Fatigue

Atlas of
Fatigue Curves
Edited by
Howard E. Boyer
Senior Technical Editor
American Society for Metals
The Materials
Information Society
ASM lnternatlonal"
Materials Park, Ohio 44073-0002
www.asminternational.org
Preface
This Atlas was developed to serve engineers who are looking for fatigue
data on a particular metal or alloy. In the past, the first step to locating this
data was an expensive and time-consuming search through the technical literature. Now, many ofthe important and frequently referenced curves are presented together in this one volume. They are arranged by standard alloy designationsand are accompanied by a textual explanation offatigue testing and
interpretation of test results. In each case, the individual curve is thoroughly
referenced to the original source.
Having these important curves compiled in a single book will also facilitate the computerization of these data. Plans are currently under way also to
make the data presented in this book available in ASCII files for analysis by
computer programs.
The Atlas of Fatigue Curves is obviously not complete, in that many
more curves could be included. Persons wishing to contribute curves to this
compilation for inclusion in future revisions should contact the Editors,
Technical Books, American Society for Metals, Metals Park, Ohio 44073.
Contents
Fatigue Testing
Introduction
I
Fatigue Crack Initiation 4
Fatigue Crack Propagation
12
SECTION 1: S-N Curves That Typify Effects of Major Variables

I-I.
1-2.
1-3.
1-4.
1-5.
1-6.
1-7.
1-8.
1-9.
1-10.
I-II.
1-12.
1-13.
1-14.
1-15.
27
S-NCurves Typical for Steel 27

S-NCurves Typical for Medium-Strength Steels 28
S-NDiagrams Comparing Endurance Limit for Seven Alloys
30
Steel: Effect of Microstructure 31
Steel: Influence of Derating Factors on Fatigue Characteristics
32
Steel: Correction Factors for Various Surface Conditions
33
Fatigue Behavior: Ferrous vs Nonferrous Metals
34
Comparison of Fatigue Characteristics: Mild Steel vs Aluminum Alloy
35
Carbon Steel: Effect of Lead as an Additive
36
Corrosion Fatigue: General Effect on Behavior
37
Effect of Corrosion on Fatigue Characteristics of Several Steels
38
Steel: Effect of Hydrogen on Fatigue Crack Propagation 39
Relationship of Stress Amplitude and Cycles to Failure 40
Strain-Life and Stress-Life Curves 41
Fatigue Plot for Steel: Ultrasonic Attenuation vs Number of Cycles 42
SECTION 2: Low-Carbon Steels: Flat-Rolled, Weldments and Tubes

2-1.
2-2.
2-3.
2-4.
2-5.
2-6.
2-7.
2-8.
2-9.
2-10.
2-11.
2-12.
2-13.
2-14.
2-15.
2-16.
2-17.
2-18.
2-19.
2-20.
2-21.
2-22.
2-23.
2-24.
2-25.
43
Typical S-N Curve for Low-Carbon Steel Under Axial Tension 43

AISI 1006: Effects of Biaxial Stretching and Cold Rolling
44
AISI 1006: Weldment; FCAW, TIG Dressed
45
AISI 1006: Weldment; Shear Joints 46
AISI 1006: Weldment; Lap-Shear Joints 47
AISI 1015: Effect of Cold Working 48
A533 Steel Plate: Fatigue Crack Growth Rate
49
A514F Steel Plate: Fatigue Crack Growth Rates
50
A514F and A633C: Variation in Fatigue Crack Growth Rate With Orientation 51
A514F: Scatterbands of Fatigue Crack Growth Rate
52
A633C Steel Plate: Scatterbands of Fatigue Crack Growth Rates
53
Low-Carbon Steel Weldment: Effects of Various Weld Defects
54
Low-Carbon Steel Weldment: Effect of Weld Reinforcement and Lack of
Inclusions
55
Low-Carbon Steel Weldment: Effect of Weld Reinforcement and Lack of
Penetration 56
Low-Carbon Steel Weldment: Computed Fatigue Strength; Weldment Contained
Lack of Fusion
57
Low-Carbon Steel Weldment: Effect of Reinforcement and Undercutting
58
Low-Carbon Steel: Transverse Butt Welds; Effect of Reinforcement
59
A36/E60S-3 Steel Plate: Butt Welds 60
A514F/EllO Steel: Bead on Plate Weldment
61
A36 and A514 Steel Plates: Butt Welded
62
A36 Plate Steel: Butt Welded
63
Low-Carbon Steel Tubes: Effect of Welding Technique
64
Low-Carbon Steel: Effect of Applied Anodic Currents in 3% NaCI
65
Low-Carbon Steel: Effect of pH in NaCI and NaOH
66
Low-Carbon Steel: Effect of Carburization and Decarburization 67
Contents
VI
2-26. A514B Steel: Effect of Various Gaseous Environments on Fatigue Crack

Propagation 68
2-27. Cast 1522 and 1541 Steels: Effect of Various Surface Conditions
69
2-28. Cast A216 (Grade WCC) Steel: Fatigue Crack Growth Rate
70
SECTION 3: Medium-Carbon Steels, Wrought and Cast

3-1.
3-2.
3-3.
3-4.
3-5.
3-6.
3-7.
3-8.
3-9.
3-10.
3-11.
3-12.
AISI 1030 (Cast) Compared With AISI 1020 (Wrought)

71
AISI 1035: Effect of Gas and Salt Bath Nitriding
72
AISI 1040: Cast vs Wrought
73
AISI 1045: Relationship of Hardness and Strain-Life Behavior 74
AISI 1141: Effect of Gas Nitriding 75
Medium-Carbon Steels: Interrelationship of Hardness, Strain Life and Fatigue
Life 76
Medium-Carbon Steel: Effect of Fillet Radii
77
Medium-Carbon Steel: Effect of Keyway Design
78
Medium-Carbon Steel: Effect of Residual Stresses
79
Medium-Carbon Cast Steel: Effect of Changes in Residual Stress
80
Medium-Carbon Cast Steel: S-NProjection (Effect of Applied Stress)
81
Medium-Carbon Cast Steel: Effect of Applied Stress (Shot Blasting)
82
SECTION 4: Alloy Steels: Low- to High-Carbon, Inclusive

4-1.
4-2.
4-3.
4-4.
4-5.
4-6.
4-7.
4-8.
4-9.
4-10.
4-11.
4-12.
4-13.
4-14.
4-15.
4-16.
4-17.
4-18.
4-19.
4-20.
4-21.
4-22.
4-23.
4-24.
4-25.
4-26.
4-27.
4-28.
4-29.
4-30.
4-31.
4-32.
4-33.
4-34.
4-35.
4-36.
4-37.
71
83
Medium-Carbon Alloy Steels, Five Grades: Effect of Martensite Content 83

Medium-Carbon Alloy Steels, Six Grades: Hardness vs Endurance Limit
84
Medium-Carbon Alloy Steels: Effect of Specimen Orientation
85
4027 Steel: Carburized vs Uncarburized
86
4120 Steel: Effect of Surface Treatment in Hydrogen Environment
87
4120 Steel: Effect of Surface Treatment in Hydrogen Environment
88
4120 Steel: Effect of Various Surface Treatments on Fatigue Characteristics in Air
vs Hydrogen
89
4130 Steel: Fatigue Crack Growth Rate vs Temperature in Hydrogen
90
4135 and 4140 Steels: Cast vs Wrought
91
4135 and 4140 Steels: Cast vs Wrought
92
4140,4053 and 4063 Steels: Effect of Carbon Content and Hardness
93
4140 Steel: Effect of Direction on Fatigue Crack Propagation 94
4140 Steel: Effect of Cathodic Polarization 95
Cast 4330 Steel: Effects of Various Surface Conditions 96
4340 Steel: Scatter of Fatigue Limit Data
97
4340 Steel: Strength vs Fatigue Life 98
4340 Steel: Total Strain vs Fatigue Life 99
4340 Steel: Stress Amplitude vs Number of Reversals
100
4340 Steel: Effect of Periodic Overstrain
101
4340 Steel: Estimation of Constant Life
102
4340 Steel: Effect of Strength Level on Constant-Life Behavior
103
4340 Steel: Notched vs Unnotched Specimens
104
4340 Steel: Effect of Decarburization
105
4340H Steel: Effect of Inclusion Size
106
4340 Steel: Influence of Inclusion Size
107
4340 Steel: Effect of Hydrogenation; Static Fatigue
108
4340 Steel: Effect of Hydrogen
109
4340 Steel: Effect of Nitriding
110
4340 Steel: Effect of Nitriding and Shot Peening
III
4340 Steel: Effect of Induction Hardening and Nitriding
112
4340 Steel: Effect of Surface Coatings
113
4340 Steel: Effect of Temperature on Constant-Lifetime Behavior
114
4520H Steel: Effect of Type of Quench
115
4520H Steel: Effect of Shot Peening
116
117
4620 Steel: P/M-Forged
118
4620 Steel: P/M-Forged at Different Levels
119
Contents
4-38.
4-39.
4-40.
4-41.
4-42.
4-43.
4-44.
4-45.
4-46.
4-47.
4-48.
4-49.
4-50.
4-51.
4-52.
4-53.
4-54.
4-55.
4-56.
4-57.
4-58.
4-59.
4-60.
4-61.
4-62.
4-63.
4-64.
4625 Steel: P/M vs Ingot Forms

120
4640 Steel: P/M-Forged
121
High-Carbon Steel (Eutectoid Carbon): Pearlite vs Spheroidite
122
52100 EF Steel: Surface Fatigue; Effect of Finish and Additives
123
124
52100 EF Steel: Surface Fatigue; Effect of Surface Finish and Speed
52100 EF Steel: Surface Fatigue; Effect of Lubricant Additives
125
52100 EF Steel: Surface Fatigue; Effect of Lubricant Viscosity, Slip Ratio and
Speed
126
52100 EF Steel: Rolling Ball Fatigue; Effect of Oil Additives
127
52100 Steel: Carburized vs Uncarburized
128
8620H Steel: Carburized; Results From Case and Core
129
8620H Steel: Effect of Variation in Carburizing Treatments
130
131
8622 Steel: Effect of Grinding
132
Cast 8630 Steel: Goodman Diagram for Bending Fatigue
133
Cast 8630 Steel: Effect of Shrinkage
134
Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue
135
Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue
136
Cast 8630 Steel: Effect of Shrinkage on Plate Bending
137
Cast 8630 vs Wrought 8640
138
8630 and 8640 Steels: Effect of Notches on Cast and Wrought Specimens
139
Nitralloy 135 Steel: Effect of Nitriding
140
AMS 6475: Effects of Welding
141
Medium-Carbon, ICr-Mo-V Steel Forging: Effect of Cycling Frequency
142
EM 12 Steel: Effect of Temperature on Low-Cycle Fatigue
143
Cast 0.5Cr-Mo-V Steel: Effects of Dwell Time in Elevated-Temperature
Testing
144
Cast 0.5Cr-Mo-V Steel: Effect of Environment at 550C (1022 OF) 145
Cast C-0.5Mo Steel: Effect of Temperature and Dwell Period on Cyclic Endurance
at Various Strain Amplitudes
146
SECTION 5: HSLA Steels

5-1.
5-2.
5-3.
5-4.
5-5.
5-6.
5-7.
5-8.
5-9.
5-10.
5-11.
5-12.
5-13.
5-14.
5-15.
5-16.
5-17.
5-18.
147
HI-FORM 50 Steel vs 1006 147

HI-FORM 50 Steel vs 1006: Stress Response
148
HI-FORM 50 Steel Compared With 1006, DPI and DP2
149
HSLA vs Mild Steel: Torsional Fatigue
150
Proprietary HSLA Steel vs ASTM A440
151
Comparison of HSLA Steel Grades BE, JF and KF for Plastic Strain Amplitude
vs Reversals to Failure
152
Comparison of HSLA Steel Grades BE, JF and KF for Total Strain Amplitude
vs Reversals to Failure
153
Comparison of a Dual-Phase HSLA Steel Grade With HI-FORM 50: Total Strain
Amplitude vs Reversals to Failure
154
AISI 50 XF Steel: Effects of Cold Deformation
155
AISI 80 DF Steel: Effects of Cold Deformation
156
Comparison of Three HSLA Steel Grades, Cb, Cb-V and Cb-V-Si: Strain Life
From Constant Amplitude
157
Comparison of Stress Responses: DPI vs DP2 Dual-Phase HSLA Steels
158
Dual-Phase HSLA Steel Grade: Stress Response for As-Received vs
Water-Quenched
159
Dual-Phase HSLA Steel Grade: Stress Response for As-Received vs Gas-JetCooled
160
S-N Comparison of Dual-Phase HSLA Steel Grades DPI and DP2 With
1006 161
Comparison of Dual-Phase HSLA Steel DP2 With HI-FORM 50
162
Comparison of Cyclic Strain Response Curves for Cb, Cb-V and Cb-V-Si Grades
of HSLA Steel
163
Fatigue Crack Propagation Rate: Effect of Temperature for Two HSLA Steel
Grades
164
VII
viii
Contents
5-19. Effect of R-Ratio and Test Temperature on Crack Propagation of HSLA Steel
Grade I
165
5-20. Effect of Test Temperature on Fatigue Crack Propagation Behavior for Two
HSLA Steel Grades
166
5-21. Stress-Cycle Curves for Weldments of Different HSLA Steel Grades
167
5-22. Weldments (FCA W): SAE 980 X Steel vs 1006 168
5-23. Weldments (TIG): DOMEX 640 XP Steel Welded Joints vs Parent Metal
169
5-24. Weldments (FCAW Dressed by TIG): Fatigue Life Estimates Compared With
170
Experimental Data for SAE 980 X Steel
5-25. SAE 980 X Steel Weldment (FCAW): Smooth Specimen vs TIG-Dressed vs
As-Welded
171
5-26. SAE 980 X Steel Weldment (FCAW): Lap-Shear Joints
172
5-27. Microalloyed HSLA Steels: Properties of Fusion Welds
173
5-28. Microalloyed HSLA Steels: Properties of Spot Welds
174
SECTION 6: High-Strength Alloy Steels

