] C.R.Stidhametal
Low earth orbital atomic oxygen environmental simulation facility for space
materials evaluation
10-13 May 1993, Vol. 1, p649-663
[2027] C.Barbieretal
'ATOX: The ESTEC atomic oxygen simulation facility'
Preparing for the Future, Vol. 2, No. 2, June 1992
[2028] S.K.Rutledgeetal
'Evaluation of atomic oxygen resistant protective coatings for fiberglass-
epoxy composites in LEO'
8-11 May 1989, p1163-1174
[2029] J.H.Sandersetal
'Application and performance of silicon-based atomic oxygen protective
coatings'
8-11 May 1989, p1152-1162
[2030] S.L.Oldham
'Atomic oxygen resistant, thermally stable epoxies'
21st International SAMPE Technical Symposium, 25-28 September 1989,
p984-993
[2031] C.Arnoldetal
'Siloxane modified cyanate ester resins for space applications'
9-12 March 1992, p128-136
[2032] S.Palsule
'Anoxic molecular composites: The third-generation polymers for aerospace
applications'
Lightweight Structures '94, ESTEC, Noordwijk
March 1994, ESA WPP-070, p577-580
[2033] C.R.Maag
'Low earth orbit debris effects on materials'
IAF-91-284 Paper from 42nd Congress of the International Astronautical
Federation, Montreal, Canada, 5-11 October 1991
[2034] M.Allbrooks&D.Atkinson
'The magnitude of impact damage on LDEF materials'
NASA-CR-188258, July 1992
[2035] J.Zhang&D.Rex
'The distribution of debris impacts over satellite surfaces with special regard
to the long duration exposure facility LDEF'
Z. Flugwiss. Weltraumforsch. 17 (1993), p170-180
[2036] M.Lambert
'Shielding against orbital debris - A challenging problem'
Proceedings of an International Symposium on Space Applications of
Advanced Structural Materials, ESTEC, Noordwijk
March 1990, ESA SP-303, p319-328
346
ECSSEHB3220Part2A
20March2011
[2037] M.Lambert
Shielding against natural and man-made space debris: A growing challenge
ESA Journal 1993, Vol. 17, p31-42
[2038] E.Schneider&M.Lambert
'Space debris simulation facilities and high speed diagnostic methods'
Lightweight Structures 94, ESTEC, Noordwijk
March 1994, ESA WPP-070, p385-390
[2039] HG.Reimerdesetal
Ballistic limit equations for the Columbus-Double Bumper Shield concept
Proceedings of the First European Conference on Space Debris Darmstadt,
April 1993, ESA SD-01, p433-439
[2040] HG.Reimerdes
Meteoroid and debris protection study (MDPS) Executive summary
ESTEC Contract No. 8732/90/NL/PP, ERNO Raumfahrttechnik GmbH
contractor report CR(P) 3615, October 1992
[2041] M.Lambert
Shielding against orbital debris - A comparison between different shields
Internal ESTEC Working Paper No. 1740
[2042] N.L.Johnson&D.S.McKnight
Artificial space debris
Orbit Book Company, Malabar, Florida, USA, 1987
[2043] ESAPSS01202:Preservation,storage,handlingandtransportationof
ESAspacecrafthardware
[2044] ESAPSS01609:Radiationdesignhandbook.
20.17.2 ECSS documents

[See:ECSSwebsite]

ESAPSS01202: Preservation,storage,handling&
transportationofESAspacecraft
hardware.
ESAPSS01609: Radiationdesignhandbook.
ECSSQST70 Materials,mechanicalpartsand
processes;previouslyESAPSS0170.
ECSSQ7071 Dataforselectionofspacematerialsand
ECSSQST7002 Thermalvacuumoutgassingtestforthe
screeningofspacematerials;previously
ESAPSS01702
ECSSQST7004 Thermaltestingfortheevaluationof
spacematerials,processes,mechanical
partsandassemblies;previouslyESA
347
ECSSEHB3220Part2A
20March2011
PSS01704.
ECSSQST7006 ParticleandUVradiationtestingfor
spacematerials;previouslyESAPSS01
706.
ECSSQST7029 Determinationofoffgassingproducts
frommaterialsandassembledarticlesto
beusedinamannedspacevehiclecrew
compartment;previouslyESAPSS01
729.
ECSSEHB3223 Threadedfastenershandbook;previously
ESAPSS03208.
ECSSEST3208 Materials
NOTE ECSS-Q-ST-70 and ECSS-Q-70-71 provides information on materials

and process selection for space applications.
348
ECSSEHB3220Part2A
20March2011
21
Bonded joints
21.1 Introduction
Design, calculation information and data on adhesively bonded joints are presented. General
guidelinesontheselectionofmostpracticaljointtypesarealsoprovided.
[Seealso:ECSSE3221Adhesivebondinghandbookforadvancedstructuralmaterials;previously
ESAPSS03210]
21.2 Adhesives
21.2.1 General
The use of adhesives demands detailed technical specifications. Joint design should be be carefully
studied,takingintoaccounttemperatureandarea.
Ingeneral,adhesivesshouldbeabletoexhibitseveralproperties,including:
Straincapability;toaccommodatejointsbetweendissimilarmaterials.
Cureataslowatemperatureaspractical.
CoefficientsofthermalexpansionofadhesivesareintheCTErangeofthejointparts.
minimalMoistureeffects.
Thicknesscannotbetoolarge.
Compatibilitywithadherends.
Conformtooutgassingrequirements,[See:ECSS-Q-ST-70-02].
21.2.2 Types of adhesives
21.2.2.1 General
Adhesivesarebroadlyclassifiedaseither:
Film,or
Paste.
349
ECSSEHB3220Part2A
20March2011
Mostsuppliersdonotprovidetheinformationneededtoenabletheselectionofthemostappropriate
adhesivefordesignwork.Thisincludestheeffectofmoisture,whichshouldbecarefullyconsidered
intheselectionprocess.
Table21.2.1 lists some chemical groups ofadhesives that have been used in spaceapplications,Ref.
[211].TheirtypicalcharacteristicsareshowninTable21.2.2,Ref.[211].
Table 21.2-1 - Adhesives: Common types
Maximum Service Shear Strength

Adhesive Type
Temperature (C) (MPa)
Epoxy 90 36
Epoxyphenolic 200 25 - 40
Epoxypolyamide 100 25
Polyimide 400 14 - 18
Silicone 340 14
Phenolic-Polyamide 150 36
Phenolic-Vinyl 100 36
Polyamide 100 40 - 60
Key: Shear strengths at room temperature
350
ECSSEHB3220Part2A
20March2011
Table21.22Adhesives:Typicalcharacteristics
Vacuum Outgassing % Tensile Strength Tensile Modulus Elongation Thermal Expansion
Adhesive Trade Name -6 -1
TML CVCM (MPa) (MPa) (%) (10 C )
Aerobond 2143 1.22 0 42.7 1240 1 58.9
FM 53 1.6 0.01 44.0 2101 1 88.9
FM 73 1.47 0 52.4 2136 3 74.0
FM 400 0.71 0.02 41.3 7235 1 46.6
A-452 2.24 0.26 57.9 041 3 80.3
C7/W 1.43 0.02 45.5 2019 3 66.4
Plastilock 717A 1.53 0.58 55.1 075 3 75.6
Araldite AV100 1.1 0.07 1,4.5 2604 - 59.9
Araldite AV138 0.7 0.02 17.7 9164 - 54
Redux 312 1.1 0.05 43.4 - - -
Hysol 9309 1.33 0 37.0 2439 3 48.4
Hysol 9601 1.37 0.3 61.3 2556 1 34.4
Epibond 123/945 0.91 0 26.87 5305 1.3 52.4
Epibond 123/931 0.92 0 62.7 965 1 44.5
RTV 566 0.2 0.02 6.89 - - 648
RTV-S 691 0.2 0.01 6.2 - - 648-129.6
AF 126 1.59 0.27 22.7 2564 12 78.7
AF 143 1.52 0 40.23 2198 2 84.6
AF 163-2M 1.29 0.26 49.6 2060 6 96.1
AF 163-K 1.65 0.23 37.2 2039 2 86.0
EC 2214 1.02 0.11 50.9 5870 1 57.2
EC 2216 1 0.02 20.7 - - -
Metlbond 329 85 0.02 61.3 6856 1.5 78.3
Metlbond 1133 1.67 0 44.1 1915 3 107.3
351
ECSSEHB3220Part2A
20March2011
21.2.2.2 Epoxy
Epoxyresinadhesivesarethemostcommontypesforspaceapplications.Theyareobtainedaseither:
Twopartmixtures(resinplushardener)forselfcuringatroomtemperatures,or
Onepart(resin)foraheatcuringprocess.
Epoxy resins without hardeners can be cured simply by keeping the assembly at the stated
temperatureforthespecifiedtimewithoutanincreaseinpressure.Inthiscase,thecuringtemperature
canbeabout180Candthecuringtimeisverylong.
Inordertoobtainthemaximumstrengthatelevatedservicetemperatureandaveryshortcuringtime
atlowtemperature,itisadvisabletouseahardener,withorwithoutanaccelerator.Inthiscase,the
curingtimecanbeasshortas20minutesat145Cunderapressureof0.7MPa.
Modifiedepoxyresins,i.e.epoxyphenolicandepoxypolyamide,adhesivescanbecuredforabout1
hourbetween130Cand175C.
21.2.2.3 Polyimides
Polyimideresinadhesivescureattemperaturesbetween250Cand400C.Apostcureisnecessaryto
obtainthehigheststrengthsandarangeofservicetemperatureoftheorderof290Cto400C.
21.2.2.4 Fluorocarbon (Viton)

Fluorocarbonpolymerresinadhesivesareusedforspecialhightemperatureapplications.
21.2.2.5 Polyamides
Polyamideresininadhesivesproducesmodifiedepoxyresinsandphenolicpolyamideresins.
21.2.2.6 Silicones
Siliconeadhesiveshaveexcellentbehaviourathightemperaturesbutrelativelylowstrength.Forthis
reason they are mixed with other resins to providehightemperature stabilitywith high mechanical
strength.Forexample,epoxysiliconeadhesivecanhavemaximumservicetemperaturesof340C(at
continuousexposure)and510C(atintermittentexposure).
21.2.2.7 Phenolic
Phenolicresinadhesivesforspaceapplicationsareusedinmixtureswithotherresins,including:
Phenolicpolyamide: Shear strength of the order of 36 MPa and excellent strength at high
temperatures,e.g.1.4MPaat150C.
Phenolicvinyl (or vinyl phenolic): Shear strength of the order of 36 MPa, but the range of
servicetemperatureisverylowanddegradesquicklyabove100C.
21.2.3 Adhesives for joining different materials