6-1.
6-2.
6-3.
6-4.
6-5.
6-6.
6-7.
176
HY-130 Steel: Effect of Notch Radii

176
300 M Steel: Effect of Notch Severity on Constant-Lifetime Behavior
177
TRIP Steels Compared With Other High-Strength Grades
178
Corrosion Fatigue: Special High-Strength Sucker-Rod Material
179
Corrosion Fatigue Cracking of Sucker-Rod Material
180
181
Hydrogenated Steel: Effect of Baking Time on Hydrogen Concentration
Hydrogenated Steel: Effect of Notch Sharpness
182
SECTION 7: Heat-Resisting Steels
183
7-1. 0.5%Mo Steel: Effect of Hold Time in Air and Vacuum at Different
Temperatures
183
7-2. DIN 14 Steel (1.5 Cr, 0.90 Mo, 0.25 V): Effect of Liquid Nitriding
184
7-3. 2.25Cr-1.0Mo Steel: Influence of Cyclic Strain Range on Endurance Limit in
Various Environments
185
7-4. 2.25Cr-1.0Mo Steel: Effect of Elevated Temperature
186
7-5. 2.25Cr-I.OMo Steel: Effect of Elevated Temperature and Strain Rate
187
7-6. 2.25Cr-1.0Mo Steel: Effect of Temperature on Fatigue Crack Growth Rate
188
7-7. 2.25Cr-1.0Mo Steel: Effect of Cyclic Frequency on Fatigue Crack Growth
Rate
189
7-8. 2.25Cr-1.0Mo Steel: Fatigue Crack Growth Rates in Air and Hydrogen
190
7-9. 2.25Cr-1.0Mo Steel: Effect of Holding Time
191
7-10. Cast 2.25Cr-1.0Mo Steel, Centrifugally Cast: Fatigue Properties at 540C
(1000 OF) 192
7-11. HII Steel: Crack Growth Rate in Water and in Water Vapor
193
7-12. 9.0Cr-1.0Mo Steel: Creep-Fatigue Characteristics
194
7-13. 9.0Cr-1.0Mo Modified Steel: Stress Amplitudes Developed in Cycling
195
7-14. 9.0Cr-1.0Mo Modified Steel: Effect of Deformation
196
SECTION 8: Stainless Steels
197
8-1. Type 301 Stainless Steel: Scatter Band for Fatigue Crack Growth Rates
197
8-2. Type 301 Stainless Steel: Effects of Temperature and Environment on Fatigue
Crack Growth Rate
198
8-3. Type 304 Stainless Steel: Effect of Temperature on Frequency-Modified
Strains
199
8-4. Type 304 Stainless Steel: Fatigue Crack Growth Rate-Annealed and Cold
Worked
200
8-5. Type 304 Stainless Steel: Effect of Humidity on Fatigue Crack Growth Rate
201
8-6. Type 304 Stainless Steel: Effect of Aging on Fatigue Crack Growth Rate
202
8-7. Type 304 Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rate
203
8-8. Type 304 Stainless Steel: Damage Relation at 650C (1200 OF) 204
Contents
8-9. Type 304 Stainless Steel: Fatigue Crack Growth Rate at Room and Subzero
Temperatures
205
8-10. Types 304 and 304L Stainless Steel: Effect of Cryogenic Temperatures on Fatigue
Crack Growth Rate 206
8-11. Type 304 Stainless Steel: Fatigue Crack Growth Rate in Air With Variation in
Waveforms
207
8-12. Type 304 Stainless Steel: Effect of Hold Time on Cycles to Failure 208
8-13. Type 304 Stainless Steel: Effect of Hold Time and Continuous Cycling on Fatigue
Crack Growth Rates
209
8-14. Type 304 Stainless Steel: Effect of Cyclic Frequency on Fatigue Crack Growth
Rate 210
8-15. Type 304 Stainless Steel: Effect of Frequency on Fatigue Crack Growth
Behavior 211
8-16. Type 304 Stainless Steel Welded With Type 308: Fatigue Crack Growth Rates
212
8-17. Types 304 and 310 Stainless Steel: Effect of Direction on S-N 213
8-18. Types 304, 316, 321, and 348 Stainless Steel: Effects of Temperature on Fatigue
Crack Growth Rates
214
8-19. Type 309S Stainless Steel: Effect of Grain Size on Fatigue Crack Growth
Rate 215
8-20. Type 310S Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rate
216
8-21. Type 316 Stainless Steel: Growth Rate of Fatigue Cracks in Weldments
217
8-22. Type 316 Stainless Steel: Fatigue Crack Growth Rates-Aged vs Unaged
218
8-23. Type 316 Stainless Steel: Fatigue Crack Growth Rates-Effect of Aging
219
8-24. Type 316 Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rate
220
8-25. Type 316 Stainless Steel: Effect of Cyclic Frequency on Fatigue Crack Growth
Rate 221
8-26. Type 316 Stainless Steel: Fatigue Crack Growth Rate in the Annealed
Condition 222
8-27. Type 316 Stainless Steel: Effect of Environment (Sodium, Helium, and Air) on
Cycles to Failure
223
8-28. Types 316 and 321 Stainless Steel: Effects of Gaseous Environments on Fatigue
Crack Growth Rates
224
8-29. Type 32I Stainless Steel: Effect of Hold Time on Fatigue Crack Growth Rates
225
8-30. Type 403 Stainless Steel: Effect of Environment on Fatigue Crack Growth
Rate
226
8-3I. Type 403 Modified Stainless Steel: Scatter of Fatigue Crack Growth Rates
227
8-32. Type 422 Stainless Steel: Fatigue Crack Growth Rates in Precracked
Specimens
228
8-33. Type 422 Stainless Steel: Fatigue Strength-Longitudinal vs Transverse
229
8-34. Type 422 Stainless Steel: Effect of Temperature on Fatigue Strength
230
8-35. Type 422 Stainless Steel: Effects of Delta Ferrite on Fatigue Strength 231
8-36. 17-4 PH Stainless Steel: Fatigue Crack Growth Rates in Airvs Salt Solution 232
8-37. 15-5 PH Stainless Steel: Fatigue Crack Growth Rates in Air vs Salt Solution 233
8-38. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates at Room
Temperature 234
8-39. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates in Air and Sump Tank
Water 235
8-40. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates at Subzero
Temperatures
236
8-41. PH 13-8 Mo Stainless Steel: Constant-Life Fatigue Diagram
237
8-42. Types 600 and 329 Stainless Steel: S-NCurves for Two Processing Methods
238
8-43. Grade 21-6-9 Stainless Steel: Effect of Temperature on Fatigue Crack Growth
Rates
239
8-44. Kromarc 58 Stainless Steel: Effect of Cryogenic Temperatures on Weldments
240
8-45. Pyromet 538 Stainless Steel: Effects of Welding Methods on Fatigue Crack
Growth Rates
241
8-46. Duplex Stainless Steel KCR 171: Corrosion Fatigue
242
ix
Contents
SECTION 9: Maraging Steels
243
9-1. Grades 200, 250, and 300 Maraging Steel: S-N Curves for Smooth and 'Notched
Specimens
243
9-2. Grade 300 Maraging Steel: Fatigue Life in Terms of Total Strain
244
SECTION 10: Cast Irons

10-1.
10-2.
10-3.
10-4.
10-5.
10-6.
10-7.
10-8.
10-9.
10-10.
10-11.
10-12.
10-13.
10-14.
10-15.
10-16.
10-17.
10-18.
10-19.
10-20.
10-21.
10-22.
10-23.
10-24.
10-25.
10-26.
10-27.
245
Fatigue of Cast Irons as a Function of Structure-Sensitive Parameters

245
Gray Iron: Fatigue Life, and Fatigue Limit as a Function of Temperature
246
Gray Iron: S-N Curves for Unalloyed vs Alloyed
247
Gray Iron: Effect of Environment
248
Class 30 Gray Iron: Modified Goodman Diagram for Class 30 249
Class 30 Gray Iron: Fatigue Crack Growth Rates for Class 30 250
Gray Irons: Torsional Fatigue for Various Tensile Strength Values
251
Gray Irons: Torsional Fatigue Data for Five Different Compositions
252
Gray Irons: Thermal Fatigue-Effect of Aluminum Additions
253
Gray Irons: Thermal Fatigue-Effect of Chromium and Molybdenum
Additions
254
Gray Irons: Thermal Fatigue-Room Temperature and 540C (1000 OF)
255
Gray Irons: Thermal Fatigue Properties-Comparisons With Ductile Cast Iron
and Carbon Steel
256
Cast Irons: Thermal Fatigue Properties for Six Grades
257
Ductile Iron: Effect of Microstructure on Endurance Ratio-Tensile Strength
Relationship
258
Ductile Iron: Effect of Microstructure on Endurance Ratio-Tensile Strength
Relationship
259
Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades, Using V-Notched
Specimens
260
Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades, Using Unnotched
Specimens
261
Ductile Iron: Fatigue Diagrams for Bending Stresses and Tension-Compression
Stresses
262
Ductile Iron: Effect of Surface Conditions-As-Cast vs Polished Surface
263
Ductile Iron: Fatigue Limit in Rotary Bending as Related to Hardness
264
Ductile Iron: Effect of Rolling on Fatigue Characteristics
265
Ductile Iron: Effect of Notches on a 65,800-psi-Tensile-Strength Grade
266
Ductile Iron: Fatigue Crack Growth Rate Compared With That of
Steel
267
Malleable Iron: S- N Curve Comparisons of Four Grades
268
Pearlitic Malleable Iron: Effect of Surface Conditions on S-N Curves
269
Pearlitic Malleable Iron: Effect of Nitriding 270
Ferritic Malleable Iron: Effect of Notch Radius and Depth
271
SECTION 11: Heat-Resisting Alloys

II-I.
11-2.
11-3.
11-4.
11-5.
11-6.
11-7.
11-8.
11-9.
11-10.
II-II.
11-12.
11-13.
11-14.
272
A286: Effect of Environment

272
A286: Effect of Frequency on Life at 593C (1095 OF) 273
A286: Fatigue Crack Growth Rates at Room and Elevated Temperatures
274
Astroloy: S-N Curves for Powder vs Conventional Forgings
275
Astroloy: Powder vs Conventional Forgings Tested at 705C (1300 OF)
276
FSX-430: Effect of Grain Size on Cycles to Cracking 277
FSX-430: Effect of Grain Size on Fatigue Crack Propagation Rate
278
HS-31: Effect of Testing Temperature 279
IN 738 LC Casting Alloy: Standard vs HIP'd Material
280
IN 738 LC: Effect of Grain Size on Cycles to Failure 281
IN 738 LC: Effect of Grain Size on Cycles to Cracking 282
IN 738 LC: Effect of Grain Size on Fatigue Crack Propagation Rate
283
IN 738 LC: Fatigue Crack Growth Rate at 850C (1560 OF) 284
Inconel 550: Axial Tensile Fatigue Properties in Air and Vacuum at 1090 K 285
Contents
11-15.
I 1-16.
11-17.
II-18.
11-19.
1I-20.
I 1-21.
1I-22.
11-23.
11-24.
I1-25.
11-26.
11-27.
11-28.
II-29.
11-30.
1I-31.
I 1-32.
I 1-33.
11-34.
11-35.
11-36.
11-37.
1I-38.
11-39.
11-40.
11-41.
I 1-42.
11-43.
11-44.
1I-45.
I 1-46.
11-47.
Inconel625: Effect of Temperature on Cycles to Failure 286

Inconel 706: Effect of Temperature on Fatigue Crack Growth Rate
287
Inconel "7I3C": Effect of Elevated Temperatures on Fatigue Characteristics
288
Inconel "7I3C" and As-Cast HS-31: Comparison of Two Alloys for Number of
Cycles in Thermal Fatigue to Initiate Cracks 289
Inconel 718: Effect of Frequency on Fatigue Crack Propagation Rate 290
Inconel 718: Relationship of Fatigue Crack Propagation Rate With Stress
Intensity 29I
Inconel 718: Relationship of Fatigue Crack Growth Rate With Load/Time Waveforms
292
Inconel 718: Fatigue Crack Growth Rate in Air vs Helium
293
Inconel 718: Effect of Environment on Fatigue Crack Growth Rate
294
Inconel 718: Fatigue Crack Growth Rate in Air Plus 5% Sulfur Dioxide
295
lnconel 7I8: Fatigue Crack Growth Rate in Air at Room Temperature
296
Inconel 718: Fatigue Crack Growth Rate in Air at 316C (600 OF) 297
Inconel 718: Fatigue Crack Growth Rates at Cryogenic Temperatures
301
Inconel 718 and X-750: Fatigue Crack Growth Rates at Cryogenic
Temperatures
302
Inconel X-750: Effect of Temperature on Fatigue Crack Growth Rates
303
Jethete M I52: Interrelationship of Tempering Treatment, Alloy Class, and Testing
Temperature With Fatigue Characteristics 304
Lapelloy: Interrelationship of Hardness and Strength With Fatigue
Characteristics
305
MAR-M200: Effect of Atmosphere on Cycles to Failure
306
MAR-M509: Correlation of Initial Crack Propagation and Dendrite Arm
Spacing 307
MAR-M509: Correlation Between Number of Cycles Required to Initiate a Crack
and Dendrite Arm Spacing 308
MERL 76, P/M: Axial Low-Cycle Fatigue Life of As-HIP'd Alloy at 540C
(1000 OF) 309
Nickel-Base Alloys: Effect of Solidification Conditions on Cycles to Onset
of Cracking 310
Rene 95 (As-HIP): Cyclic Crack Growth Behavior Under Continuous and HoldTime Conditions
3I I
Rene 95: Effect of Temperature on Fatigue Crack Growth Rate
312
313
S-8 I6: Effect of Notches on Cycles to Failure at 900C (1650 F)
Udimet 700: Fatigue Crack Growth Rates at 850C (1560 OF) 314
U-700 and MAR-M200: Comparison of Fatigue Properties
315
Waspaloy: Stress-Response Curves
316
X-40: Effect of Grain Size and Temperature on Fatigue Characteristics
317
Cast Heat-Resisting Alloys: Ranking for Resistance to Thermal Fatigue
318
SECTION 12: Aluminum Alloys
319
12-1. Corrosion-Fatigue Properties of Aluminum Alloys Compared With Those of

Other Alloys 319
12-2. Comparisons of Aluminum Alloys With Magnesium and Steel: Tensile Strength
vs Endurance Limit
320
12-3. Aluminum Alloys (General): Yield Strength vs Fatigue Strength
321
12-4. Comparison of Aluminum Alloy Grades for Crack Propagation Rate
322
12-5. Alloy 1100: Relationship of Fatigue Cycles and Hardness for HO and H 14
Tempers
323
12-6. Alloy 1100: Interrelationship of Fatigue Cycles, Acoustic Harmonic Generation
and Hardness
324
12-7. Alloy 2014-T6: Notched vs Unnotched Specimens; Effect on Cycles to Failure
325
12-8. Alloy 2024-T3: Effect of Air vs Vacuum Environments on Cycles to Failure
326
12-9. Alloy 2024-T4 Alclad Sheet: Effect of Bending on Cycles to Failure
327
xi
xii
Contents
12-10. Alloy 2024-T4: High-Cycle vs Low-Cycle Fatigue

328
12-11. Alloy 2024-T4: Relationship of Stress and Fatigue Cycles 329
12-12. Alloy 2024-T4: Dependence of the Average Rocking Curve Halfwidth 7J on Distance From the Surface 330
12-13. Alloys 2024 and X2024: Effect of Alloy Purity on Cycles to Failure
331
12-14. Alloys 2024 and 2124: Relationship of Particle Size and Fatigue
Characteristics
332
12-15. Alloys 2024-T4 and 2124-T4: Comparison of Resistance to Fatigue Crack
Initiation 333
12-16. Alloys 2024-TJ and 7075-T6: Summary of Fatigue Crack Growth Rates
334
12-17. Alloys 2024-T4 and 7075-T6: Effect of Product Form and Notches
335
12-18. Alloys 2024-T351 and 7075-T73XXX: Comparison of P / M Extrusions and
Rod
336
12-19. Alloy 2048-T851: Longitudinal vs Transverse for Axial Fatigue
337
12-20. Alloy 2048-T851: Notched vs Unnotched Specimens at Room and Elevated
Temperatures
338
12-21. Alloy 2048-T851: Fatigue Crack Propagation Rates in LT and TL
Orientations
339
340
12-22. Alloy 2048-T85I: Modified Goodman Diagram for Axial Fatigue
12-23. Alloy 2219-T851: Dependence of Relaxation Behavior on the Cyclic Hardening
Parameter
341
12-24. Alloy 2219-T851: Effect of Strain Amplitude on the Relaxation of Residual
Surface Stress With Fatigue 342
12-25. Alloy 2219-T851: Relationship of Fatigue Cycles to Different Depth Distributions
of Surface Stress
343
12-26. Alloy 2219-T851: Probability of Fatigue Failure
344
12-27. Alloys 3003-0, 5154-H34 and 6061-T6: Effect of Alloy on Fatigue Characteristics
of Weldments
345
12-28. Alloy 5083-0 Plate: Effect of Orientation on Fatigue Crack Growth Rates
346
12-29. Alloy 5083-0 Plate: Effect of Temperature and Humidity on Fatigue Crack Growth
Rates
347
12-30. Alloys 5086-H34, 5086-H36, 6061-T6, 7075-T73 and 2024-T3: Comparative
Resistance to Axial-Stress Fatigue
348
12-31. Alloys 5083-0/5183: Fatigue Life Predictions and Experimental Data Results for
Double V-Butt Welds
349
12-32. Alloys 5083-0/5183: Predicted Effect of Stress Relief and Stress Ratio on Fatigue
Life of Butt Welds
350
12-33. 7XXX Alloys: Cyclic Strain vs Crack Initiation Life 351
12-34. Alloy 7050: Influence of Alloy Composition and Dispersoid Effect on Mean
Calculated Fatigue Life 352
353
12-35. Alloy 7050: Effect of Grain Shape on Cycles to Failure
12-36. Alloy 7075 (TMP, T6 and T651): Effect of Thermomechanical Processing on Cycles
to Failure
354
12-37. Alloys 7075 and 7475: Effect of Inclusion Density on Cycles to Failure
355
356
12-38. Alloy 7075: Effect of TMT on Cycles to Failure
12-39. Alloys 7075 and 7050: Relative Ranking for Constant Amplitude and Periodic
Overload
357
12-40. Alloy 7075: Effect of Environment and Mode of Loading 358
12-41. Alloy 7075-T6: Effects of Corrosion and Pre-Corrosion
359
360
12-42. Alloy 7075-T73: Effect of a 3.5% NaCl Environment on Cycles to Failure
12-43. Alloy 7075: Effect of Cathodic Polarization on Fatigue Behavior
361
12-44. Alloy 7075-T6: Effect of Surface Treatments and Notch Designs on Number of
Cycles to Failure
362
364
12-45. Alloy 7075-T6: Effect of R-Ratio on Fatigue Crack Propagation
12-46. Alloy 7075: Effect of Predeformation on Fatigue Crack Propagation Rates
365
12-47. Alloys 7075 and 2024-T3: Comparative Fatigue Crack Growth Rates for Two
Alloys in Varying Humidity
366
367
12-48. Alloy 7075-T65I: Fatigue Life as Related to Harmonic Generation
12-49. Alloys 7075-T6 and 7475-T73: Effect of Laser-Shock Treatment on Fatigue
Properties
368
12-50. Alloy 7075-T6: Effect of Laser-Shock Treatment on Hi-Lok Joints
369
Contents
12-51.
12-52.
12-53.
12-54.
14-55.
12-56.
12-57.
12-58.
12-59.
12-60.
12-61.
12-62.
12-63.
12-64.
12-65.
12-66.
12-67.
12-68.
12-69.
12-70.
12-71.
12-72.
12-73.
12-74.
12-75.
12-76.
Alloy 7075 (High Purity): Effect of Iron and Silicon on Cycles to Failure
370
Alloy X-7075: Effect of Grain Size on Cycles to Failure
371
Alloy X-7075: Effect of Grain Size on Stress-Life Behavior
372
Alloy X-7075: Effect of Environment; Air vs Vacuum
373
Alloy X-7075: Effect of Environment on Two Different Grain Sizes 374
Alloy X-7075: Effect of Grain-Boundary Ledges on Cycles to Failure 375
Alloys X-7075 and 7075: Effects of Chromium Inclusions on Fatigue Crack
Propagation
376
Alloy 7475-T6: S-N Diagram for a Superplastic Fine-Grain Alloy
377
Alloy 7475: Effect of Alignment of Grain Boundaries on Cycles to Failure
378
Alloy 7475-T6: Superplastic vs Nonsuperplastic, as Related to Fatigue Crack
Growth
379
Alloys X-7075 and 7075: Effect of Chromium-Containing Inclusions on Cycles to
Failure
380
Aluminum Forging Alloys: Stress Amplitude vs Reversals to Failure
381
AI-5Mg-0.5Ag: Effect of Condition on Fatigue Characteristics
382
AI-Zn-Mg and AI-Zn-Mg-Zr: Effect of Grain Size on Strain-Life Behavior
383
AI-Zn-Mg: Strain-Life Curves of a Large-Grained Alloy
384
Aluminum With a Copper Overlay: Stress Amplitude vs Cycles to Failure
385
P/M Alloys 7090 and 7091 vs Extruded 2024 386
P / M Alloys 7090 and 709I vs 1/ M 7050 and 7075 Products
387
P/M Aluminum Alloys: Typical Fatigue Behavior
388
P / M Aluminum Alloys: Comparison With Specimens Made by Ingot
Metallurgy
389
P/M Aluminum Alloys: Comparison With Forged 7175 for Cycles to
Failure
390
Various Aluminum Alloys: Comparison of Grades for Corrosion-Fatigue Crack
Growth Rates; Air vs Salt Water 391
Various Aluminum Alloys: Comparison of Grades for Corrosion-Fatigue Crack
Growth Rates in Salt Water
392
Various Aluminum Alloys: Wrought vs Cast, and Influence of Casting Method on
Fatigue Life 393
Aluminum Casting Alloy AL-195: Interrelationship of Fatigue Properties With
Degree of Porosity
394
Aluminum Casting Alloy LM25-T6: Squeeze Formed vs Chill Cast; Effect on
Reversals to Failure
395
SECTION 13: Copper Alloys
396
13-1. Copper: Effect of Air and Water Vapor on Cycles to Failure

396
13-2. Copper: Applied Plastic-Strain Amplitude vs Fatigue Life 397
13-3. Copper Alloy CI 1000 (ETP Wire): Effect of Temperature on Fatigue
Strength
398
13-4. Copper Alloy C26000 (Cartridge Brass): Influence of Grain Size and Cold Work on
Cycles to Failure
399
13-5. Copper Alloy C83600 (Leaded Red Brass): S-N Curves; Scatter Band
400
13-6. Copper Alloy C86500 (Manganese Bronze): S-N Curves; Scatter Band
401
13-7. Copper Alloys C87500 and C87800 (Silicon Brasses): S-N Curves; Scatter
Band
402
13-8. Copper Alloy C92200 (Navy "M" Bronze): S-N Curves; Scatter Band
403
13-9. Copper Alloy C93700 (High-Leaded Tin Bronze): S-NCurves; Scatter Band
404
13-10. Copper Alloy No. 192: Effect of Salt Spray on Tubes
405
13-1 I. Copper Alloy 955: Goodman-Type Diagram 406
SECTION 14: Magnesium Alloys
407
14-1. Magnesium Casting Alloy QE22A-T6: Effects of Notches and Testing

Temperature
407
14-2. Magnesium Casting Alloy QH2 IA-T6: S- N Curves; Effects of Notches and Testing
Temperature
408
14-3. Mg-AI-Zn Casting Alloys: Effects of Surface Conditions on Fatigue
Properties
409
xiii
xiv
Contents
SECTION 15: Molybdenum
410
15-1. Molybdenum: Fatigue Limit Ratio vs Temperature
SECTION 16: Tin Alloys
410
411
16-1. Tin-Lead Soldering Alloy: S-N Data for Soldered Joints