Figure21.2.1showsthetypesofadhesiveswhichareapplicableforjoiningdifferentmaterials.
Wherenoadhesivesareshown,thematerialscanbebondedwithdifficulty.Bondingispossibleafter
asurfacetreatment,whichincreasesthecriticalsurfacetension.
352
ECSSEHB3220Part2A
20March2011
Figure21.21Adhesivesforbondingdifferentmaterials
353
ECSSEHB3220Part2A
20March2011
21.3 Design of bonded joints
21.3.1 Basic considerations
21.3.1.1 Adherends
Todesignabondedjoint,knowledgeofthepartstobejoined(theadherends)isnecessary,including:
Ultimatestrengthsforcomposites,
Yieldstrengthsformetals,
Elasticitymoduli,
Coefficientsofthermalexpansion,
Maximumelongation,and
Chemicalcomposition.
21.3.1.2 Loading
Allthepossibleloadingmodesthatthebondedjointistowithstandduringitsservicelifeshouldbe
considered.
21.3.1.3 Environment
Both the working environment and the exposure duration should be known, e.g. thermal and
moistureconditions,andcombinationsthereof.
21.3.2 Basic guidelines

Fundamentalpointstobeconsideredbeforeajointisdesignedare:
Thewholejointareawithstandstheloadsactingonthejoint.
The loads applied result mainly in shear stresses. Cleavage, peel stresses and stress
concentrationsatfreeendsareeitheravoidedorminimised.
Fromexperience,thebestadhesivebondshaveanadhesivelayerthicknessbetween0.1mmand
0.25mm.
Inspectionofthejointduringitsworkinglifeiscrucial.
21.3.3 Failure modes

Thepredominantfailuremodesdependupon:
overlaplength,
adhesivethickness,and
fibreorientationinthelayerofcompositeadjacenttotheadhesive.
354
ECSSEHB3220Part2A
20March2011
Forexample:
Long lap lengths and a thin adhesive film, giving a lengthtothickness ratio of 50, tend to
producetensionorcompressionfailuresintheadhesives.
Lengthtothicknessratiosof25tendtoproduceshearfailuresintheadhesivesorinterlaminar
areasdependingonthejointconfiguration,[Seealso:21.4forjointconfiguration].
21.3.4 Features
21.3.4.1 Polymer composites

Jointsbetweentwoormorecompositelayerscanbeachievedby:
Secondary bonding,usinganadhesive(oftenafilm),
Co-curing:
usingtheadhesivepropertyoftheprepreg,
byincorporatinganadhesiveintoacompositeassemblypriortocuring.
The objective and function of an adhesive joint is to provide load transfer between two adjacent
components by the most efficient means. The anisotropic properties of composites should be taken
intoaccount,suchas:
stiffnessandstability,
thermalexpansion,
loadbearingrequirements.
Bondingismoreappropriateforlightlyloadedjointsinwhichitshighefficiencycharacteristicscanbe
fullyutilised.
Damagefromdrillingormachiningmakesorthotropicmaterials(suchascomposites)susceptibleto:
interlaminarshear,
delamination,and
plypeeling.
Consequently, bonding is usually preferred over mechanically fastened joints between advanced
compositecomponentsortometalparts.
Sincemechanicalloadsonjointsinspacecraftareoftenlow,theimportantcharacteristicsare:
dimensionalstability,
thermalexpansion,and
lowweight.
Thermal cycling from 132C to +121C can impose high stresses on the bondlines of adhesively
bondedjoints.
355
ECSSEHB3220Part2A
20March2011
21.3.4.2 Advantages
Themainadvantagesofbondedjointsare:
Loaddistributiononalargearea;avoidingloadconcentrations.
Highjointefficiencyindex;relationshipofstrengthandmassofthejointregion.
Lowpartcount.
Nostrengthdegradationofbasiclaminatebyuseofcutoutsormachiningoperations.
Lowpotentialcost.
Corrosionproblemsareminimisedforbondingdissimilarmaterials.
21.3.4.3 Disadvantages
Thedisadvantagesofbondedjointscanbesummarisedas:
Theycanbedifficulttoinspectnondestructively.
Designsshouldavoidorminimisepeelloadings.
Efficientstructuralbondsrelyonaccuratematingofadherends.
Theyarepermanentandcannotbedisassembled.
Thermalcyclingandhighhumiditycanaffecttheirstrength.
Specialsurfacepreparationsareneededpriortobonding.
21.3.4.4 Non-polymer composites

The family of fibre metal laminates FML, i.e. ARALL and GLARE, comprise of fibrereinforced
adhesivelayersthatareusedtobondtogethersheetsofconventionalmetals,[See:46.17].Components
madefromthesehybridmateriallaminatescanalsobeassembledusingadhesivebonding.
Adhesive bonding is widely used in the aerospace industry for the assembly of conventional
structural metal alloys, e.g. sheet aluminium components. The structural integrity of honeycomb
panels relies on adhesively bonded joints between the face skins and the core, [See also: 26.16 and
ECSSEHB3221].
Thepottingofinsertsandotherattachmentpointsintohoneycombpanelsusesadhesivepastes,often
thickenedtoimprovehandlingcharacteristics,[See:ECSSEHB3222].
Bondingisusedtoassemblesolarpanelsandalsowithinsomeelectricalandelectronicsubassemblies,
[See:ECSSQ7071].
[See also: Chapter 58 for joint selection factors associated with metal and ceramicbased advanced
materials]
356
ECSSEHB3220Part2A
20March2011
21.4 Joint configuration
21.4.1 Basic configurations
21.4.1.1 General
ExamplesofeachtypeareshowninFigure21.4.1,Ref.[211].
Thesearebasicconfigurationsandnotnecessarilypracticaldesigns.
21.4.1.2 Single lap joints

Featuresofsinglelapjointscanbesummarisedas:
Itisthesimplestbondedjointconfiguration.
Thehigheststressconcentrationsinthistypeofjointareatthefreeendsofthebond.
Thecentreofthejointtransmitsverylittleload.
The bevelled and radiused singlelap joints improve on simple joint behaviour, minimising
stressconcentrations.
21.4.1.3 Double lap joints

Featuresofdoublelapjointsare:
Thesearemorecomplexthansinglelapjoints.
Theyeliminatemostofthebendingandpeelstresses.
Numerousapplicationsarepossibleforjointsofthistype.
21.4.1.4 Strap joints

Featuresofstrapjointscanbesummarisedas:
Thistypeofjointisamoresophisticatedsolutionthandoublelap,butitisbasedonthesame
principle.
The doublelap advantages are improved by applying special design to each application or
problem.
21.4.1.5 Stepped-lap joints

Featuresofsteppedlapjointscanbesummarisedas:
Theyareverydifficulttomachine.
Theyarenotapplicabletothinadherends.
Are used to improve the shear strength at joint whose adherends are of very thick section.
Nevertheless,theplystackingsequenceduetodifferentplyorientationsisagreatproblem
sinceitcanresultinnon0orientedfibreareasonthesurface,fornonunidirectionalcomposite
bondingcases.
Theyproduceverygoodstressdistributions.
357
ECSSEHB3220Part2A
20March2011
Theyhaveahighjointefficiencyindex(strengthtoweight).
21.4.1.6 Scarf joints

Scarfjointsareverysimilartosteppedlapjoints,butareeasiertomachine.
21.4.1.7 Shim insert joints

Shiminsertjointsareusedtoincreasejointsection,e.g.forinserts.
21.4.1.8 Other configurations

Thesecanbedescribedasdesignsthatprovideimprovementsonthebasicconfigurationsforspecial
applications.
358
ECSSEHB3220Part2A
20March2011
Figure21.41Bondedjoints:Basicconfigurations
359
ECSSEHB3220Part2A
20March2011
21.4.2 Orientation of surface fibres

Thedirectionofthesurfacelayeroffibresinajointisspecifiedinthecompositelayup,sothatthey
areinthedirectionoftheload.
Figure21.4.2indicatesthefibreorientationaccordingtoloaddirections,Ref.[213].
Figure21.42Bondedjoints:Orientationofsurfacefibre
360
ECSSEHB3220Part2A
20March2011
21.5 Environmental factors for bonded joints
21.5.1 General
Temperatureandmoistureareconsideredindesigningcompositestructures,owingtotheirinfluence
oncompositecomponentproperties.
21.5.2 Effect of moisture

Absorbed moisture moves through the adhesive resin by capillary action producing a gradual
softeningplasticisation,causing:
swelling,and
loweringoftheglasstransitiontemperature,Tg.
Experimental results indicate that moisture absorption levels in excess of 0.6% can be withstood
withoutanydecreaseinadhesivestrength.
Figure 21.5.1 compares the variation of adhesive shear stress along the joint, due to a constant end
load,Ref.[214].
361
ECSSEHB3220Part2A
20March2011
Key: [a] Adhesive uniformly dry

[b] Adhesive softened by moisture absorption at edge
[c] Adhesive partially dried due to change in external environment
max: Adhesive maximum shear strain
Figure21.51Effectofmoistureonstressdistributionsinadhesivebondedjoints
Figure21.5.2(A)showsasimpletransitionwherethereisadryadhesiveontheoutsidepart,inFigure
21.5.2(B)therearetwostaggeredtransitionswithamoistadhesiveontheoutside.Inthesecondcase,
therearethreeratherthantwoadhesivestatesresponsiblefordefiningtheoutermostadhesiveshear
strain,Ref.[214].Thesephenomenahavebeenobservedforuniformdoublelapjoints.
362
ECSSEHB3220Part2A
20March2011
Bothillustrationsdrawntothesamescale
Figure21.52Influenceofmoistureabsorption/desorptiononpeakadhesive
shearstrains
Figure 21.5.3 shows that, for a bonded doubler or wide overlap joint, continuing penetration of the
joint by moisture has little effect on the most critically loaded adhesive near the end of the overlap
oncethemoisturehaspenetratedpastthatedgearea,Ref.[214].
363
ECSSEHB3220Part2A
20March2011
Figure21.53Effectofprogressivemoistureabsorptiononbondstrain
21.5.3 Effect of temperature