411
16-2. Babbitt: Variation of Bearing Life With Babbitt Thickness
412
16-3. SAEI2 Bearing Alloy: Effect of Temperature on Fatigue Life 413
SECTION 17: Titanium and Titanium Alloys
414
17-1. Unalloyed Titanium, Grade 3: S-N Curves for Annealed vs Cold Rolled
414
17-2. Unalloyed Titanium, Grade 4: S-N Curves for Three Testing
Temperatures
415
17-3. Ti-24V and Ti-32V: Stress Amplitude vs Cycles to Failure 416
17-4. Ti-5AI-2.5Sn: Effects of Notches and Types of Surface Finish
417
17-5. Ti-5AI-2.5Sn and Ti-6AI-4V: Fatigue Crack Growth Rates
418
17-6. Ti-6AI-6V-2Sn: Effects of Machining and Grinding 419
17-7. Ti-6AI-6V-2Sn (HIP): S-N Curves for Titanium Alloy Powder Consolidated
by HIP
420
17-8. Ti-6AI-6V-2Sn (HIP): S-N Curves for Annealed Plate vs HIP
421
17-9. Ti-6AI-2Sn-4Zr-2Mo: Bar Chart Presentation on Effects of Machining and
Grinding 422
17-10. Ti-6AI-2Sn-4Zr-2Mo: Constant-Life Fatigue Diagram
423
17-11. Ti-6AI-2Sn-4Zr-6Mo: Low-Cycle Axial Fatigue Curves
424
17-12. Ti-8Mo-2Fe-3AI: S-NCurves; 'Solution Treated and Aged Condition
425
17-13. Ti-IOV-2Fe-3AI: S-N Curves; Notched vs Unnotched Specimens in Axial
Fatigue 426
17-14. Ti-IOV-2Fe-3AI and Ti-6AI-4V: Comparison of Fatigue Crack Growth
Rates
427
17-15. Ti-IOV-2Fe-3AI: S-N Curve; Notched Bar Fatigue Life for a Series of Forgings
Compared With Ti-6AI-4V Plate
428
17-16. Ti-I3V-IICr-3AI: Constant-Life Fatigue Diagrams
429
17-17. Ti-6AI-4V: Effect of Condition and Notches on Fatigue Characteristics 430
17-18. Ti-6AI-4V: Effect of Direction on Endurance
431
17-19. Ti-6AI-4V: Effect of Isothermally Rolled vs Extruded Material on Cycles to
Failure 432
17-20. Ti-6AI-4V: Comparison of Wrought vs Isostatically Pressed Material for Cycles
to Failure 433
17-21. Ti-6AI-4V: Effect of Fretting and Temperature on Cycles to Failure
434
17-22. Ti-6AI-4V (Beta Rolled): Effect of Finishing Operations on Cycles to
Failure 435
17-23. Ti-6AI-4V: Effect of Yield Strength on Stress-Life Behavior 436
17-24. Ti-6AI-4V: Effect of Stress Relief on Cycles to Failure
437
17-25. Ti-6AI-4V: Interrelationship of Machining Practice and Cutting Fluids on Cycles to
Failure 438
17-26. Ti-6AI-4V: Relative Effects of Machining and Grinding Operations on Endurance
Limit
439
17-27. Ti-6AI-4V: Effects of Various Metal Removal Operations on Endurance
Limit
440
17-28. Ti-6AI-4V: Effect of Texture on Fatigue Strength 441
17-29. Ti-6AI-4V: Effect of Complex Texture on Cycles to Failure 422
17-30. Ti-6AI-4V: Effect of Texture and Environment on Cycles to Failure 443
17-31. Ti-6AI-4V: Fatigue Crack Growth Rates
444
17-32. Ti-6AI-4V: Fatigue Crack Growth Rates for ISR Tee, and Extrusions 445
446
17-34. Ti-6AI-4V: Effect of Final Cooling on Fatigue Crack Growth Rates
447
17-35. Ti-6AI-4V: Effect of Dwell Time on Fatigue Crack Growth Rates
448
17-36. Ti-6AI-4V: Fatigue Crack Growth Data
449
17-37. Ti-6AI-4V P / M: Comparison of HIP'd Material With Alpha-Beta Forgings for
Cycles to Failure 450
xv
Contents
17-38. Ti-6AI-4V PI M: Comparisons of HIP'd Material With Annealed Plate for Cycles
to Failure
45 I
452
17-39. Ti-6AI-4V P/M: Effect of Powder Mesh Size on Fatigue Properties
17-40. Ti-6AI-4V P/M: Comparison of Blended Elemental, Prealloyed and Wrought
Material for Effect on Cycles to Failure
453
454
17-41. Ti-6AI-4V: P/M Compacts vs 11M Specimens: Cycles to Failure
17-42. Ti-6AI-4V: Comparison of Specimens Processed by Various Fabrication Processes
for Cycles to Failure
455
456
17-43. Ti-6AI-4V: Comparison of Fatigue Crack Growth Rate, PI M vs II M
17-44. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for Cycles to Failure
457
17-45. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for Cycles to Failure
458
17-46. Ti-6AI-4V EB Weldments: Base Metal Compared With Flawless Weldments
459
17-47. Ti-6AI-4V EB Weldments: Effects of Porosity on Cycles to Failure
460
17-48. Ti-6AI-4V Gas Metal-Arc Weldments: Effects of Porosity on Cycles to
Failure 461
17-49. Ti-6AI-4V: Unwelded vs Electron Beam Welded Material for Cycles to
Failure
462
463
17-50. Ti-6AI-4V: S-N Diagram for Laser-Welded Sheet
464
17-51. Ti-6AI-4V (Cast): S-N Diagram for Notched Specimens
SECTION 18: Zirconium
465
18-1. Zirconium 702: Effects of Notches and Testing Temperature on Cycles to

Failure
465
SECTION 19: Steel Castings
466
(For other data on steel castings see Sections 3,4 and 5, on carbon and
alloy steels.)
19-1. Steel Castings (General): Effect of Design and Welding Practice on Fatigue
Characteristics
466
19-2. Steel Castings (General): Effects of Discontinuities on Fatigue
Characteristics
467
SECTION 20: Closed-Die Forgings
468
(See also under specific grades of alloys.)
20-1. Closed-Die Steel Forgings: Effect of Surface Condition on Fatigue Limit
SECTION 21: Powder Metallurgy Parts
468
469
(See also under specific alloys.)

21-1. P/M: Relation of Density to Fatigue Limit and Fatigue Ratio
469
470
21-2. PI M: Relation of Fatigue Limit to Tensile Strength for Sintered Steels
21-3. PI M (Nickel Steels): As-Sintered vs Quenched and Tempered for Cycles to
Failure
471
2 I-4. PI M (Nickel Steels): Relation Between Fatigue Limit and Tensile Strength for
Sintered Steels
472
21-5. P/M (Nickel Steels): Effect of Notches on Cycles to Failure for the As-Sintered
Condition 473
21-6. PI M (Nickel Steels): Effect of Notches on Cycles to Failure for the Quenched and
Tempered Condition
474
21-7. P/M (Low-Carbon, 1-5%Cu): Effects of Notches and Nitriding on Cycles to
Failure
475
2 I-8. PI M (Sintered Iron, Low-Carbon, No Copper): Effect of Density and Nitriding on
Cycles to Failure
476
21-9. P/M: Effect of Nitriding on Ductile Iron and Sintered Iron (3%Cu) for Cycles to
Failure 477
SECTION 22: Composites
478
22-1. Brass/ Mild Steel Composite: Comparison of Brass-Clad Mild Steel With Brass and
Mild Steel for Cycles to Failure
478
22-2. Stainless Steell Mild Steel Composite: Comparison of Stainless-Clad Mild Steel
With Stainless Steel and Mild Steel for Cycles to Failure
479
xvi
Contents
SECTION 23: Effects of Surface Treatments
480
23-1. Carbon and Alloy Steels (Seven Grades): Effects of Nitrocarburizing on Fatigue
Strength
480
23-2. Carbon and Alloy Steels (Seven Grades): Effects of Tufftriding on Fatigue
Characteristics
481
23-3. Carbon and Alloy Steels (Six Grades): Effects of Nitriding on Fatigue
Strength 482
23-4. Carbon-Manganese Steel: Effects of Nickel Coating on Fatigue Strength 483
SECTION 24: Test Results for Component Parts 484

24-1. Coil Springs, Music Wire (Six Sizes): Data Presented by Means of a Goodman
Diagram
484
24-2. Coil Springs: S-N Data for Oil-Tempered and Music Wire Grades
485
24-3. Coil Springs: Effects of Shot Peening on Cycles to Failure 486
24-4. Coil Springs, 8650 and 8660 Steels: Relation of Design Stresses and Probability of
Failure 487
.
24-5. Coil Springs, HSLA Steels: Effects of Corrosion on Cycles to Failure 488
24-6. Leaf Springs, 5160 Steel: Maximum Applied Stress vs Cycles to Failure 489
24-7. Front Suspension Torsion Bar Springs, 5160H Steel: Distribution of Fatigue
Results for Simulated Service Testing 490
24-8. Gears, Carburized Low-Carbon Steel: Relation of Life Factor to Required
Life 491
24-9. Gears, Carburized Low-Carbon Steel: Bending Stress vs Cycles to Failure 492
24-10. Gears, Carburized Low-Carbon Steel: Effect of Shot Peening on Cycles to
Failure 493
24-11. Gears, Carburized Low-Carbon Steel: Probability-Stress-Life Design Curves
494
24-12. Gears, 8620H Carburized: Bending or Contact Stress vs Cycles to Fracture or
Pitting
495
24-13. Gears, 8620H Carburized: A Weibull Analysis of Bending Fatigue Data 496
24-14. Gears, 8620H Carburized: T-N Curves for Six-Pinion, Four-Square Tests
497
24-15. Hypoid Gears, 8620H Carburized: Minimum Confidence Level; Stress vs Cycles to
Rupture
498
24-16. Hypoid, Zero I and Spiral Bevel Gears, 8620H Carburized: S-NScatter Band and
Minimum Confidence Level 499
24-17. Spiral Bevel and Zero I Bevel Gears, 8620H Carburized: S-N Scatter Band and
Minimum Confidence Level 500
24-18. Gears, 8620H Case Hardened: Relation of Life Factor to Cycles to Rupture
501
24-19. Bevel Gears, Low-Carbon Steel Case Hardened: Relation of Life Factor to Cycles
to Rupture for Various Confidence Levels 502
24-20. Gears, AMS 6265: S-N Data for Cut vs Forged
503
24-21. Spur Gears, 8620H: S-N Data for Cut vs Forged
504
24-22. Gears and Pinions: PIM 4600V vs 4615; Weibull Distributions
505
24-23. Gears and Pinions: PIM Grades 4600V and 2000 vs 4615; Percent Failure vs
Time
506
24-24. Gear Steel AMS 6265: Parent Metal vs Electron Beam Welded
507
24-25. Gears, 42 CrMo4 (German Specification): S-N Curves for Various Profiles
508
24-26. Gears, 42 CrMo4 (German Specification): Endurance Test Results in the Weibull
Distribution Diagram
509
24-27. Bolts, 1040 and 4037 Steels: Maximum Bending Stress vs Number of Stress
Cycles
510
24-28. Bolts: S-N Data for Roll Threading Before and After Heat Treatment
511
24-29. Power Shafts, AMS 6382 and AMS 6260: Electron Beam Welded vs Silver Brazed
Joints
512
24-30. Axle Shafts, 1046, 1541 and 50B54 Steels: S-N Data for Induction Hardening vs
Through Hardening 513
24-31. Steel Rollers, 8620H Carburized: Effects of Carburizing Temperature and
Quenching Practice on Surface Fatigue
514
Contents
24-32. Steel Rollers, 8620H Carburized: Effects of Carburizing Temperature and

Quenching Practice on Surface Fatigue 515
24-33. Linkage Arm, Cast Low-Carbon Steel: Starting Crack Size vs Cycles to
Failure 516
24-34. Notched Links, Hot Rolled Low-Carbon Steel: S-N Data for Component Test
Model
517
24-35. Fuselage Brace, Ti-6AI-6V-2Sn: Fatigue Endurance of HIP-Consolidated
Powder 518
xvii
Fatigue Testing
Introduction
Fatigue is the progressive, localized, permanent structural change that occurs in materials
subjected to fluctuating stresses and strains that
may result in cracks or fracture after a sufficient
number of fluctuations. Fatigue fractures are
caused by the simultaneous action of cyclic
stress, tensile stress and plastic strain. If anyone
of these three is not present, fatigue cracking will
not initiate and propagate. The cyclic stress
starts the crack; the tensile stress produces crack
growth (propagation). Although compressive
stress will not cause fatigue, compression load
may do so.
The process of fatigue consists of three stages:
Initial fatigue damage leading to crack nucleation and crack initiation
Progressive cyclic growth of a crack (crack
propagation) until the remaining un cracked
cross section of a part becomes too weak to
sustain the loads imposed
Final, sudden fracture of the remaining
cross section
Fatigue cracking normally results from cyclic
stresses that are well below the static yield
strength of the material. (In low-cycle fatigue,
however, or if the material has an appreciable
work-hardening rate, the stresses also may be
above the static yield strength.)
Fatigue cracks initiate and propagate in regions where the strain is most severe. Because
most engineering materials contain defects and
thus regions of stress concentration that intensify
strain, most fatigue cracks initiate and grow
from structural defects. Under the action of cyclic loading, a plastic zone (or region of deformation) develops at the defect tip. This zone of high
deformation becomes an initiation site for a fatigue crack. The crack propagates under the applied stress through the material until complete
fracture results. On the microscopic scale, the
most important feature of the fatigue process is
nucleation of one or more cracks under the influ-
ence of reversed stresses that exceed the flow

stress, followed by development of cracks at persistent slip bands or at grain boundaries.
Prediction of Fatigue Life

The fatigue life of any specimen or structure is
the number of stress (strain) cycles required to
cause failure. This number is a function of many
variables, including stress level, stress state, cyclic wave form, fatigue environment, and the
metallurgical condition of the material. Small
changes in the specimen or test conditions can
significantly affect fatigue behavior, making analytical prediction of fatigue life difficult. Therefore, the designer may rely on experience with
similar components in service rather than on
laboratory evaluation of mechanical test specimens. Laboratory tests, however, are essential in
understanding fatigue behavior, and current
studies with fracture mechanics test specimens
are beginning to provide satisfactory design
criteria.
Laboratory fatigue tests can be classified as
crack initiation or crack propagation. In crack
initiation testing, specimens or parts are subjected to the number of stress cycles required for
a fatigue crack to initiate and to subsequently
grow large enough to produce failure.
In crack propagation testing, fracture mechanics methods are used to determine the crack
growth rates of preexisting cracks under cyclic
loading. Fatigue crack propagation may be
caused by cyclic stresses in a benign environment, or by the combined effects of cyclic stresses
and an aggressive environment (corrosion fatigue).
Fatigue Crack Initiation

Most laboratory fatigue testing is done either
with axial loading, or in bending, thus producing
only tensile and compressive stresses. The stress
usually is cycled either between a maximum and
a minimum tensile stress, or between a maximum
tensile stress and a maximum compressive stress.
Fatigue Testing
1100
The latter is considered a negative tensile stress,

is given an algebraic minus sign, and therefore is
known as the minimum stress.
The stress ratio is the algebraic ratio of two
specified stress values in a stress cycle. Two
commonly used stress ratios are the ratio, A, of
the alternating stress amplitude to the mean
stress (A = Sal S m) and the ratio, R, of the minimum stress to the maximum stress (R = Sminl
.;
Smax)'
If the stresses are fully reversed, the stress ratio

R becomes -1; if the stresses are partially reversed, R becomes a negative number less than 1.
If the stress is cycled between a maximum stress
and no load, the stress ratio R becomes zero. If
the stress is cycled between two tensile stresses,
the stress ratio R becomes a positive number less
than 1. A stress ratio R of 1indicates no variation
in stress, making the test a sustained-load creep
test rather than a fatigue test.
Applied stresses are described by three parameters. The mean stress, S m' is the algebraic
average of the maximum and minimum stresses
in one cycle, S m = (S max + Smin) / 2. In the completely reversed cycle test, the mean stress is zero.
The range of stress, S" is the algebraic difference
between the maximum and minimum stresses in
one cycle, S, = Smax - Smin' The stress amplitude,
S a' is one half the range of stress, Sa = S,/ 2 =
(Smax - Smin)/2.
During a fatigue test, the stress cycle usually is

maintained constant so that the applied stress
conditions can be written Sm SO' where S mis the
static or mean stress, and Sa is the alternating
stress, which is equal to half the stress range.
Nomenclature to describe test parameters involved in cyclic stress testing are shown in Fig. 1.
Smax I---~:------;...----r---,----
Time
Ol-\-.........,~-----Ic-I-'--~r--f--+----
Smin
L..-
--""-
......._ - ' -_ _
Fig. 1 Nomenclature to describe test

parameters involved in cyclic stress
testing
S-N Curves. The results offatigue crack initiation tests usually are plotted as maximum stress,
minimum stress, or stress amplitude to number
of cycles, N, to failure using a logarithmic scale
for the number of cycles. Stress is plotted on
either a linear or a logarithmic scale. The result-
I
150
2340 steel
o~ '"'
48 HRC
~ (unnotchedl'r
1000
<ll
900
:2:
800
a..
<I)"
700
Q)
600
0.
<ll
<Il
<Il
U5
500
400
300
200
125 'iii
1::-- - - -
f-
_.
'1...
.:.!.
1 --
Fatigue limit Sf
<I)"
2340 steel-= 100

Q)
48 HRC _
u
(notched)
.~
-cr....
Fatigue limit, Sf
l
Aluminum alloy
"""Q; ~ ,}075-T6
j-a~-
75
0.
E
<ll
50
<Il
<Il
~,
U5
25
Stress ratio (R) = - 1
Number of cycles to fracture, N

Fig. 2 Typical S-N curves for constant
amplitude and sinusoidal loading
ing plot of the data is an S-N curve. Three typical

S-N curves are shown in Fig. 2.
The number of cycles of stress that a metal can
endure before failure increases with decreasing
stress. For some engineering materials such as
steel (see Fig. 2) and titanium, the S-N curve becomes horizontal at a certain limiting stress.
Below this limiting stress, known as the fatigue
limit or endurance limit, the material can endure
an infinite number of cycles without failure.
Fatigue Limit and Fatigue Strength. The horizontal portion of an S-N curve represents the
maximum stress that the metal can withstand for
an infinitely large number of cycles with 50%
probability of failure and is known as the fatigue
(endurance) limit, Sf' Most nonferrous metals do
not exhibit a fatigue limit. Instead, their S-N
curves continue to drop at a slow rate at high
numbers of cycles, as shown by the curve for
aluminum alloy 7075-T6 in Fig. 2.
For these types of metals, fatigue strength
rather than fatigue limit is reported, which is the
stress to which the metal can be subjected for a
specified number of cycles. Because there is no
standard number of cycles, each table of fatigue
strengths must specify the number of cycles for
which the strengths are reported. The fatigue
strength of nonferrous metals at 100million (108)
or 500 million (5 X 108) cycles is erroneously
called the fatigue limit.
Low-Cycle Fatigue. For the low-cycle fatigue
region (N< 104 cycles) tests are conducted with
controlled cycles of elastic plus plastic strain,
Introduction
rather than with controlled load or stress cycles.
Under controlled strain testing, fatigue life behavior is represented by a log-log plot of the total
strain range, dE, versus the number of cycles to
failure (Fig. 3).
The total strain range is separated into elastic
and plastic components. For many metals and
alloys, the elastic strain range, dE eo is equal to the
stress range divided by the modulus of elasticity.
The plastic strain range, dE p' is the difference between the total strain range and the elastic strain
range.
Stress-Concentration Factor. Stress is concentrated in a metal by structural discontinuities,
such as notches, holes, or scratches, which act as
stress raisers. The stress-concentration factor,
K" is the ratio of the area test stress in the region
of the notch (or other stress concentrators) to the
corresponding nominal stress. For determination of K" the greatest stress in the region of the
notch is calculated from the theory of elasticity,
or equivalent values are derived experimentally.
The fatigue notch factor, Kf> is the ratio of the
fatigue strength of a smooth (unnotched) specimen to the fatigue strength of a notched specimen at the same number of cycles.
Fatigue notch sensitivity, q, for a material is
determined by comparing the fatigue notch factor, K J, and the stress-concentration factor, K"
for a specimen of a given size containing a stress
concentrator of a given shape and size. A common definition of fatigue notch sensitivity is q =
(KJ - l)f(K, - 1), in which q may vary between
zero (where K J = 1) and 1 (where KJ = K,). This
value may be stated as percentage.

In large structural components, the existence
of a crack does not necessarily imply imminent
failure of the part. Significant structural life may
remain in the cyclic growth of the crack to a size
at which a critical failure occurs. The objective of
fatigue crack propagation testing is to determine
the rates at which subcritical cracks grow under
cyclic loadings prior to reaching a size critical for
fracture.
The growth or extension of a fatigue crack
under cyclic loading is principally controlled by
maximum load and stress ratio. However, as in
crack initiation, there are a number of additional
factors that may exert a strong influence, including environment, frequency, temperature, and
grain direction. Fatigue crack propagation testing usually involves constant-load-amplitude cy-
10- 3
L..-_ _--I.
10- 1
..J
-l..
--L.
103
10
105
Cycles to failure
Fig.3 Typical plot of strain range versus
cycles-to-failure for low-cycle fatigue
cling of notched specimens that have been precracked in fatigue. Crack length is measured as a
function of elapsed cycles, and these data are
subjected to numerical analysis to establish the
rate of crack growth, da l d N,
Crack growth rates are expressed as a function
of the crack tip stress-intensity factor range, dK.
The stress-intensity factor is calculated from expressions based on linear elastic stress analysis
and is a function of crack size, load range, and
cracked specimen geometry. Fatigue crack growth
data are typically presented in a log-log plot of
da/dNversus s (Fig. 4).
:J.K. ksivTn.
5.
10
15
20
25
30 35
1.-----r-"T"T"--r--r-~--,.,..._,_-,-,
10- 1 I---+_--+----;:--et-'~-__l
10- 3
Q)
10- 2
>-
c::
10- 4 <:"tl
10- 3
~
10 -5
10-4
I-~=---+_--+----+---__l
o~
10- 5
L..-_ _.L-_ _..I.--- _ _-L._ _- '
10
20
30
10- 6
40
J.K. MPa\ m
Fig. 4 Fatigue crack propagation rate
data in 7075-T6 aluminum alloy (R <0)
LIVE GRAPH
Click here to view
Fatigue Testing

Crack initiation tests are procedures in which
a specimen or part is subjected to cyclic loading
to failure. A large portion of the total number of
cycles in these tests is spent initiating the crack.
Although crack initiation tests conducted on
small specimens do not precisely establish the fatigue life of a large part, such tests do provide
data on the intrinsic fatigue crack initiation behavior of a metal or alloy. As a result, such data
can be utilized to develop criteria to prevent fatigue failures in engineering design. Examples of
the use of small-specimen fatigue test data can be
found in the basis of the fatigue design codes for
boilers and pressure vessels, complex welded, riveted, or bolted structures, and automotive and
aerospace components.
Fatigue Cracking
Fatigue cracks normally result from cyclic
stresses that are below the yield strength of the
metal. In low-cycle fatigue, however, the cyclic
stress may be above the static yield strength, especially in a material with an appreciable workhardening rate. Generally, a fatigue crack is initiated at a highly stressed region of a component
subjected to cyclic loading of sufficient magnitude. The crack then propagates in progressive
cyclic growth through the cross section of the
part until the maximum load cannot be carried,
and complete fracture results.
Crack Nucleation. A variety of crystallographic features have been observed to nucleate
fatigue cracks. In pure metals, tubular holes that
develop in persistent slip bands, slip band extrusion-intrusion pairs at free surfaces, and twin
boundaries are common nucleation sites. Grain
boundaries in polycrystalline metals, even in the
absence of inherent grain boundary weakness,
are crack nucleation sites. At high strain rates,
this appears to be the preferred site. Nucleation
at grain boundaries appears to be a geometrical
effect, whereas nucleation at twin boundaries is
associated with active slip on crystallographic
planes immediately adjacent and parallel to the
twin boundary.
The foregoing processes also occur in alloys
and heterogeneous materials. However, alloying
and commercial production practices introduce
segregation, inclusions, second-phase particles,
and other features that disturb the structure. All
of these phenomena have a significant influence

on the crack nucleation process. In general, alloying that (1) enhances cross slip, (2) enhances
twinning, or (3) increases the rate of work hardening will stimulate crack nucleation. On the
other hand, alloying usually raises the flow stress
of a metal, thus offsetting its potentially detrimental effect on fatigue crack nucleation.
Crack Initiation. Fatigue cracks initiate at
points of maximum local stress and minimum
local strength. The local stress pattern is determined by the shape of the part and by the type
and magnitude of the loading. In addition to the
geometric features of a part, features such as surface and metallurgical imperfections can act to
concentrate stress locally. Surface imperfections
such as scratches, dents, burrs, cuts, and other
manufacturing flaws are the most obvious sites
at which fatigue cracks initiate. Except for instances where internal defects or special surfacehardening treatments are involved, fatigue cracks
initiate at the surface.
Relation to Environment. Corrosion fatigue
describes the degradation of the fatigue strength
of a metal by the initiation and growth of cracks
under the combined action of cyclic loading and
a corrosive environment. Because it is a synergistic effect of fatigue and corrosion, corrosion fatigue can produce a far greater degradation in
strength than either effect acting alone or by superposition of the singular effects. An unlimited
number of gaseous and liquid mediums may affect fatigue crack initiation in a given material.
Fretting corrosion, which occurs from relative
motion between joints, may also accelerate fatigue crack initiation.
Fatigue Testing Regimes

The magnitude of the nominal stress on a cyclically loaded component frequently is measured by the amount of overstress-that is, the
amount by which the nominal stress exceeds the
fatigue limit or the long-life fatigue strength of
the material used in the component. The number
of load cycles that a component under low overstress can endure is high; thus, the term highcycle fatigue is often applied.
As the magnitude of the nominal stress increases, initiation of multiple cracks is more
likely. Also, spacing between fatigue striations,
which indicate the progressive growth of the
crack front, is increased, and the region of final
fast fracture is increased in size.

Low-cycle fatigue is the regime characterized
by high overstress. The arbitrary, but commonly
accepted, dividing line between high-cycle and
low-cycle fatigue is considered to be about 104 to
105 cycles. In practice, this distinction is made by
determining whether the dominant component
of the strain imposed during cyclic loading is
elastic (high cycle) or plastic (low cycle), which in
turn depends on the properties of the metal as
well as the magnitude of the nominal stress.
Presentation of Fatigue Data. High-cycle fatigue data are presented graphically as stress (S)
versus cycles-to-failure (N) in S-N diagrams or
S- N curves. These are described in the Introduction to this Section along with the symbols and
nomenclature commonly applied in fatigue testing. Because the stress in high-cycle fatigue tests
is usually within the elastic range, the calculation
of stress amplitude, stress range, or maximum
stress on the S-axis is made using simple equations from mechanics of materials; i.e., stress calculated using the specimen dimensions and the
controlled load or deflection applied axially, in
flexure, or in torsion.
Figure 5 illustrates a stress-strain loop under
controlled constant-strain cycling in a low-cycle
fatigue test. During initial loading, the stressstrain curve is O-A-B. Upon unloading, yielding
begins in compression at a lower stress C due to
the Bauschinger effect. In reloading in tension, a

hysteresis loop develops. The dimensions of this
loop are described by its width df (the total
strain range) and its height da (the stress range).
The total strain range df consists of an elastic
strain component df e = dalE and a plastic strain
component df p
The width of the hysteresis loop depends on
the level of cyclic strain. When the level of cyclic
strain is small, the hysteresis loop becomes very
narrow. For tests conducted under constant df,
the stress range da usually changes with an increasing number of cycles.
The common method of presenting low-cycle
fatigue data is to plot either the plastic strain
range, df p' or the total strain range, df, versus N.
When plotted using log-log coordinates, a straight
line can befit to the dfp-Nplot. The slope of this
line in the region where plastic strain dominates
has shown little variation for the large number of
metals and alloys tested in low-cycle fatigue, the
average value being Y2. This power-law relationship between dfpand Nis known as the CoffinManson relationship. Figure 6 is an example of
the typical presentation of low-cycle fatigue test
results.
Classification of Fatigue
Testing Machines
Fatigue test specimens are primarily described
by the mode of loading:
Direct (axial) stress

Plane bending
Rotating beam
Alternating torsion
Combined stress
Testing machines, however, may be universaltype machines that are capable of conducting all
of the above modes ofloading, depending on the
fixturing used.
Fig. 5 Stress-strain loop for constantstrain cycling
Fatigue Testing Machine Components