Temperatureeffectsbythemselvesarenotimportantiftheadhesiveisusedwithinitsdefinedservice
temperaturerange.
21.5.4 Combined moisture and temperature

The worst case occurs under the influence of both temperature and moisture, because a high
temperatureincreasestheabsorptionanddiffusionofmoisture.
Guidelinestoavoidthedetrimentaleffectsofmoistureandtemperatureare:
Define accurately the worst case of environmental conditions, which the bonded joint has to
withstand.
Defineexactlytheservicetemperaturerange.
364
ECSSEHB3220Part2A
20March2011
Select the most adequate type of adhesive resin, e.g. epoxy, polyimide, silicone, taking into
account both the worst case conditions and the service temperature range of the particular
adhesive.
21.6 Bonding defects
21.6.1 General
Themostcommonbondingdefectsare:
Flaws.
Porosity.
Alsoimportantare:
Variationinthickness,
Undercuring,
Variationinresinfraction,and
Variationindensity.
Theimportanceofthedefectsdependsonsuchfactorsas:
Theextenttowhichdefectsarepresent.
Theirconsequences(criticalornot).
Whetherdefectsarerandomorlocallyconcentrated.
Whetherdefectsareindicativeofdegradationprocessesornot.
Whetherdefectsareindicativeofmaterialdeficienciesornot.
21.6.2 Description of bonding defects
21.6.2.1 Flaws and porosity

Anybondingdefectsleadtoloadredistributionthroughtheadhesivefilm.Thisideamightsuggestan
increaseinpeakstressesatdiscontinuitypointsintheadhesivefilm.However,theactualeffectisless
than expected and only an imperceptible increase in stress occurs in most cases. This assessment is
validonlywhentheflawsizeisproportionallysmallwithrespecttojointsize.
Figure21.6.1givesthepossiblyacceptableflawsizesaccordingtoaZone1(criticalzone)andaZone2
(lesscriticalzone),asclassifiedinRef.[214].Itindicatestheflawsizesthatareacceptableinprimary
structures,dependingonthezoneinwhichtheflawappears.
The limits for secondary structures in use throughout the aerospace industry are currently more
stringentthanthosegiveninFigure21.6.1.
365
ECSSEHB3220Part2A
20March2011
Figure 21.6-1 - Bonded joints: Examples of acceptable bond flaw sizes
Figure21.6.2depictsthreedifferentexamplesofstructuraljointsandtheircriticalzones,Ref.[214].
366
ECSSEHB3220Part2A
20March2011
Key: Zone 1: Critical.

Zone 2: Less critical.
Figure 21.6-2 - Typical quality zoning for bonded joints
367
ECSSEHB3220Part2A
20March2011
Figure21.6.3showthenormaladhesivestressdistributioninadefectfreebondedjoint.
ThejointconfigurationisthesamefromFigure21.6.3toFigure21.6.6inclusive.
Figure 21.6-3 - Adhesive shear stresses in bonded joints
TheeffectofanedgeflawontheshearstressdistributionisshowninFigure21.6.4.Whentheposition
ofthesamesizedflawischanged,theeffectonshearstressdistributionisshowninFigure21.03.5.
Figure 21.6-4 - Adhesive stresses in flawed bonded joints
368
ECSSEHB3220Part2A
20March2011
Figure21.6.6illustratestheeffectofalargecentralflawonshearstressdistribution.
InFigure21.6.7theeffectsofflawsinshortandlongoverlapbondsunderthesameappliedloadare
compared. Of particular interest is the vertical gradient over which the maximum adhesive shear
stress induced is reduced from its failure value to a much lower value, i.e. independent of overlap
length.
369
ECSSEHB3220Part2A
20March2011
Theeffectofflawsisinconsequential,exceptwhentotaleffectiveoverlapisbarelysufficienttocarry
entireloadplastically.
Figure 21.6-7 - Adhesive bonded joints: Effect of debond flaws on flexibility
Moredetailedanalysesareneededtocoverthecasesinwhichbondflawsaresolargethatthereisa
possibility of complete tearing of the adhesive film. Such analyses need to account for two
dimensional load redistribution around flaws as well as the onedimensional redistribution
consideredhere.
21.6.2.2 Variation in thickness

Experiencehasshownthatthebestresultsinbondingofcompositesareobtainedwhentheadhesive
filmshaveathicknessrangingfrom0.12mmto0.25mm,Ref.[214].Inpracticethisoptimumthickness
often cannot be achieved. Figure 21.6.8 shows the variation in peak induced adhesive stress with
adhesivethicknessattheendsofthebond,Ref.[214].
[Seealso:Variationinresinfraction]
370
ECSSEHB3220Part2A
20March2011
Figure 21.6-8 - Variation of peak induced adhesive stress with thickness of adhesive at
ends of overlap
371
ECSSEHB3220Part2A
20March2011
21.6.2.3 Undercuring
Undercuringoccursduetoinsufficienttimeortemperatureforadequatecuringoftheadhesive.
21.6.2.4 Variation in resin fraction

Thisoccurswherethereisanunevendistributionofadhesiveonthebondsurfaces;eitheranexcessive
orscantquantity.Thedefectisusuallyconsideredasavariationinthickness.
21.6.2.5 Variation in density

Thisoccursduetothepresenceofflawsandporosity,butalsoavariationinresinfraction.
21.6.3 Inspection of bonded joints

NondestructivetestmethodscommonlyusedtodetectdefectsaredescribedinTable21.6.1.Someof
thesemethodscanbedifficulttouseinpractice.
Table21.6.2summarisesthepossibilityofdefectdetectioninbondedjointsbyvarioustechniques.
[Seealso:Chapter34fordetailsofNDTinspectiontechniques]
Table21.61Bondedjoints:DescriptionofNDTmethods
Method Description
Very effective for examining the uniformity of adhesive joints
Radiography (X-ray) and intimacy of contact in bonded areas when the adhesives
used are not radiation transparent.
Used to determine adhesive build up or variation in resin
Radiography (Neutron)
fraction.
Radiation (Gamma) Used to detect changes in thickness or density of adhesive.
Thermochromic or photochromic compounds are added to the
Thermography adhesive system. Used to determine the degree of adhesive
curing.
Sound waves ranging in frequency from 1 to 10 MHz are used
to find:
changes in thickness
detect porosity
delaminations or un-bonded areas.
Ultrasonic
There are three basic ultrasonic systems:
Pulse-echo reflector: To detect flaws or/and
delaminations.
Transmission: To detect flaws.
Resonant frequency: To detect un-bonded areas.
Applicable to determine flaws or (through sonic microflows),
Acoustic Emission
porosity, undercure or un-bonded areas and variation in
(AE)
density.
Infrared (IR) Used to find delaminations, un-bonded areas and porosity.
Dye Penetrant Used to detect flaws, porosity and delaminations.
Used to detect porosity, undercured areas, delamination and
Induced Current
variations in thickness.
372
ECSSEHB3220Part2A
20March2011
Table21.62Bondedjoints:Detectionmethodsforvariousdefects
Radiography Dye Induced
Defect Thermography Ultrasonic AE IR
X ray Neutron Gamma Penetrant Current
Unbond
Under Cure
Variation in
Resin Fraction
Density
Thickness
Porosity
Flaws
Delamination *
Key: * If delamination is in the same orientation as X-ray beam.
373
ECSSEHB3220Part2A
20March2011
21.7 Bonded joint failure modes
21.7.1 Typical failure modes

Thefailurebehaviourofbondedjointsisanimportantconsiderationinordertoachieveanoptimum
design.
Thefailuremodedependsupondesignparameters,suchas,Ref.[216]:
jointconfiguration
overlaplength
adherendthickness,and
fibreorientationadjacenttotheadhesive.
Figure 21.7.1 shows five components of bonded joint strength which, if at fault, produce the typical
failuremodes.
Figure21.71Componentsofbondedjointstrength
21.7.2 Loading modes

Theloadingmodes,ortypesofstress,whichabondedjointcanundergoare:
peel,
tensile,
tensileshear,
torsion,
cleavage;forthicksections.
374
ECSSEHB3220Part2A
20March2011
Figure21.7.2illustrateseachoftheseloadingmodes.
Figure21.72Bondedjoints:Loadingmodesortypesofstresses
21.8 Calculation of bonded joint strength

Themethodsgivenforcalculatingjointstrength,[See:21.9foranalysisofjointconfigurations],arenot
closedformmathematicalsolutions.Nevertheless,theintroductionofvarioussimplificationsmakesit
possible to produce solutions which can readily be used for the various joint configurations, if it is
assumedthatthebehaviouroftheadherendandtheadhesiveiselastic.
ThecalculationmethodsarebasedmainlyontheworkofVolkersen,Ref.[215]andGolandReissner,
Ref.[216].
Twobasicapproachestothestudyofjoints,includingplasticityeffects,arefoundinHartSmith,Ref.
[217]orESDU,Ref.[218].
[Seealso:Chapter16forinformationoncomputerbasedjointanalysis]
[See:16.9,16.11,16.13,16.14forESDUpackages;16.39forESAComp]
375
ECSSEHB3220Part2A
20March2011
21.9 Analysis of joint configurations
21.9.1 Analytical notation
N Tensile load
Ncr Critical load
E Young's Modulus of adherends
G Adhesive shear modulus
K, , W, f Factors
dS Variation of surface energy
dE Variation of elastic strain energy
dP Variation of potential energy of peeling load
l Length of peeling
L Length of overlap
b Width
T Temperature
t Thickness of lap sheet
ta Thickness of adhesive
Poisson's ratio
m Average shear stress
max Maximum shear stress
Theoretical shear stress
Stiffness Ratio
[See:21.8foradescriptionofthemethodsused]
21.9.2 Single lap shear joint
21.9.2.1 Symmetrical
Figure21.9.1definestheanalyticalnotation.
376
ECSSEHB3220Part2A
20March2011
Figure21.91Analysis:Notationforsymmetricalsinglelapshearjoint
ThetensileloadNproducesashearstressmintheadhesivelayer.
N
m [21.91]
L
max K m [21.92]
where:
K 1 WL1 3 cothWL 31 [21.93]

4

2 1 xy G
W
12
[21.94]
Ett a
1

3 1 1 2
L
1 2 2 tanh
xy N

[21.95]
t 2 Et

xy xy yx [21.96]
Therelationshipbetweenthetheoreticalandaverageshearstressisgivenas:
377
ECSSEHB3220Part2A
20March2011

1 2 1 xy yx G
12
L1 3
cosh 2WX

31 [21.97]
m 4 Ett
sinh WL
a

Theassumptionsmadeare:
Theadhesiveflexuralrigidityisnegligible.
ThebehaviourofthelaminateandadhesiveduetothetensileloadNiselasticandisotropic.
Bendingeffectscanbeignored.
The normal and shear strains in the transverse direction of the laminate are negligible with
respecttothoseintheadhesive.
21.9.2.2 Analysis of R-degree peeling

Figure21.92Analysis:NotationforsinglelapjointRdegreepeeling
Figure21.09.3showsthedegreepeelingstrength
378
ECSSEHB3220Part2A
20March2011
Figure21.93degreepeelingstrength
Thecriticalconditionforpeelingis:
dS dE dP 0 [21.9-8]
With being the surface energy to cause adhesive fracture, Equation [21.99] is obtained from
Equation[21.91]:
1 cos 2
2
N cr 2bt1 1 cos N cr 4 b 2
0 [21.9-9]
E
11t E t
2 2
SolvingthisequationforNcr:
1 cos 2 2
2bt1 1 cos 4b t1 1 cos 16 b
2 2 2

E1 t 1 E 2 t 2
N cr 1
[21.9-10]
cos 2
2

E1 t 1 E 2 t 2
379
ECSSEHB3220Part2A
20March2011
21.9.3 Double lap shear joint

Figure21.94Analysis:Notationfordoublelapshearjoint
ThetensileloadNproducesashearstressmineachadhesivelayer:
N
m 2 L [21.9-11]
max Kt m [21.9-12]
where:

1 cosh WL
K [21.9-13]
WL
sinh WL
12
2G
W [21.9-14]
t a t 2 E 2
1

1 t2 E2 [21.9-15]
2 t1 E1
and: *thegreaterofor(1)
Therelationshipbetweenthetheoreticalandaverageshearstressisgivenas:
380
ECSSEHB3220Part2A
20March2011
1 coshWL
WL coshWX 1 sinhWX [21.9-16]
m sinhWL
Ingeneral,t(thickness)andL(length)valuesresultinWL<4,hence:
max *WL m [21.9-17]
When L
8
canbeapplied,Equation[21.918]isobtained:
t2
N
K

[21.9-18]

t 2 crit
where:K=K1K2andK1,K2aredefinedas:
2a
K1 2G [21.9-19]
ta
E2
K2 [21.9-20]
t2
381
ECSSEHB3220Part2A
20March2011
21.9.4 Double-lap shear joint under mechanical and temperature

loads
21.9.4.1 Joint with standard overlap length

Figure21.95Analysis:Notationfordoublelapjoint(standardoverlaplength)
Thedistributionofshearstressintheadhesiveduetoloadcanbedescribedbythedistributionshown
inFigure21.9.6.
Figure21.96Shearstressdistributionversustheadhesivelength(E1t1=E2t2)for
singlelapjointwithouteccentricity
Thetransferofforcefromoneflangetotheotherisconcentratedatbothedgesofthelapjoint.This
theoryleadstopeakshearstressesatx/L=0andx/L=1.Astheshearstressesattheedgesarezero;
theactualshearstresspeaksattheedgesaresmaller.
Theshearstressconcentrationfactor,usingthemaximumshearstress,isdefinedas:
382
ECSSEHB3220Part2A
20March2011
f max [21.9-21]
m
wheremisdefinedas:
N
2
m bL
[21.9-22]
As given in Ref. [2110], the factor f is constant for different joints when the correlation factor K is
constant.
GLb L GL2
K [21.923]
Etb t a Ett a
Thestresspeaksatbothedgesofthebondingcanbeobtainedbymeansoff1andf2.
f2canbecalculatedasafunctionofcorrelationfactorKandthestiffnessratioby:
f K 2 2 1 cosh 2 K 2 [21.9-24]
2
2 sinh 2 K 2
with:
E1t1 E 2 t 2
2 [21.9-25]
E1t1
Thefunctionf2=f(K2)isshowninFigure21.9.7fordifferentstiffnessratios2.
383
ECSSEHB3220Part2A
20March2011
Figure21.97Stressconcentrationfactorf2asafunctionofK2withparameter2
Thepeakshearstressesobtainedbyusingthelinearelastictheoryarereducedwhenthecharacteristic
shear stress/strain curve is non linear. This reduction is important only when a static analysis is
performed.Forlifeendurancecalculationsonlythelinearelasticregionofthecharacteristiccurveis
used.
384
ECSSEHB3220Part2A
20March2011
21.9.4.2 Large overlap length and adherends of the same materials and
thicknesses
Whenlapsheetsareofthesamematerial,E1=E2,t1=t2,therefore:
E1t1 E 2 t 2 2
2 [21.926]
E1t1
and:
max K 2 1 cosh 2 K 2 K 2 coth K 2 [21.9-27]

m 2 sinh 2 K 2 2 2
WhenK2>10,theterm coth K 2 isapproximately1,resultingin:

2
max K 2 GL2
L
G
[21.928]
m 2 2 Ett a 2 Ett a
G
max L m 2 Ett a
[21.929]
N
2
m Lb
[21.930]
N G
max 2b 2 Ett a
[21.931]
For long overlap lengths, the maximum shear stress is independent of the length of the joint.
Consideringthattheloadtransferisconcentratedattheendsofthejoint,theshearstressisreducedto
zerointhemiddlepartoftheoverlapzone,asshowninFigure21.9.7.
As the lap joint length increases, the mean shear stress decreases. This leads to higher shear stress
concentrationfactors.However,forK>10,thisbehaviourislinearasshowninFigure21.9.8.
385
ECSSEHB3220Part2A
20March2011
Figure21.98Shearstressdistributionforlargeoverlaplengths
21.9.4.3 Shear stresses in the adhesive due to temperature

Shear stresses occur in the adhesive due to different coefficients of thermal expansion, T , of the
adherends.
Forthiskindofloadingitisimpossibletodefineacomparativecorrelationfactor.Thusitseemstobe
reasonabletoformulatetheshearstressasafunctionofthenormalisedoverlaplengthx/Lduetoa100
Ktemperatureincrease;asshowninFigure21.9.9andFigure21.9.10.
386
ECSSEHB3220Part2A
20March2011
Figure21.99Stressconcentrationfactorforlargeoverlaplengths
387
ECSSEHB3220Part2A
20March2011
Figure21.910Shearstressdistributionversusthenormalisedbondinglengthx/L
forabondedjointbetweenCFRPHTunidirectional(60vol.%)andaluminium,
titanium,GFRPquasiisotropic,andGFRPunidirectionalmaterials
388
ECSSEHB3220Part2A
20March2011
Themaximumshearstressesare:
independentoftheoverlaplengthL,
but:
directly dependent on the product of the difference in temperature and the difference in the
coefficientofthermalexpansionoftheadherends.
Thetheorydoesnotconsiderthattheshearstressatx/L=0andx/L=1iszero.
Figure21.911Shearstressdistributionversusthenormalisedbondinglength
x L forabondedjointbetweenCFRPHTquasiisotropic(60vol.%)and
aluminium,titanium,GFRPquasiisotropic,andGFRPunidirectionalmaterials
389
ECSSEHB3220Part2A
20March2011
21.9.5 Single taper scarf joint

Figure21.912Analysis:Notationforsingletaperscarfjoint
N cos N cos
m L

t
sin [21.932]
max K m [21.933]
where:
2
1
tan 1
K E a
cot 1 [21.9-34]
4 E
and:
a xz E a [21.9-35]
E
Ingeneral,<20and Ea 1
E
Hence:K1andmaxm.
390
ECSSEHB3220Part2A
20March2011
21.9.6 Double taper scarf joint
21.9.6.1 Symmetrical
Figure21.913Analysis:Notationforsymmetricaldoubletaperedscarfjoint
N cos N cos N cos sin

m 2L

2t
2 sin
t
[21.936]
max K m [21.937]
where:
1
tan 1
K E a
cot 1 [21.938]
4 E
and:
a xz E a [21.939]
E
Thiscaseisalmostthesameasthesingletaperscarfjointwiththedifferencethatthenecessarybond
lengthishalfthatofthebondlengthofdoublelapshearjoint.
391
ECSSEHB3220Part2A
20March2011
21.9.7 Stepped lap joint
21.9.7.1 Recessed and simple


Figure21.914Analysis:Notationforsteppedlapjoint(recessedandsimple)
21.9.7.2 Three or fewer steps

For three or fewer steps, this type of joint can be analysed by taking each lap as a single lap of
thickness,t.
Therecessedandsimpleconfigurationsaresimilar,withtheonlydifferencethattheloadtransfer(in
thecaseofthreeorfewersteps)occursgenerallyaboutthecommonmidplane.Foreachstep:
N
m L
[21.940]
max K m [21.941]
where:
1
K WL1 3 coth WL 31 [21.942]
4

2 1 xy G
W
12
[21.943]
Ett a
392
ECSSEHB3220Part2A
20March2011
1

3 1 1 2
L
1 2 2 tanh
xy N

[21.944]
t 2 Et

xy xy yx [21.945]
Foreachstep:
NisN1,N2orN3
LisL1,L2orL3
tist1,t2ort3
Therearestressconcentrationsinthefirstcase(threeorfewersteps).
21.9.7.3 Four or more steps

Forfourormoresteps,theanalysisofdoublelapjointscanbecarriedoutbytakingthestepthickness
t1andt1/2,[See:Doublelapshearjoint].Figure21.9.15definestheanalyticalnotation.
Figure21.915Analysis:Notationforrecessedandsimplescarfjoints(fouror
moresteps)
393
ECSSEHB3220Part2A
20March2011
21.10 Bonded joint design curves and test data
21.10.1 General
The design curves present information on the shear stresses to be expected in bonded joints for
differentgeometries,L/t.Staticanddynamicloadsareconsideredtogetherwithtemperatureeffects.
21.10.1.1 Design curves