Whether simple or complex, all fatigue testing
machines consist of the same basic components:
a load train, controllers, and monitors. The load
train consists of the load frame, gripping devices,
test specimen, and drive (loading) system. Typical load train components in an electrohydraulic
axial fatigue machine are shown in Fig. 7.
The load frame is the structure of the machine
that reacts to the forces applied to the specimen
by the drive system.
Fatigue Testing
Ia-...
IU
r--..... ......
Q.
'"
"1
<Ii
01
10- 1
r-, <,
........ --,
~
~
C
1Il
.~
1Il
<II
a:::
10-
"- ..........
~ ~ O.n
00
~ r-,
10
.......
"
Cycles to failure, N
Fig. 6
Low-cycle fatigue curve (t1p versus N) for type 347 stainless steel
Fig. 7 Schematic of the load train in an

electrohydraulic axial fatigue machine
The drive system is the most significant feature

of a fatigue testing system and usually is electrically powered. The simplest systems use electric
motors to act on test specimens via cams, levers,
or rotating grips. In electrohydraulic machines,
the motors drive hydraulic pumps to provide
service pressure for control of the motion and
force of a hydraulic piston actuator. Electromagnetic excitation can be used to excite a mass
or inertia system to load a specimen.
Control Systems. The controls and controllers

manually or automatically initiate power and
test, adjust, and maintain the controlled test parameter(s). Controllers also terminate the test at
a predefined status (failure,' load drop, extension, or deflection limit). The control of timevarying deflection or displacement can be obtained in mechanical systems by cam-operated
deflection levels, a rotating eccentric mass, or
hydraulically through a piston limited by stops.
Control in most simple machines and drive
systems is obtained via the open-loop mode. In
such systems, the magnitude of force and displacement initially set by the control system remains constant throughout the test.
Sensors are required to measure the load,
strain, displacement, deflection, and cycle count.
Some devices provide an output signal to the
controller, or to a readout device in the case of
uncontrolled parameters. Common sensors are
load cells (resistance strain gage bridges calibrated to load) inserted in the load train. Pressure transducers are used in hydraulic or pneumatic actuator devices.
Loading fixtures to alter the mode of loading
provide versatility. Fixtures can be designed to
convert the axial force provided by a hydraulic
actuator to perform four-point bending or torsion testing. Similarly, fixtures attached to an
oscillating platen of a rotating-eccentric-masstype machine can facilitate axial, bending, and
torsion fatigue testing of specimens.
{a}
{e}
{bl
{dl
lei
Ifl
(g)
(a) Standard grip body for wedge-type grips. (b) V-grips for rounds for use in standard grip body. (c) Flat grips for
specimens for use in standard grip body. (d) Universal open-front holders. (e) Adapters for special samples (screws,
bolts, studs, etc.) for use with universal open-front holders. (f) Holders for threaded samples. (g) Snubber-type wire
grips for flexible wire or cable.
Fig. 8
Grip designs used for axial fatigue testing
Grips. Proper gripping is not simply the attachment of the test specimen in the load train.
Grip failure sometimes occurs prior to specimen
failure. Frequently, satisfactory gripping evolves
after specimen design development. Care must
be taken in grip design and specimen installation
in the grips to prevent misalignment. The grips
shown in Fig. 8 are typical of those used for axial
fatigue tests.
Axial (Direct-Stress)
Fatigue Testing Machines
The direct-stress fatigue testing machine subjects a test specimen to a uniform stress or strain
through its cross section. For the same cross section, an axial fatigue testing machine must be
able to apply a greater force than a static bending
machine to achieve the same stress.
Electromechanical systems have been developed for axial fatigue studies. Generally, these
are open-loop systems, but often have partial
closed-loop features to continuously correct
mean load.
In crank and lever machines, a cyclic load is
applied to one end of the test specimen through a
deflection-calibrated lever that is driven by a
variable-throw crank. The load is transmitted to
the specimen through a flexure system, which
provides straight-line motion to the specimen.
The other end of the specimen is connected to a
hydraulic piston that is part of an electrohydraulically controlled load-maintaining system that
senses specimen yielding. This system automatically and steplessly restores the preset load
through the hydraulic piston.
Servohydraulic closed-loop systems offer optimum control, monitoring, and versatility in fatigue testing systems. These can be obtained as
component systems and can be upgraded as required. A hydraulic actuator typically is used to
apply the load in axial fatigue testing.
Electromagnetic or magnetostrictive excitation is used for axial fatigue testing machine
drive systems, particularly when low-load amplitudes and high-cycle fatigue lives are desired in
short test durations. The high cyclicfrequency of
operation of these types of machines enables testing to long fatigue lives (> 108 cycles) within
weeks.
Bending Fatigue Machines
The most common types of fatigue machines
are small bending fatigue machines, In general,
these simple, inexpensive systems allow laboratories to conduct extensive test programs with a
low equipment investment.
Cantilever beam machines, in which the test
specimen has a tapered width, thickness, or diameter, result in a portion ofthe test area having
uniform stress with smaller load requirements
than required for uniform bending or axial fatigue of the same section size.
Rotating Beam Machines. Typical rotating
beam machine types are shown in Fig. 9. The
R. R. Moore-type machines (Fig. 9a) can operate
up to 10000 rpm. In all bending-type tests, only
the material near the surface is subjected to the
maximum stress; therefore, in a small-diameter
specimen, only a very small volume of material is
under test.
Torsional Fatigue Testing Machines
Torsional fatigue tests can be performed on
axial-type machines using the proper fixtures if
the maximum twist required is small. Specially
Fatigue Testing
~Load
(b)
(a)
(a) Four-point loading R.R. Moore testing machine. (b) Single-end rotating cantilever testing machine.
Fig. 9 Schematic of rotating beam fatigue testing machines

Program
- - - - - Torque feedback - -
---I
I
I
L----
Angular
display feedback
Hydraulic
power
supply
I
I
...-----i----r--....,
I
I
Specimen
Hydraulic
service
manifold
Displacement
transducer
.,.,C:::}:;;:;;;:;;;:;:l:::;~~,.
Rota ry actu ato r
Fig. 10 Schematic of a servohydraulic torsional fatigue testing machine
designed torsional fatigue testing machines consist of electromechanical machines, in which linear motion is changed to rotational motion by
the use of cranks, and servo hydraulic machines,
in which rotary actuators are incorporated in a
closed-loop testing system (Fig. 10).
Special-Purpose Fatigue
Testing Machines
To perform fatigue testing of components that
are prone to fatigue failure (gears, bearings, wire,
etc.), special devices have been used, sometimes
as modifications to an existing fatigue machine.
Wire testers are a modification of rotating beam
machines, in which a length of the test wire is
used as the beam and is deflected (buckled) a

known amount and rotated.
Rolling contact fatigue testers usually are
constant-load machines in which a Hertzian contact stress between two rotating bearings is applied until occurrence of fatigue failure by pitting
or spalling is indicated by a vibration or noise
level in the system. Rolling contact fatigue of ball
and roller bearings under controlled lubrication
conditions is a specialized field of fatigue testing.
Multiaxial Fatigue Testing Machines
Many special fatigue testing machines have
been designed to apply two or more modes of
loading, in or out of phase, to specimens to de-

D
termine the properties of metals under biaxial or

triaxial stresses.
( ~=====-t-$
Fatigue Test Specimens

A typical fatigue test specimen has three areas:
the test section and the two grip ends. The grip
ends are designed to transfer load from the test
machine grips to the test section and may be
identical, particularly for axial fatigue tests. The
transition from the grip ends to the test area is
designed with large, smoothly blended radii to
eliminate any stress concentrations in the transition.
The design and type of specimen used depend
on the fatigue testing machine used and the objective of the fatigue study. The test section in the
specimen is reduced in cross section to prevent
failure in the grip ends and should be proportioned to use the upper ranges of the load capacity ofthe fatigue machine; i.e., avoiding very low
load amplitudes where sensitivity and response
of the system are decreased. Several types of fatigue test specimens are illustrated in Fig. 11.
~R
4.8 mm (3116 in.)
D, selected on basis of ultimate strength

of material R, 12.7 mm (0.50 in.)
(a)
30 mm (13/16 in.)"]
50 ~m_1
-$-O~2m.)
Tapered D, 12.7 mm (0.50 in.)
Ib)
I'" 90
mm (3:6
i ~ 19
mm (% in.)
~~-R~
'\ I
12 mm (0.48 in.)
Effect of Stress Concentration

Fatigue strength is reduced significantly by the
introduction of a stress raiser such as a notch or
hole. Because actual machine elements invariably contain stress raisers such as fillets, keyways, screw threads, press fits, and holes, fatigue
cracks in structural parts usually initiate at such
geometrical irregularities.
An optimum way of minimizing fatigue failure
is the reduction of avoidable stress raisers
through careful design and the prevention of accidental stress raisers by careful machining and
fabrication. Stress concentration can also arise
from surface roughness and metallurgical stress
raisers such as porosity, inclusions, local overheating in grinding, and decarburization.
The effect of stress raisers on fatigue is generally studied by testing specimens containing a
notch, usually a V-notch or a U-notch. The presence of a notch in a specimen under uniaxial load
introduces three effects: (1) there is an increase or
concentration of stress at the root of the notch,
(2) a stress gradient is set up from the root of the
notch toward the center of the specimen, and (3)
a triaxial state of stress is produced at the notch
root.
The ratio of the maximum stress in the region
of the notch (or other stress concentration) to the
corresponding nominal stress is the stress-con-
:cD
D, 5 to 10 mm (0.20 to 0.40 in.) selected on

basis of ultimate strength of material
R, 90 to 250 mm (3.5 to 10 in.)
[c]
.25 mm (1.0 in.) D
38 mm (1V2 in.)
~---~
~3-~
~
.
5"43'
38 mm (1'12 in.)
(d)
D, selected on basis of ultimate strength of material
R, 75 to 250 mm (3 to 10 in.)
Ie)
(a)Torsional specimen. (bl Rotating cantilever beam specimen. (c) Rotating beam specimen. (d) Plate specimen for
cantilever reverse bending. Ie) Axial loading specimen.
Fig. 11
Typical fatigue test specimens
10
Fatigue Testing
cent ration factor, K,(see the Introduction to this
Section). In some situations, values of K,can be
calculated using the theory of elasticity, or can be
measured using photoelastic plastic models.
The effect of notches on fatigue strength is determined by comparing the S-N curves of notched
and unnotched specimens. The data for notched
specimens usually are plotted in terms of nominal stress based on the net cross section of the
specimen. The effectiveness of the notch in decreasing the fatigue limit is expressed by the
fatigue-notch factor, K p This factor is the ratio
of the fatigue limit of unnotched specimens to the
fatigue limit of notched specimens.
For materials that do not exhibit a fatigue
limit, the fatigue-notch factor is based on the fatigue strength at a specified number of cycles.
Values of KJhave been found to vary with (1)
severity of the notch, (2) type of notch, (3) material, (4) type of loading, and (5) stress level.
Effect of Test Specimen Size

It is not possible to predict directly the fatigue
performance oflarge machine members from the
results oflaboratory tests on small specimens. In
most cases, a size effect exists; i.e., the fatigue
strength of large members is lower than that of
small specimens. Precise determination of this
phenomenon is difficult. It is extremely difficult
to prepare geometrically similar specimens of increasing diameter that have the same metallurgical structure and residual stress distribution
throughout the cross section. The problems in fatigue testing of large specimens are considerable,
and few fatigue machines can accommodate
specimens with a wide range of cross sections.
Changing the size of a fatigue specimen usually results in variations oftwo factors. First, increasing the diameter increases the volume or
surface area of the specimen. The change in
amount of surface is significant, because fatigue
failures usually initiate at the surface. Secondly,
for plain or notched specimens loaded in bending
or torsion, an increase in diameter usually decreases the stress gradient across the diameter
and increases the volume of material that is
highly stressed.
Experimental data on the size effect in fatigue
typically show that the fatigue limit decreases
with increasing specimen diameter. Horger's
data for steel shafts tested in reversed bending
(Table 1) show that the fatigue limit can be appreciably reduced in large section sizes.
Table 1 Effect of specimen size on the fatigue

limit of normalized plain carbon steel in
reversed bending
Specimen diameter
mm
in.
Fatigue limit
MPa
ksi
7.6
38
152
248
200
144
0.30
l.50
6.00
36
29
21
Surface Effects and Fatigue

Generally, fatigue properties are very sensitive
to surface conditions. Except in special cases
where internal defects or case hardening is involved, all fatigue cracks initiate at the surface.
Factors that affect the surface of a fatigue specimen can be divided into three categories: (1) surface roughness or stress raisers at the surface, (2)
changes in the properties of the surface metal,
and (3) changes in the residual stress condition of
the surface. Additionally, the surface may be
subjected to oxidation and corrosion.
Surface Roughness. In general, fatigue life increases as the magnitude of surface roughness
decreases. Decreasing surface roughness minimizes local stress raisers. Therefore, special attention must be given to the surface preparation
of fatigue test specimens. Typically, a metallographic finish, free of machining grooves and
grinding scratches, is necessary. Figure 12 illustrates the effects that various surface conditions
have on the fatigue properties of steel.
Effect of Mean Stress