The information included on the various joint configurations under static and dynamic loading are
summarisedinTable21.10.1,whichalsoservesasanindextothedesigncurves.
Table21.101Bondedjoints:Summaryofdatacurves
Load Adherends Adhesive Temperature ( See: Figure
C)
Single Lap Joint
C/EC/E High Mod. RT and 180 Figure 21.10.1
C/E St High Mod. RT and 180 Figure 21.10.2
C/E St Med. Mod RT Figure 21.10.6
Static C/ESt or C/E Metlbond 329-7 RT and 180 Figure 21.10.4
C/EC/E Metlbond 329-7 RT, 120 and Figure 21.10.5
C/EC/E Metlbond 329-7 180 Figure 21.10.6
RT and 180
C/EC/E RT Figure 21.10.8
Dynamic Metlbond 329-7 180 Figure 21.10.9
C/EC/E
Double Lap Joint
Static C/E Ti Med. Mod - Figure 21.10.10
Symmetrical Scarf Joint
C/E Ti RT and 180 Figure 21.10.11
Static Metlbond 329-7 RT and 180 Figure 21.10.12
C/E Ti
C/E Ti RT and 180 Figure 21.10.13
Dynamic Metlbond 329-7 RT and 180 Figure 21.10.14
C/E Ti
Key: C/E: Carbon/epoxy : Shell 951
St: Steel : Intermediate strength
Ti: Titanium
394
ECSSEHB3220Part2A
20March2011
Adherends:Carbon/epoxytocarbon/epoxy
Adhesive:Highmodulus
Temperature:RTand180C
Figure21.101Designcurve:Singlelapshearjoint
395
ECSSEHB3220Part2A
20March2011
Adherends:Carbon/epoxytosteel
Adhesive:Highmodulus
Figure 21.10-2 - Design curve: Single lap shear joint
396
ECSSEHB3220Part2A
20March2011
Adherends:Carbon/epoxytosteel
Adhesive:MediummodulusShell951.
Temperature:RT
Figure21.103Designcurve:Singlelapshearjoint
Adhesive:Metlbond3297.
Figure21.104Designcurve:Lowerbondtensionsinglelapshearjointstrength
versuslaplength
397
ECSSEHB3220Part2A
20March2011
Temperature:RT,120Cand180C
Figure21.105Designcurve:Lowerbondtensionsinglelapshearjointstrength
versuslaplength
Figure21.106Designcurve:Lowerbondtensionsinglelapcompressionloaded
jointstrengthversuslaplength
398
ECSSEHB3220Part2A
20March2011
Adherends:Intermediatestrengthcarbon/epoxytocarbon/epoxy
Temperature:180C
Figure21.107Designcurve:SinglelaptensionfatigueSNcurve
Adhesive:Metlbond3297
Temperature:180C
Figure21.108Designcurve:SinglelapshearjointtensionfatigueSNcurve
399
ECSSEHB3220Part2A
20March2011
Adherends:Carbon/epoxytocarbon/epoxyortitanium
Adhesive:Shell951.Mediummodulus
Temperature:
Figure21.109Designcurve:Doublelapshearjoint
400
ECSSEHB3220Part2A
20March2011
Adherends:Carbon/epoxytotitanium
Figure21.1010Designcurve:Symmetricaldoublescarfjointstrength
401
ECSSEHB3220Part2A
20March2011
Adherends:Carbon/epoxytotitanium
Figure21.1011Designcurve:Symmetricaldoublescarfjointstrength
402
ECSSEHB3220Part2A
20March2011
Adherends:Intermediatestrengthcarbon/epoxytotitanium
Figure21.1012Designcurve:Symmetricaldoublescarfjoint,tensionfatigueSN
curve
403
ECSSEHB3220Part2A
20March2011
Figure21.1013Designcurve:SymmetricaldoublescarfjointtensionfatigueSN
curve
404
ECSSEHB3220Part2A
20March2011
Figure21.1014Designcurve:SymmetricaldoublescarfjointtensionfatigueSN
curve
21.11 Acoustic fatigue of bonded configurations

Astudyhasbeenmadeontheresponseofbondedcompositejointstorandomvibration,Ref.[2111].
ThecompositeusedwasHerculesAS4/35016inabalancedmultidirectional0/90/45construction.
BondingwasundertakenwithepoxypasteadhesiveHysolEA93462.
Thejointconfigurationsevaluatedwere:
Cocuredriser/skinjoint
Bondedrib/skinjoint,
Bondedrivetedrib/skinjoint,
Beadedskin(costlyconstruction),
Integratedriserskinjoint.
These are shown in Figure 21.11.1, Ref. [2111]. Vibration occurred at the natural frequency of each
specimen (<300 Hz). The root mean square, RMS, deformation, in m, was determined for each
configurationandfailuresaidtohaveoccurredifthetestfrequencyfellby5%.
405
ECSSEHB3220Part2A
20March2011
Figure21.111BondedjointconfigurationsinHerculesAS4/35016
Theconclusionsdrawnwerethatthe:
Integratedskinconceptshowedlittlesignofdamageafter106cycles.
Beaded skin construction exhibited a loss of stiffness after 105 cycles, although little damage
wasvisible.
Cocuredriser/skinandrib/skin(bothrivetedandunriveted)jointsexhibitedpeelingbefore106
cycles.
Riser/skin joints in bismaleimide Narmco 5250/T300 proved more prone to cracking than the epoxy
equivalents.
406
ECSSEHB3220Part2A
20March2011
21.12 References
21.12.1 General
[211] GeorgeLubin
'Handbook of Composites'
Van Norstrand Reinhold. Edition 1982
[212] D.P.Bashford:FulmerResearchLabs.(UK)
'Guidelines on basic aspects of jointing technology for advanced fibre
reinforced plastics used in space structures'
September 1984, ESA Contract No. 5497/83/NL/PB (SC)
[213] M.M.Schwartz
Composite Materials Handbook'
McGraw Hill. Edition 1983
[214] ASTMSTP749
[215] O.Volkersen
Luftfahrtforschung 15, 41; 1938
[216] M.Goland&E.Reissner
J. Appl. Mech., Trans. ASME 66, A17; 1944
[217] L.J.HartSmith
'Design and analysis of adhesive bonded joints'
McDonnell Douglas Co. Report No. 6059A; 1972
[218] ESDU79016
'Inelastic shear stresses and strains in the adhesives bonding lap joint loaded
in tension or shear', 1979
[219] R.Uhrig
'Elastostatik und Elastokinetik in Matrizenschreibweise'
Springer Verlag, Berlin, 1973
[2110] H.Hertel
'Leichtbau'
Springer Verlag, Berlin, 1960
[2111] SONACAS.A
'Shaker test results on typical structural CFRP joints under narrow band
random fatigue'
Report ESTEC/TP01/BE/R/M1, 7 July 1992
Work undertaken on Work Order No. 21
ESTEC Contract 7090/87/NL/PP
407
ECSSEHB3220Part2A
20March2011
21.12.2 ECSS documents

[See: ECSS website]
ECSSQ7071 Dataforselectionofspacematerialsand
ECSSQST7002 Thermalvacuumoutgassingtestforthe
screeningofspacematerials;previously
ESAPSS01702.
ECSSQST7029 Determinationofoffgassingproducts
frommaterialsandassembledarticlesto
beusedinmannedspacevehiclecrew
compartments;previouslyESAPSS01
729.
ECSSEHB3221 Adhesivebondinghandbook;
previouslyESAPSS03210.
ECSSEHB3222 Insertdesignhandbook;previouslyESA
PSS031202.
408
ECSSEHB3220Part2A
20March2011
22
Mechanically fastened joints
22.1 Introduction
Composites,likeotherstructuralmaterials,arejoinedtocreateusefullargerstructures.Thefunction
ofjointelementsistoconnecttwoormorepartstogetherinordertotransferloadbetweenthem.The
mannerinwhichtheseprocessesareperformedisadeterminingfactorintheefficiencyandsuitability
ofsuchcomponents.
The efficiency of a joint can be assessed in relation to its strength, mass and cost of manufacture. A
compromiseisinvariablynecessarybecausemaximisingstrengthusuallyleadstoincreasedmassor
increasedfabricationcosts.
The designer should consider the joint in relation to the whole structure, including dimensional
stability,andtheloadingconditionsimposedonthejoint.
Theanisotropicnatureofcompositesmeansitisnecessarytodesignjointsverycarefullyinorderto
avoidinterlaminarorshearfailure.
[See also: ECSSEHB3223 Threaded fastener handbook for further information on design;
previouslyESAPSS03208;ECSSQST7046forprocurementofthreadedfasteners]
22.2 Basic considerations for design
22.2.1 Advantages
Themainadvantagesofmechanicaljointsarethat:
Nospecialsurfacepreparationorultracleanhandlingoperationsarenecessary.
No adverse or irreversible strength effects occur due to thermal cycling and high humidity,
whereasbondedjointscanbe.
Nounusualinspectionproblemsforjointquality.
Disassemblyispossible,withoutdestructionofthesubstrate.
22.2.2 Disadvantages
Thedisadvantagesofmechanicaljointscanbesummarisedas:
Machiningofholesisnecessary,weakeningthecomponent.
409
ECSSEHB3220Part2A
20March2011
Stressisconcentratedatthebearingsurfaces,therebycreatingstressraiserswhichcaninitiate
failure.
Bonded joints are generally stronger, unless mechanical joints are heavy or used for joining
thicklaminates.
Massisaddedtothejoint,reducingtheoveralljointefficiency.
22.3 Factors affecting design
22.3.1 General
Mechanical fastening of laminated composites is an effective joining technique providing that
considerationisgivento:
Tensileandbendingstressesofcomponent.
Strengthandflexibilityofthefastener.
Lossoftensilestrengthinthecomponentduetothedrilling.
Preloadinthefastener.
Sheardistributioninthejoint.
Frictionbetweenparts.
Residualstresses.
Allowablebearingstresses.
Typesoffasteners.
Fatigue.
Jointselectionisbasedonconsiderationof:
Componentrequiredstrength;bothstaticandfatigue.
Reliability.
Easeoffabrication.
Cost.
Specialjointfunctions,e.g.removable,replaceable.
Allofthesevariablescanbegroupedintothecategories:
Material parameters:
fibretypeandform,i.e.unidirectionalorwovenfabric.
resintype
fibreorientation
laminatestackingsequence
410
ECSSEHB3220Part2A
20March2011
fibrevolumefraction
Fastener parameters,[Seealso:22.4]
fastenertype,i.e.boltorrivet.
fastenersize
clampingforce
washersize
holesizeandtolerance
Design parameters:
jointtype,e.g.singlelap,doublelap.
laminatethicknessandtolerance
geometry,e.g.pitch,edgedistance,holepattern.
loaddirection
loadingrate
staticordynamicload
failurecriteria
22.3.2 Material parameters

Thematerialsconsideredarecarbonandaramidfibrereinforcedepoxies.Themajorityofcomposites
usedinspacestructuresarestackedunidirectionalplylaminates.
Aramid laminates have poorer compressive and interlaminar properties than carbon composites
becauseofthefibrillatedstructureofaramidfibres.Thisreducesthebearingstrengthandsomakes
mechanicalfasteningunattractive.
Little published work is available on mechanicallyfastened ARP aramid composites. The use of
wovencarbonfibrefabriclaminatesisincreasingandtheseareamenabletomechanicalfastening.
As laminate thickness and hence loadbearing demands increase, there is an increasing tendency to
considermechanicalfasteningoveradhesivebonding.
Combined bolted and bonded joints can negate the deficiencies of either one alone, e.g. bolting can
restrain peeling whereas bonding maintains component location, resists creep or dimensional
changes.Thecombinedapproachgivesaredundantloadpath,assumingbothloadpathsaredesigned
towithstandthefullload.
Optimisingstackingsequencescanhaveasignificanteffectonjointperformance,suchas:
Theinclusionof45pliesreducesstressconcentrationfactorsaroundholesinCFRP.
Optimumtensilepropertiesareobtainedwitharatioof0to45pliesof2:1,whilstoptimum
shearstrengthneedsaratioof1:1.
Forglassfibrereinforcedplastics,Ref.[22-5]recommendsputtingthe90fibreonthesurfaceof
thelaminate.
The combination of 0, 45, 90 is an improvement in terms of achieved joint strength to
laminatestrength.
411
ECSSEHB3220Part2A
20March2011
Figure22.3.1summarisestheeffectofplyconfigurationonbearingstresses,Ref.[224].
Figure22.31Boltedjoints:BearingstressesforvariousCFRPplyconfigurations
412
ECSSEHB3220Part2A
20March2011
22.3.3 Fastener parameters

Thechoiceofmechanicalfastenersisbetween:
Boltsarepreferredbecausetheyofferthegreateststrengthofallmechanicallyfastenedjointsin
composites,[See:22.4].
Rivets,wherejointshaveadequatestrengthinhighdutyCFRPlaminateswiththicknessesupto
about 3 mm.The type of rivet, [See:22.6], the use of washers, dimensions ofcountersink and
imposedclampingloadsallaffecttheultimatestrengthofthejoint.Laminateswithfewerthan
four plies cannot be riveted. Little use of riveted joints is envisaged within space structures,
unlessverylightlyloaded,i.e.secondaryortertiarystructures.
Screws,whichcauseextensivedamagetothelaminate.
22.3.4 Design parameters
22.3.4.1 Joint types

The choice of joint type is dependent on the particular application. Some examples are shown in
Figure22.3.2,Ref.[221].
Figure22.32Mechanicallyfastenedjoints:Examplesofconfigurations
When using experimental data, a close correlation between idealised test joints and the actual
configuration being designed is necessary. Likewise, all loading modes and eccentricities should be
considered.Thisusuallynecessitatesconsiderabletestingofboltedspecimenstoproducedesigndata.
413
ECSSEHB3220Part2A
20March2011
22.3.4.2 End-distance effects
A minimum value of e/d ratio is needed for full bearing strength, [See: Figure 22.4.1]. The ratio is
alwaysgreaterthan3forpseudoisotropic[0,45,90,45]layups.For45layups,itisgreater
than5(thisdiffersfromtheenddistanceratioof2.0recommendedbyECSSEHB3223formetallic
materials).
22.3.4.3 Width effects

From work on single hole specimens, it has been found that [0, 45] lay ups in CFRP need a
minimumw/dof4(w/dof8forall45),whichcanprobablybeequatedtominimumpitchdistances
formultiholejoints,[See:Figure22.4.1].In[0,45]layupstheminimume/dandw/ddependonthe
relativeproportionsof0and45plys.
22.3.4.4 Row and pitch distance

Thefactorstobeconsideredare:
Holepatterns:Atandemrowispreferabletoaparallelrow,[See:Figure 22.3.3,Ref.[22-1]],and
theuseofstaggeredpitchwithtandemrowshelps.
Generally, a pitch of 5d (5 times the fastener diameter) is acceptable for joints of large safety
factors,while4disaminimumvalue.
Multiple rows of fasteners: The pitch on outer rows is greater than that on inner rows, since
largetensileloadingcanthusbetoleratedinthejoint.
Whereeccentricityofthejointisconsidered,e.g.anunsymmetricaltypefasteningproducesbending
stressinthejoint,thenmultiplerowsoffastenerscanminimise:
damageduetobending,but
axialloadcarryingcapacity.
Insuchacase,asinglerowoffastenersdisplacedeverysecondfastenercanbeanimprovement.
The recommended distance (back pitch) between adjacent rows is no less than 4d. High back pitch
providesthedisplacementneededtominimisestressescausedbyjointeccentricity.
414
ECSSEHB3220Part2A
20March2011
Key:
a. Rowofbolts:Rows1and2arestaggered;Rows2and3areuniformrectangularpattern.
b. Twoboltsintandem(lineofbolts).
c. Twoboltsinparallel(rowofbolts).
Figure22.33Mechanicallyfastenedjoints:Definitionofjointgeometry
415
ECSSEHB3220Part2A
20March2011
22.3.4.5 Diameter thickness effect
Extremelythinlaminates(t0.76mm)needtobereinforcedtoavoidd/tratiogreaterthan4.
ForCFRP,ad/tratio3isgenerallysatisfactoryformostthicknesses.
22.3.4.6 Failure criteria and modes

The definition of failure varies widely which, in turn, confuses the value of design allowable
strengths. A partially accepted failure criterion is based on the deformation of the bolt hole under
load. However, agreement on the degree of permanent hole deformation is lacking with values
varying from 0.5% to 4% of the hole diameter being suggested. A safety factor of 2 on the ultimate
strengthappearstobeaminimumandisclosetothestressatwhichdamageinitiates.
Mechanicallyfastenedjointscandisplayanumberoffailuremodes,asshowninFigure22.3.4.Failure
is normally gradual, with progressive damage eventually resulting in complete loss of loadbearing
capacity.
Thedefinitionofjointstrengthcanbecomeconfused,thefailurecriteriacanbe:
Theonsetofdamage,
Degreesofdamageor
Ultimately,themaximumloadbearingcapacity.
416
ECSSEHB3220Part2A
20March2011
Figure22.34Boltedjoints:Failuremodes
Forjointscombiningcompositeandmetal,eachmaterialhasitsownfailurecharacteristics,e.g.
Compositeshavealowstrain,brittlemodeoffailure.
Metals show greater ductility and plastic deformation than composites; depending on their
compositionandcondition.
For bolted joints, full bearing strength is only developed if failure due to tension and shear, [See:
Figure22.3.4],issuppressedbyprovidingadequateenddistanceandwidth(orpitch),andifsufficient
restraintispresenttopreventlocalinterlaminarfailure.
Simple bolted joints rarely achieve more than 50% of base laminate strength. Higher joint strengths
canbeobtainedby:
localisedreinforcement,and
modifiedloaddistribution.
Thesetechniquesreducejointefficiencyintermsofmassandincreaseproductioncosts.
417
ECSSEHB3220Part2A
20March2011
Figure22.3.5showsexamplesoflocallyreinforcedjoints.
Figure22.35Localreinforcementsforboltedjoints
22.4 Bolted joints
22.4.1 Material parameters

[See:22.3]
22.4.2 Fastener parameters

Boltsofferthegreateststrengthformechanicallyfastenedjointsincomposites.Unlessovertightened,
nodamageisdonetothecompositeduringassembly.
Machining of holes in composites needs care to achieve a repeatable high quality. If machining
damagesthecompositearoundtheholethenjointperformanceanddurabilitycanbeimpaired.
Intermsofoptimisingstrength,thefactorstoconsiderare:
Theboltisagoodfitintothe(preferably)reamedhole.
Increasingthebolttighteningtorque,andhenceclampingforce,increasesthebearingstrength.
Optimumlevelsofclampingexist.
The hole in the washer (ideally, reamed) is a close fit to the bolt. Washer size can influence
strength.
418
ECSSEHB3220Part2A
20March2011
Bewareofgalvanic corrosion.
AproblemidentifiedwithmechanicallyfastenedCFRPforaircraftconstructionsisgalvaniccorrosion,
[See: 20.3]. For CFRP and aluminium, it can be particularly acute because their electrochemical
potentialsareverydifferent.
Titaniumfastenersarebetter,thoughexpensive,forCFRP.Whilstcorrosionisnotnormallyaproblem
for deployed satellites, reusable vehicles and launchers have to operate in Earths environment,
exposingstructurestogalvaniccorrosionproblems.
Whilstbolttighteningcanbeusedtoincreasejointbearingstrengths,itisnotadvisabletorelyonthis
increment of strength when designing constructions. Some relaxation of clamped composites occurs
overaperiodoftimeandthestrengthisthenlost.Theadditionalstrengthduetoclampingisignored
duringdesign,itspresenceisanunaccountedsafetymarginbenefit.
Figure22.4.1showsthebearingstrengthofboltedcarbonfibrereinforcedlaminates,Ref.[222].
Figure22.41Bearingstrengthofboltedcarbonfibrereinforcedlaminates
There are many different bolts applicable to composites, included in these are countersunk variants
whichneedadditionalcompositemachiningwhichinturnaffectsstrengthandfailuremodes.Ashort
review of fasteners for composites was published based experiences within the American aircraft
manufacturers;andidentifiesseveralsourcesoffasteners.
[See:22.3forcommentsonothertypesofmechanicalfasteners]
[See:ECSSQST7046forprocurementofthreadedfasteners]
419
ECSSEHB3220Part2A
20March2011
22.4.3 Design parameters