A series of fatigue tests can be conducted at
various mean stresses, and the results can be
plotted as a series of S-N curves. A description of
applied stresses and S-N curves can be found in
the Introduction to this Section. For design purposes, it is more useful to know how the mean
stress affects the permissible alternating stress
amplitude for a given life (number of cycles).
This usually is accomplished by plotting the allowable stress amplitude for a specific number of
cycles as a function of the associated mean stress.
At zero mean stress, the allowable stress amplitude is the effective fatigue limit for a specified
number of cycles. As the mean stress increases,
the permissible amplitudes steadily decrease. At
a mean stress equal to the ultimate tensile
strength of the material, the permissible amplitude is zero.
The two straight lines and the curve shown in
11

1000 ,...---,....----,----r---,-------r-----r------,-----,...-----,
800
900
ro 700
a..
~
~
~
600
r~~;;::~~;;;;t;;~==~===~~=~===t===~==~
100 'iii
~--_t_---~2.......,_1_--~~=_---=""""
90 -""
vi
.......= _ - _ + - - _ 1 _ - - - - _ t _ - _ _ l
500
W ~
70
~
~
>
60
400
Q)
50
>
~
>
~ 300 t-----t-----t---+----+----+--"""""~--~""""""----+--___I
40
~
~
Qi
Qj
~
U;
~
200 I----t----+---+---+-----+---+---+--~.,___f-----~
10
20
50
100
200
30
500
Life, 1000 cycles

Fig. 12
Effect of surface conditions on the fatigue properties of steel (302 to 321 HB)
Fig. 13 represent the three most widely used empirical relationships for describing the effect of
mean stress on fatigue strength. The straight line
joining the alternating fatigue strength to the
tensile strength is the modified Goodman law.
Goodman's original law included the assumption that the fatigue limit was equal to one third
of the tensile strength; this has since been generalized to the relation shown in Fig. 13, using the
fatigue strength as determined experimentally.
Stress Amplitude. Because stress amplitude
varies widely under actual loading conditions, it
is necessary to predict fatigue life under various
stress amplitudes. The most widely used method
of estimating fatigue under complex loading is
provided by the linear damage law. This is a hypothesis first suggested by Palmgren and restated by Miner, and is sometimes known as
Miner's rule.
The assumption is made that the application
of n.cycles at a stress amplitude S;, for which the
average number of cycles to failure is N;, causes
an amount of fatigue damage that is measured by
the cumulative cycles ratio n;/N;, and that failure will occur when "'i.(n;/ N;) = 1.
This method is not applicable in all cases, and
numerous alternative theories of cumulative linear damage have been suggested. Some considerations of redistribution of stresses have been
clarified, but there is as yet no satisfactory approach for all situations.
Fatigue strength, S
CI)"
vi
~
~
OJ
C
Gerber's parabola
Modified Goodman line

Tensile
strength, Su
Mean stress, Sm
As shown by the modified Goodman line. Gerber's parabola. and Soderberg line. See text for discussion.
Fig. 13 Effect of mean stress on the alternating stress amplitude
The effect of varying the stress amplitude (linear damage) can be evaluated experimentally by
means of a test in which a given number of stress
cycles are applied to a test piece at one stress amplitude. The test is then continued to fracture at a
different amplitude. Alternatively, the stress can
be changed from one stress amplitude to another
at regular intervals; such tests are known as
block, or interval, tests. These tests do not simulate service conditions, but may serve a useful
purpose for assessing the linear damage law and
indicating its limitations.
12
Fatigue Testing
Corrosion Fatigue
Stress-intensity factor range UK). ksi\.
20
10
Corrosion fatigue is the combined action of

repeated or fluctuating stress and a corrosive environment to produce progressive cracking. Usually, environmental effects are deleterious to fatigue life, producing cracks in fewer cycles than
would be required in a more inert environment.
Once fatigue cracks have formed, the corrosive
aspect also may accelerate the rate of crack
growth.
In corrosion fatigue, the magnitude of cyclic
stress and the number of times it is applied are
not the only critical loading parameters. Timedependent environmental effects also are of
prime importance. When failure occurs by corrosion fatigue, stress-cycle frequency, stresswave shape, and stress ratio all affect the cracking processes.
100
q
I
I
a
~
6~
t----
da
dN
~ C(~Kln
I~
dO
: W
'0
Ie?
if
sx;
I
5
~~o
I
I
I
'I
.9' r;~
i
Region 1:
slow crack
I- growth
Fatigue Crack
Propagation
fil.
50
Region 3:
rapid
unstable
crack
growlh
I
rI
r--
I
I
1
Region 2: power-taw behavior
I,c
1;1
I,j>
I
Fatigue failure of structural and equipment

components due to cyclic loading has long been a
major design problem and the subject ofnumerous investigations. Although considerable fatigue data are available; the majority has been
concerned with the nominal stress required to
cause failure in a given number of cyclesnamely, S-N curves. Usually, such data are obtained by testing smooth or notched specimens.
With this type of testing, however, it is difficult to
distinguish between fatigue crack initiation life
and fatigue crack propagation life.
Preexisting flaws or crack-like defects within a
material reduce or may eliminate the crack initiation portion of the fatigue life of the component.
Fracture mechanics methodology enhances the
understanding of the initiation and propagation
of fatigue cracks and assists in solving the problem of designing to prevent fatigue failures.
Fatigue Crack
Propagation Test Methods
The general nature of fatigue crack propagation using fracture mechanics techniques is
summarized in Fig. 14. A logarithmic plot of the
crack growth per cycle, da/ dN, versus the stressintensity factor range, I:!..K, corresponding to the
load cycle applied to a specimen is illustrated.
The da/ dN versus I:!..K plot was constructed of
20
10
30
40
50 60
Suess-tntensnv teeter range UK), MPa \
80 100
Yield strength of 470 MPa (70 ksi). Test conditions: R =

0.10; ambient room air, 24C (75 OF).
Fig. 14 Fatigue crack growth behavior

of ASTM A533 B1 steel
data on five specimens of ASTM A533 HI steel

tested at 24 0 C (75 0 F). A plot of similar shape is
anticipated with most structural alloys; the absolute values of da/dNand I:!..K, however, are dependent on the material.
Results of fatigue crack growth rate tests for
nearly all metallic structural materials have
shown that the da/ dN versus I:!..K curves have
three distinct regions. The behavior in Region I
(Fig. 14) exhibits a fatigue crack growth threshold, I:!..K"" which corresponds to the stressintensity factor range below which cracks do not
propagate.
At intermediate values of I:!..K (Region II in
Fig. 14), a straight line usually is obtained on a
log-log plot of I:!..K versus da/ dN. This is described by the power-law relationship:
da
dN
= C(I:!..K)"
where C and n are constants for a given material

and stress ratio.
13

Fatigue crack growth rate data for some steels
show that the primary parameter affecting
growth rate in Region II is the stress-intensity
factor range and that the mechanical and metallurgical properties of these steels have negligible
effects on the fatigue crack growth rate in a
room-temperature air environment. Data for
four martensitic steels fall within a single band,
as shown in Fig: 15. The upper bound of scatter
can be obtained from:
da
dN = 0.66 X
1O-8(~K)2.25
where a is given in inches, and ~K is given in

ksiyTr;.
For some steels, the stress ratio and mean
stress have negligible effects on the rate of crack
growth in Region II. Also, the frequency of cyclic
loading and the waveform (sinusoidal, triangular, square, trapezoidal) do not affect the rate of
crack propagation per cycle of load for some
steels in benign environments.
At high ~Kvalues (Region III in Fig. 14), unstable behavior occurs, resulting in a rapid increase in the crack growth rate just prior to complete failure of the specimens. There are two
possible causes of this behavior. First, the increasing crack length during constant load testing causes the peak stress intensity to reach the
fracture toughness, K'n of the material, and the
unstable behavior is related to the early stages of
brittle fracture. Second, the growing crack reduces the uncracked area of the specimen sufficiently for the peak load to cause fully plastic
limit load behavior. The first possibility is operative for high-strength, low-toughness metals, in
which specimen sizes normally used for fatigue
crack growth rate testing behave in a linear elastic manner at K levels equal to K/c. The second
possibility, plastic limit load behavior, is common for ductile metals, particularly if K/cis high.
When plastic limit load behavior causes unstable crack growth, ~K values have no meaning, because the limitations of linear elastic fracture mechanics have been exceeded. Here, the
use of the J-integral concept, crack-opening displacement, or some other elastic-plastic fracture
mechanics approach is more appropriate than
~K for correlating the data.
Standardized testing procedures for measuring fatigue crack growth rates are described in
ASTM Standard E 647. This method applies to
medium to high crack growth rates-that is,
above 10-8 tu] cycle (3.9 X 10-7 ui.] cycle). Procedures for growth rates below 10-8 in] cycle are
Stress-intensity factor range (.:lK), ksiV'Ti1."
102
10
10- 2
o
o
Ql
"0
>
E
E
12 Ni steel
10 Ni steel
HY-80 steel
HY-130 steel
Ql
10- 4 "0
>
~
c::
10- 3
......:
<:
<:
~
10- 5
2:!
s:
s:
10- 4
Ol
Ol
-"
0
-"
0
10- 6
~
U
co
L.
10- 5 t--t-l~tf--+--+--+----I
2 x 10- 6
'---_..I..-_..L-_..I..-_....L_--L--'
10
20
50
100
200
Stress-intensity factor range (.:lKl. MPaVm

Fig. 15 Summary of fatigue crack
growth data for martensitic steels
under consideration by ASTM. For applications

involving fatigue lives of up to about 106 load
cycles, the procedures recommended in ASTM E
647 can be used. Fatigue lives greater than about
106 cycles correspond to growth rates below 10-8
in] cycle, and these require special testing procedures, which are related to the threshold of fatigue crack growth illustrated in Fig. 14.
ASTM E 647 describes the use of centercracked specimens and compact specimens (Fig.
16 and 17). The specimen thickness-to-width ratio, B/W, is smaller than the 0.5 value for K/c
tests; the maximum B/W values for centercracked and compact specimens are 0.125 and
0.25, respectively. With the thinner specimens,
crack length measurements on the sides of the
specimens can be used as representations of
through-thickness crack growth behavior.
For tension-tension fatigue loading, the K;
loading fixtures frequently can be used. For this
type of loading, both the maximum and minimum loads are tensile, and the load ratio, R =
P min/Pmax' is in the range 0 < R < 1. A ratio of
14
Fatigue Testing
Two holes W/3 diam
-r
W
~----+-----J_l
!I !
I
!I
I
2a n i.sthe machined notch; a is the crack length; B is the
specimen thickness.
Fig. 16 Standard center-cracked tension specimen for fatigue crack propagation testing when the width (WI of the
specimen ';;;;75 mm (3 in.)
R= 0.1 is commonly used for developing data for

comparative purposes.
Testing often is performed in laboratory air at
room temperature; however, any gaseous or liquid environment and temperature of interest
may be used to determine the effect of temperature, corrosion, or other chemical reaction on
cyclic loading.
Data Analysis. For constant-amplitude loading, a set of crack-length versus elapsed-cycle
data (a versus N) is generated, with the specimen
loading, Pmax and P min' generally held constant.
Figure 18illustrates a typical a versus N plot. The
minimum crack-length interval, 6.a, between
data points (see Fig. 18)should be 0.25 mm (0.01
in.) or ten times the crack-length measurement
precision, which is defined as the standard deviation on the mean value of crack length determined for a set of replicate measurements. This
prevents the measurement of erroneous growth
rates from a group of data points that are spaced
too closely relative to the precision of data measurement and relative to the scatter of data.
Crack measurement intervals are recommended in ASTM E 647 according to specimen
type. For compact-type specimens:
6.a ~ 0.04 Wfor 0.25 ~!!.- ~ 0.40
6.a ~ 0.02 W for 0.40 ~ !!.- ~ 0.60

W
6.a~
0.01 Wfor -;;::: 0.60

W
For center-cracked tension specimens:

6.a ~ 0.03 Wfor 2a
W
6.a ~ 0.02 W for
< 0.60
~ > 0.60
W
Fatigue crack growth rate data can be calculated by several methods. The most commonly
used methods, however, are the secant and incremental polynomial methods. The secant method consists of the slope of the straight line connecting two adjacent data points. This method,
although simpler, results in more scatter in measured crack growth rate.
The incremental polynomial method fits a
secon~-order ,Polynomial expression (parabola)
to typically five to seven adjacent data points,
and the slope of this expression is the growth
rate. The incremental polynomial method eliminates some of the scatter in growth rate that is
inherent in fatigue testing.
Numerous relationships have been generated
to correlate crack growth rate and stress-intensity
data. The most widely accepted relationship is
that proposed by Paris. This is a linear relationship when plotted on log-log coordinates and
generally yields a reasonable fit to the data in
Region II (see Fig. 14) of the crack growth
regime.
Other relationships based on the Paris equation, such as the commonly used Forman equatl?n, are used to represent the variation of da/ dN
with other key variables, including load ratio, R
and the critical K value, K" at which rapid frac~
ture of the specimen occurs (Region III in Fig.
14). The Forman equation is:
da =
C(6.K)"
dN (1 - R)(K,. - 6.K)
where Cand n are material constants of the same

types as those in the Paris equation, but of different values. An advantage of the Forman equation is that it describes the type of accelerated
da/dNbehavior that is often observed at high
values of 6.K, which is not described by the Paris
equation.
Additionally, the Forman equation describes
the frequently observed increase in da/ dN asso-
15

Two holes
0.25Wdiam
t
t
0.6W
0.275W
t
t
0.275W
0.6W
~~_1_
1 + - - - a ----;~
( - + - - - - - - w-------;~I
)
1
Allowable thickness: W/20 s B s W/4

Minimum dimensions: W = 25 mm (1.0 in.)
an
= 0.20W
~------1.25W
-------:;~I
Fig. 17 Standard compact-type specimen for fatigue crack propagation testing (see Fig.
16 for explanation of symbols)
ciated with an increase in R from 0 toward I.