[See:22.3Designparameters]
[See also: ECSSEHB3223 for guidelines on the design of bolted joints; ECSSQST7046 for
procurementofthreadedfasteners]
22.5 Bolted joint: Analysis
22.5.1 General
JointsinCFRPelementsarenormallydesignedinmuchthesamewayasthoseinmetallicstructures
usingmechanicalfasteners.Knowledgeofthestressdistributionandthefailurestrengthishelpfulin
selectingappropriatejointsizes.
An overview of certain problems and a guideline for the preliminary design of bolted joints is
presented.
22.5.2 Analytical methods
22.5.2.1 Two-dimensional analysis

Analyticalmethodshavebeendevelopedforcalculatingthestrengthofboltedjointsonthebasisof
twodimensional analysis methods, Ref. [226], [227], [228], [229], but they ignore several effects,
including:
freeedgeeffectsatopenholes
clampingeffectswhenjointsarebolted
influenceofdelaminationonloaddistribution(plastification).
22.5.2.2 Three-dimensional analysis

Highly sophisticated FE computer codes, Ref. [2210], have been used to perform threedimensional
analyseswhichareingoodaccordancewithexperimentalresults.
22.5.3 Stress distribution
22.5.3.1 Infinite isotropic plate containing a circular unloaded hole

Ananalyticalsolutionforthestressdistributioninaninfiniteisotropicplate,containinganunloaded
hole,fromRef.[2211],andcalculatedwiththeaidofacomputerprogram,Ref.[227].
ThestressdistributionsareshowninFigure22.5.1.
420
ECSSEHB3220Part2A
20March2011
ComparisonofSpringer,Ref.[227]withthetheoreticalresultsgivenbyTimoshenko,Ref.[2211]:
Figure22.51Stress2alongthex1axisinanisotropicinfiniteplatecontaining
anunloadedcircularhole
22.5.3.2 Infinite orthotropic plate containing a circular unloaded hole

For comparison, Figure 22.5.2 gives the stress distribution in an orthotropic plate [0/90]. This
diagramisbasedonthesolutiongiveninRef.[227],[228].
421
ECSSEHB3220Part2A
20March2011
ComparisonofSpringer,Ref.[227]withthetheoreticalresultsobtainedby
NuismerWhitney,Ref.[228].
Figure22.52Stress2alongthex1axisinanorthotropicinfiniteplate[0/90]
containinganunloadedcircularhole
22.5.3.3 Finite width isotropic plate containing a circular loaded hole

Figure 22.5.3 presents the stress distribution in an isotropic plate containing a loaded hole. This
diagramisbasedonthesolutionmethodspresentedinRef.[227],[229].
422
ECSSEHB3220Part2A
20March2011
Figure22.53Stress2alongthex1axisinanisotropicplateoffinitewidth
containingaloadedhole
22.5.4 Failure prediction
22.5.4.1 General
Numerous failure criteria have been developed for laminated composites. The ESAcomp software
documentationexplainsmanyofthecriteriathatarecommonlyusedintheEuropeanspaceindustry,
Ref.[2218].Thesecriteriaaregenerallyexpressedintheformoffarfieldstrainorstresslevelsthat
cannotbeexceeded(i.e.thestrengthofthelaminatewithoutanyloadconcentrations).Whenapplying
suchcriteriatoboltedjoints,itisrecommendedthatlocaleffectsaroundtheloadedholeareaccounted
forbyusingtheWhitney/Nuismerfailurehypothesis.Thishypothesisisusedinconjunctionwithone
ofthefarfieldfailurecriterionsuchastheYamadacriterion.
22.5.4.2 Yamada failure criterion

Thiscriterionisbasedontheassumptionthatfailureoccurswheneveryplyhasfailedasaresultof
cracksalongthefibres,Ref.[2212].
Conditionswhicharemetinanyoneofthepliesare:
2 2
1
12
e 1 : no failure
e2 [22.5-1]
R tu R su e 1 : failure
1 12
where:
1and12arethelongitudinalandshearstressesinaply(xandybeingthecoordinatesparalleland
normaltothefibresintheply),
423
ECSSEHB3220Part2A
20March2011
R1tu isthelongitudinaltensilestrengthoftheply,and
R12su istheshearstrengthofasymmetric,crossplylaminatethathasthesamenumberofpliesasthe
laminateunderconsideration.
22.5.4.3 Whitney/Nuismer failure hypothesis

Thishypothesis,Ref.[2213],isusedwithanappropriatefailurecriterion,e.g.Yamada,andproposes
that failure occurs when in any one of the plies of a laminate stresses satisfy this criterion on a
characteristiccurvearoundtheloadedhole.ThischaracteristiccurveisshowninFigure22.5.4andcan
becalculatedby:
r D 2 R R
c ot oc
Rot cos [22.5-2]
where:

[22.5-3]
2 2
RotandRocarecharacteristiclengthsfortensionandcompression.Theseparameterscanbedetermined
experimentallybymeasuringthetensileandcompressivestrengthsofnotchedlaminates.
Rot and Roc depend only on the material. Therefore, the coordinates of the characteristic curve also
dependonlyonthematerialandareindependentofthegeometryandthestressdistribution.
Figure22.54Whitney/Nuismerfailurehypothesis:Characteristiccurve
424
ECSSEHB3220Part2A
20March2011
WhenthishypothesisisusedtogetherwiththeYamadacriterion,failureoccurswhen:
e1atr=rc
atanypointonthecharacteristiccurve.
ThishypothesisisgeneralandnotrestrictedtoYamadascriterion.Anyothercriterioncanbeused.
22.5.5 Experimental data
22.5.5.1 Shear load carrying capabilities of bolts in CFRP facings

Thesamplesweretestedstatically(tensionandcompression)uptofractureloadusingthetestdevice
showninFigure22.5.5,Ref.[2214].
Thetestswereperformedwith:
differentfibreorientationsofthefacing
differentmaterials(HTandHMCFRP).
laminatethickness0.25mmand0.375mm.
ThetestresultsareshowninTable22.5.1for914CT300compositeandinTable22.5.1for914CHM
composite,Ref.[2214].
Figure22.5.6comparesthebearingstressesatthefailurepoint,Ref.[2214].
425
ECSSEHB3220Part2A
20March2011
Figure22.55Boltedjoints:Testdeviceandconfigurationforinplanefacesheet
loading
426
ECSSEHB3220Part2A
20March2011
Table 22.5-1 - Bolted joints: Results of static test on material 914C T300

Lay-up 0/90 90/0 45/-45 45/0/-45 22.5/-22.5/-67.5 60/60/-60 30/-30/-90
Tf 0.25 0.25 0.25 0.375 0.375 0.375 0.375
Sample No. Load [N] Load [N] Load [N] Load [N] Load [N] Load [N] Load [N]
Tensile Comp. Tensile Comp. Tensile Comp. Tensile Comp. Tensile Comp. Tensile Comp. Tensile Comp.
1 2639 2656 2383 2355 2360 2225 4444 4409 4500 4578 4212 4005 3554 3693
2 2944 2717 (2625) 2236 2766 2253 4534 4260 5072 4401 3273 3523 4180 4401
3 2465 3067 2442 2188 2830 2475 4620 3994 4991 4645 (4887) 5042 3499 3787
4 2621 2289 2166 2197 (1779) 1686 4450 3731 4273 4036 (4410) 4252 3258 3921
5 2919 2905 2140 1981 2940 2764 4901 4621 4600 4555 3738 3996 3499 3783
6 2326 2736 2135 2039 2564 2251 4774 3868 5007 4386 4275 4107 3856 4257
7 2660 2952 - - 2474 2318 4851 4261 3276 3406 3854 3437 3380 3159
p 2653 2760 2253 2166 2660 2381 4653 4163 4531 4315 3870 3813 3604 3858
S.D. 223 254 147 136 232 208 190 313 630 373 404 309 313 405
-2
p/(td) Nmm 758 789 643 618 760 680 886 793 863 822 737 726 686 735
Failure mode buckling buckling buckling buckling bearing + bearing + bearing bearing bearing bearing bearing bearing bearing bearing
under under under under 45 shear 3x45 1x
0 0 0 0 buckling buckling
Key: : Very high load speed : Without axis clamping : Shimming around hole Tf : Facing Thickness
427
ECSSEHB3220Part2A
20March2011
Table22.52Boltedjoints:Resultsofstatictestonmaterial914CHM
Lay-up 0/90 45/-45
Facing Thickness 0.25 0.25
(Tf)
Load [N] Load [N]
Sample No.
Tensile Comp. Tensile Comp.
1 2144 2483 1950 2201
2 (1630)1 2074 2104 1958
3 (1655)1 2022 1860 2165
4 1980 2194 1964 1963
5 1998 2093 1863 2113
6 2061 2249 1962 2355
7 1544 2152 1812 1727
p 1825 2128 1931 2069
S.D. 299 169 97 206
-2
p/(t d), Nmm 521 608 552 591
shear out 1 x
tensile + buckling under tensile + shear
Failure mode buckl. under
shear out 0 out under 45
45
Key: 1) Sample was influenced by the failure of neighbouring samples.
Figure22.56Boltedjoints:BearingstressatfailurepointofCFRPfacingswith
differentlayups
428
ECSSEHB3220Part2A
20March2011
22.6 Riveted joints
22.6.1 General
Intheinitialstructuralapplicationofcompositematerials,fastenersandinstallationproceduresthat
had been optimised for metallic structures were applied. For these applications interference fit and
high preload give strength and durability improvements and the use of rivet guns to install rivets
leadstocostsavings.Experiencehasshownthatusingthesamefastenersandinstallationprocedures
incompositestructurescanproduceunacceptabledamage.
The nature of the resulting damage, the fastener implications and types of fasteners are described
here.
22.6.2 Installation damage of riveted joints
22.6.2.1 Interference
The low interlaminar strength of composite materials restricts the use of interference fit fasteners.
When permanent fasteners are forced into an interferencefit hole, the plies on the back side can
delaminateandsplitout.
Rivets are usually installed such that they completely fill the hole by expansion of the shank.
Experience with fibreglass has shown that such expansion causes delaminations and ply buckling.
Thesedifficultieshaveleadtothegeneralrequirementthatonlyclearancefit(0.000to+1.02104m)
fastenersareappropriatetocompositestructures,Ref.[2215].Thisisnottoimplythatinterferencefit
isnotdesirableorthatsuitablefastenerconceptsforinterferencearenotavailable,e.g.increasingthe
level of interference in a carbon/epoxy joint can increase the fatigue life, Ref. [2215]. Interference is
alsodesirableto:
increaseloadsharinginjointswithmultiplerowsoffasteners,
retardfuelleakageintankareas,and
providethereactivetorqueneededforonesidedinstallationsofHiLokfasteners.
22.6.2.2 Impact forces

Thepropensityofcompositestructurestodelaminateandsplitoutonthebacksidefromtheimpact
forces created by debris or tools is well known. Consequently, the use of rivet guns on composite
structuresisprohibited.
Thepulltypefasteners,likethelockboltsystems,alsodeliverimpactforcesintheformofthereaction
forcesreleasedwhenthebreakneckissevered.However,nodamagetothelaminate,attributableto
thistypeofimpact,hasbeenobserved.
22.6.2.3 Preload
Preloadistheforceexertedonthesheetsbeingjoinedbythefollowing:
Nuttorquing,
Collarswaging
Tailformation.
429
ECSSEHB3220Part2A
20March2011
With the low compressive strength of composites, a high preload results in laminate crushing. The
aerospaceindustrymainlymakesuseofbigfootfastenerstoovercomethisproblem,[See:Footprint].
There is, however, room for some optimisation of the fastener design in the future, e.g. higher
preloadscanprovidelongerfatiguelives.
22.6.3 Pull through strength
22.6.3.1 General
The low pull through strength of mechanically fastened composite joints is well known. The
prevalenceoffastenerswithenlargedfootprintsandenlargedheadsinsuchstructuresisevidenceof
thisfact.Modifiedfastenershavebeendevelopedtoimprovepullthroughcharacteristics.
22.6.3.2 Footprint
The footprint is the bearing area of the nut, collar or tail of an installed fastener. The pull through
strengthofafastenerispartiallydeterminedbythesizeofthefootprintasshowninFigure22.6.1.
Monogram developed the Bigfoot fastener, which has the same tail diameter and pull through
strength as the Hi Loks used on the opposite surface. The collar for the titanium GPL fasteners
(GrooveProportionalLockbolt)alsohasanenlargedfootprinttoavoidpullthroughproblems.
Figure22.61Rivets:Improvementinpullthroughstrengthbylargerflushhead
diameterandfootprint
430
ECSSEHB3220Part2A
20March2011
22.6.3.3 Flush head configuration
Theflushheadconfigurationsconsidered,[See:Figure22.6.1],arewherethepullthroughstrengthis
showntobeafunctionoftheheadoutsidediameter.Factorstobeconsideredare:
Thickcomposites:Theuseoftensionheadfasteners(100)forthickcompositestructuresisnow
quitecommon.
Thincomposites:Fastenerswith130shearheadreportedlysupports30%moreloadthanthe
100shearhead.The120shearheadcanalsobeapplicable.
22.6.4 Fasteners used in composite structures

AsummaryofthosefastenersusedforaerospacematerialsisshowninTable22.6.1,Ref.[2215].
Table22.61Typesoffastenersusedtojoinaerospacematerials
Load Transfer Materials Joined Head Type Fastener Type
Low C/E C/E Flush Threaded
C/E C/E Flush Blind
C/E C/E Protruding Threaded
Low C/E metal Flush Blind
C/E metal Flush Threaded
C/E metal Flush Rivet
C/E metal Protruding Threaded
High C/E metal Protruding Threaded
C/E metal Flush Threaded
C/E metal Flush Blind
C/E metal Flush Swaged
Key:C/ECarbonfibre/epoxycomposite.
431
ECSSEHB3220Part2A
20March2011
22.6.4.1 Types of rivets
VarioustypesofrivetsproposedforjoiningaerospacematerialsareshowninFigure22.6.2.
Figure 22.6-2 - Types of rivets for aerospace materials
432
ECSSEHB3220Part2A
20March2011
22.7 References
22.7.1 General
[221] FulmerResearchLaboratories
'Guidelines on basic aspects of jointing technology for advanced fibre
reinforced plastics used in space structures'
ESA Contract 5487/83/NL/PB (SC)
[222] T.A.Collings
'The strength of bolted joints in multi directional CFRP laminates'
Composites, January 1977
[223] G.Lubin
'Handbook of Composites'
Von Nostrand Reinhold Company, 1982
[224] J.Bauer
'Gegenberstellung von Vorhersagewerten und Versuchsergebnissen an
Kerben und bolzenbelasteten Bohrungen in CFK Laminaten'
DGLR Symposium, 1984
[225] E.W.Godwin&F.L.Matthews
'A review of strength of joints in fibre reinforced plastics'
Composites, July 1980
[226] C.M.S.Wong&F.L.Matthews
'A finite element analysis of single and two-hole bolted joints in fibre
reinforced plastic'
[227] G.S.Springeretal
'Strength of mechanically fastened composite joints'
Journal of Composite Materials, Vol. 16, 1982
[228] R.J.Nuismer&J.M.Whitney
'Uniaxial failure of composite laminate containing stress concentrations'
ASTM STP 593
[229] Th.DeJong
'Stresses around pin loaded holes in composite materials' Mechanics of
Composite Materials: Recent Advances
[2210] ASKAComputerCode
[2211] S.Timoshenko&S.WaimowskyKrieger
'Theory of Plates and Shells'
McGraw-Hill, New York 1940
[2212] S.E.Yamada
'Analysis of laminate strength and its distribution'
'Stress fracture criteria for laminated composite containing stress
concentrations'
433
ECSSEHB3220Part2A
20March2011
[2214] MBB/ERNORaumfahrttechnik
'Inserts in nonmetallic sandwich components'
ESA Contract 440/80/NL/AK (SC)
[2215] R.T.Coleetal
'Fastener for composite structures'
Composites, July 1982
[2216] G.P.Sendeckyj&M.D.Richardson
'Fatigue behaviour of a graphite epoxy laminate loaded through an
interference fit pin'
Proceedings of 2nd Air Force Conference on Fibrous Composites in Flight
Vehicle Design, AFFDL74 103, September 1974
[2217] J.M.Hodgkinson,D.L.deBeer&F.L.Matthews
'The strength of bolted joints in Kevlar RP'
Proc Workshop 'Composite Design for Space Applications'
ESTEC Noordwijk, Oct 1985 (ESA SP-243, Feb 1986) p53-61
[2218] M.Palater:ComponeeringInc.(Finland)
'Theoretical Background of ESAComp Analyses'
Part of ESAComp software documentation, v. 1.5 (April 1999)
22.7.2 ECSS standards

[See: ECSS website]
ECSSEHB3223 Threadedfastenershandbook;previouslyESA
PSS03208.
ECSSQST7046 Requirementsformanufacturingand
procurementofthreadedfasteners;previously
ESAPSS01746
434

ECSS E HB 32 20 - Part2A PDF

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

ECSS E HB 32 20 - Part2A PDF

Uploaded by

Copyright:

Available Formats

ECSS-E-HB-32-20 Part 2A

Change log .................................................................................................................3

10 Stress-strain relationships ................................................................................28

11 Strength prediction and failure criteria.............................................................61

12 Calculation of thermal stress and displacement ...........................................104

13 Moisture effects on composite properties .....................................................114

14 Stress concentrations and fracture ................................................................136

15 Prediction of dynamic characteristics ............................................................189

16 Computer analysis of composites ..................................................................196

17 Composite adequate design............................................................................246

18 Curing stresses: Effects and prediction.........................................................266

19 Manufacturing faults and service damage .....................................................274

20 Environmental aspects of design ...................................................................292

22 Mechanically fastened joints ...........................................................................409

circular hole, 2 L W 10 .......................................................................... 162

Table 14.9-3 - Material data ................................................................................................. 163

10.2 Elastic property prediction for UD ply from

Prediction of Ply Properties

10.3 Analytical notation for elastic constant methods

10.4 Calculation methods for elastic constants of UD ply

10.5 Longitudinal modulus

Figure 10.5-1 - Longitudinalmodulus(E1)ofglass/epoxy

Figure 10.5-2 - Longitudinalmodulus(E1)ofcarbon(HT)/epoxy

10.6 Longitudinal Poisson's ratio

10.7 Transverse modulus

10.7.2 Jones method

10.7.3 Frster/Knappe method

10.7.4 Schneider method

10.7.5 Puck method

10.7.6 Tsai method

10.7.7 HSB method

Figure 10.7-2 - Transverse modulus (E2) of carbon (HT)/epoxy composites

10.8 Transverse Poisson's ratio

10.9 Transverse shear modulus

10.9.2 Jones method

10.9.3 Frster/Knappe method

10.9.4 Schneider method

10.9.5 Puck method

10.9.6 Tsai method

10.9.7 HSB method

Figure 10.9-1 - Shear modulus (G12) of glass/epoxy composites

Figure 10.9-3 - Shear modulus (G12) of carbon (HM)/epoxy composites

Figure 10.9-4 - Shear modulus (G12) of aramid/epoxy composites

10.10 In-plane stress calculation methods

10.11 Analytical notation for in-plane methods

10.12 Stress-strain relations for unidirectional plies

10.12.1 Fibre-oriented co-ordinate system

10.13 On axis stress strain relations

10.13.2 Compliance matrix

10.13.3 Modulus matrix

10.13.4 Symmetry of compliance and modulus matrices

10.14 Stress-strain relations for a ply of arbitrary

With: m cos ; n sin

10.14.2 Off axis stiffness of a unidirectional ply

Qxx... Transformation Q11...

Figure 10.14-2 - Transformation of stiffness and compliance matrix

10.15 Stiffness matrix for a laminate

10.15.1 General laminates

10.15.2 Symmetric laminates

10.15.3 Flow chart

10.16 Calculation methods with interlaminar stresses and

10.16.1 Calculation with free-edge stresses

Figure 10.16-1 - Uniaxial loading of a composite