When it is necessary to describe the effect of K
approaching Kc> or the effect of R on da / dN, the
Forman equation can be used to represent the
da/dNbehavior. When only ~Kin Region II is
involved, the less complex Paris equation may be
used.
Cyclic Crack Growth Rate

Testing in the Threshold Regime
55
2.2
(
50
E
E 45
.r=
40
~ 35
u
1.6
(J
00
20
125
Oo c
150
1.0
175
200
225 250
<iI
.><
1.2
30
25
s:
0,
1.4 Q)c
'rP0a- J
.><
1.8
",,0
(J
Cyclic crack growth rate testing in the lowgrowth regime (Region I in Fig. 14) complicates
acquisition of valid and consistent data, because
the crack growth behavior becomes more sensitive to the material, environment, and testing
procedures under this regime. Within this regime, the fatigue mechanisms of the material that
slow the crack growth rates are more significant.
The precise definition of the cyclic crack
growth rate threshold, ~K'h' varies significantly.
i~'
OJ
2.0
275
300
Cycles. 10 N
Fig. 18 Crack growth versus constantamplitude stress cycles for a Fe-10Ni8Co-1 Mo high-strength steel
...
ro
U
16
Fatigue Testing
The most accurate definition would be the stressintensity value below which fatigue crack growth
will not occur. It is extremely expensive to obtain
a true definition of IJ.K,h, and in some materials a
true threshold may be nonexistent. Generally,
designers are more interested in the near-threshold regime, such as the IJ.Kthat corres~onds to a
fatigue crack growth rate of 10-8 to 10- 0 tn] cycle
(3.9 X 10-7 to 10-9 in.j cycle). Because the duration of the tests increases greatly for each additional decade of near-threshold data (10- 8 to 10-9
to 10-10, etc., m/cycle), the precise design requirements should be determined in advance of
the test.
Short cracks
that join long
crack behavior
q/;/"
Behavior of Short Cracks

Recently, it has been well documented that
short cracks may behave differently from large
cracks when plotted in the standard form of cyclic crack growth rate versus stress intensity.
A short crack is difficult to define. It may be
small compared to the microstructure of the
material to be studied (I to 50 /.Lm) when the concepts of continuum mechanics are of interest. It
can also be small compared to the plastic zone
size (10 to 1000 /.Lm). In this situation, linear elastic fracture mechanics might be replaced with
elastic-plastic fracture mechanics. The crack
may also be physically small (500 to 1000 /.Lm)
when crack closure, crack tip shape, environment, and growth mechanisms are of concern.
Figure 19 schematically illustrates the possible
behavior of short cracks.
Selection of Test Specimens

Selection of a fatigue crack growth test specimen is usually based on the availability of the
material and the types of test systems and crackmonitoring devices to be used. The two most
widely used types of specimens are the centercracked tension specimen and the compact-type
specimen (see Fig. 16 and 17). However, any
specimen configuration with a known stressintensity factor solution can be used in fatigue
crack growth testing, assuming that the appropriate equipment is available for controlling the
test and measuring the crack dimensions. Stressintensity factor solutions for center-cracked tension and compact-type specimens are given in
Table 2.
Consideration of the range of application of
the stress-intensity solution of a specimen configuration is very important. Many stress-inten-
//
\',
Short cracks that behave

as long cracks
./
\---
Short cracks that become

non propagating cracks
Stress-intensity factor range

Fig. 19 Typical short crack behavior
sity expressions are valid only over a range of the

ratio of crack length to specimen width (a/W).
For example, the expression given in Table 2 for
the compact-type specimen is valid for a/ W >
0.2; the expression for the center-cracked tension
specimen is valid for 2a/W< 0.95. The use of
stress-intensity expressions outside their applicable crack-length region can produce significant errors in data.
The size of the specimen must also be appropriate. To follow the rules of linear elastic fracture mechanics, the specimen must be predominantly elastic. However, unlike the requirements
for plane-strain fracture toughness testing, the
stresses at the crack tip do not have to be maintained in a plane-strain state. The stress state is
considered to be a controlled test variable. The
material characteristics, specimen size, crack
length, and applied load will dictate whether the
specimen is predominantly elastic. Because the
loading mode of different specimens varies significantly, each specimen geometry must be considered separately.
Notch Preparation. The method by which a
notch is machined depends on the specimen

Table 2 Stress-intensity factor solutions for standardized (ASTM E 647)
fatigue crack growth specimen geometries
Center-cracked tension specimens (Fig. 16)
sr
I:!.K=
Ii V
2W
sec
tra
2a; expression
. va1'"
2a 0.95
where a = id 10r-<
Compact-type specimens (Fig. 17)

_
I:!.K -
I:!.P(2
+ a)
ru;
3/2 (0.886 + 4.64a - 13.32a
By W(l-a)
+ 14.Tl 3 -
5.6a )
. va liId f or-
a 0.2
where a = -a ; expression
material and the desired notch root radius (p).

Sawcutting is the easiest method, but is generally
acceptable only for aluminum alloys. For a
notch root radius of p ~ 0.25 mm (0.010 in.) in
aluminum alloys, milling or broaching is required. A similar notch root radius in low- and
medium-strength steels can be produced by
grinding. For high-strength steel alloys, nickelbase superalloys, and titanium alloys, electrical
discharge machining may be necessary to produce a notch root radius of p ~ 0.25 mm (0.010
in.).
Precracking of a specimen prior to testing is
conducted at stress intensities sufficient to cause
a crack to initiate from the starter notch and
propagate to a length that will eliminate the effect of the notch. To decrease the amount of time
needed for precracking to occur, common practice is to initiate the pre cracking at a load above
that which will be used during testing and to subsequently reduce the load.
Load generally is reduced uniformly to avoid
transient effects. Crack growth can be arrested
above the threshold stress-intensity value due to
formation of the increased plastic zone ahead of
the tip of the advancing crack. Therefore, the
step size of the load during precracking should
be minimized. Reduction in the maximum load
should not be greater than 20% of the previous
load condition. As the crack approaches the final
desired size, this percentage may be decreased.
Gripping of the specimen must be done in a
manner that does not violate the stress-intensity
solution requirements. For example, in a singleedge notched specimen, it is possible to produce
a grip that permits rotation in the loading of the
specimen, or it is possible to produce a rigid grip.
Each of these requires a different stress-intensity

solution. In grips that are permitted to rotate,
such as the compact-type specimen grip, the pin
and the hole clearances must be designed to minimize friction. It is also advisable to consider lateral movement above and below the grips.
Gripping arrangements for compact-type and
center-cracked tension specimens are described
in ASTM E 647. For a center-cracked tension
specimen less than 75 mm (3 in.) in width, a single
pin grip is generally suitable. Wider specimens
generally require additional pins, friction gripping, or some other method to provide sufficient
strength in the specimen and grip to prohibit
failure at undesirable locations, such as in the
grips. Grips designed for compact-type specimens are illustrated in Fig. 20.
Crack-Length
Measurement Techniques
Several different techniques have been developed to monitor the initiation, growth, and instability of cracks, including optical (visual and
photographic), electrical (eddy current and resistance), compliance, ultrasonic, and acoustic
emission monitoring techniques.
Optical Crack Measurement Techniques
Monitoring of fatigue crack length as a function of cycles is most commonly conducted visually by observing the crack at the specimen surfaces with a traveling low-power microscope at a
magnification of 20 to 50X. Crack-length measurements are made at intervals such that a
nearly even distribution of da] dN versus ~K is
achieved. The minimum amount of extension be-
17
18
Fatigue Testing
p
t
(a)
~--al~
p
Fig. 20 Grips designed for fatigue crack

propagation testing of compact-type
specimens (courtesy of MTS Systems
Corp.)
tween readings is commonly about 0.25 mm

(0.10 in.).
The optical technique is straightforward and,
if the specimen is carefully polished and does not
oxidize during the test, produces accurate results. However, the process is time consuming,
subjective, and can be automated only with
complicated and expensive video-digitizing
equipment. In addition, many fatigue crack
growth rate tests are conducted in simulatedservice environments that obscure direct observation of the crack.
Compliance Method of
Crack Extension Measurement
The compliance of an elastically strained specimen containing a crack of length a measured
from the load line to the crack tip is usually expressed as the quotient of the displacement, 0,
and the tensile load, P, with the displacement
measured along, or parallel to, the load line. Fig-
(b)
(a) C(aD)
= 0D/P, (b) C(a1) = 01 /P
Fig. 21 Schematic of the relationship

between compliance and crack length
ure 21 illustrates that the more deeply a specimen

is cracked, the greater the amount of 0 measured
for a specific value of tensile load. Compliance
can also be defined for shear and torsional loads
applied to cracked specimens, and crack extension under these loading modes can be similarly
determined.
Specimen load is simultaneously measured by
an electronic load cell and conditioner / amplifier
system, and the output is directed to the same
data-acquisition system. A generalized schematic
of the circuits involved is shown in Fig. 22.
19
Displacement
gage
Load
a, a 1 a2
-+----... )
....
'Lr
x-v recorder
Specimen
Gage
condition
/P'~
J=1
Load
cell
I
10 V dc
Load cell
condition
10 V dc
Fig. 22
Components of a compliance measurement system
The required sensitivity of the systems depends on specimen geometry and size; in general,
noise-free, amplified output on the order of I V
dc per I mm (0.04 in.) of deflection is satisfactory. Similarly, for the load range applied to the
specimen, an approximately I V de change in
signal from the load cell is required for accurate
calculation of the compliance.
Electric Potential
Crack Monitoring Technique
The electrical potential, or potential drop,
technique has gained increasingly wide acceptance in fracture research as one of the most accurate and efficient methods for monitoring the
initiation and propagation of cracks. This method relies on the fact that there will be a disturbance in the electrical potential field about any
discontinuity in a current-carrying body, the
magnitude of the disturbance depending directly
on the size and shape of the discontinuity.
For the application of crack growth monitoring, the electric potential method entails passing
a constant current (maintained constant by external means) through a cracked test specimen
and measuring the change in electrical potential
across the crack as it propagates. With increasing
crack length, the uncracked cross-sectional area
of the test piece decreases, its electrical resistance
increases, and thus the potential difference between two points spanning the crack rises. By
monitoring this potential increase, Va' and comparing it with some reference potential, Vo , the
crack length to width ratio, a/ W, can be determined through the use of the relevant calibration
curve for the particular test piece geometry
concerned,
Crack Growth Studies. By far the most useful
application of the electrical potential method has
been in measurements of crack length during
crack propagation, where it has been utilized to
monitor almost all mechanisms of subcritical
crack growth and most notably to follow fatigue
crack growth. Typical crack propagation rates
derived from direct current potential measurements are shown in Fig. 23 for tests on a 2.25CrIMo steel in air, gaseous hydrogen, and hydrogen sulfide environments.
Electromechanical
Fatigue Testing Systems
The primary function of electromechanical fatigue testers is to apply millions of cycles to a test
piece at oscillating loads up to 220kN (50000 lbf)
to investigate fatigue life, or the number of cycles
to failure under controlled cyclic loading conditions. Variables associated with fatigue-life tests
are frequency of loading and unloading, amplitude of loading (maximum loads and minimum
loads), and control capabilities. The fundamental data output requirement is the number of cycles to failure, as defined by the application,
A variety of electromechanical fatigue testers
have been developed for different applications.
Forced-displacement, forced-vibration, rota-
20
Fatigue Testing
Table 3
Comparison of electromechanical fatigue systems
Parameter
Forced displacement
Tension
Compression
Reverse stress
Bending
Frequency range
Load range
Forced vibration
.
Yes
Yes
.
.
Yes
.
Yes
.
Fixed
. Typically < 450 N
1001bf)
Type:
Control
Mode
Open-loop
Displacement
.
.
Maximum deflection
Advantages
Yes
Yes
Yes
Yes
Fixed, 1800 rpm
Up to 220 kN
(500001bf)
Open-loop
Load
25.4 mm (l.00 in.

Versatile, efficient,
durable
Fixed frequency, Iii
ited control (open
loop)
Simple, straightforward
No load control, very
limited applications
(soft samples)
Disadvantages
Alternating stress intensity (:>K), ksi vln,

678910
'"
0
l>
Environment
Moist air
Dry hydrogen
Air
Dry hydrogen
Hydrogen sulfide
Molstair
Dry hydrogen
::
I~rt
V~
I 'I II ,
20
Frequency. Hz
50
50
5
5
5
0.05
0.05
0.1
0.1
0.1
0.75
0.75
40
~#
nO
#
00 '"
fill'
t4,f-
~
lJ
.. ..
60 10 80
I
10" ,
10-.
.
..
1)
10- 1
s:
'U
One lattice -..

spacing
-- 10-.
per cycle
0
0
>
u
.~
- 10'
oOd9
~.'l>~ ~
rP
rPJ pO
od
!II
50
~I
I"'"
'"
i-
30
10
ern
..
Threshold :>K",
t
6
8 9 10
20
30
40
50 60 10 80 90
Alternating stress intensity (:>K), MPa\m

Data derived from direct current potential measurements in martensitic 2.25Cr-1 Mo steel (SA542-C12) at R =0.05
to 0.75 in air, hydrogen. and hydrogen sulfide at ambient temperature.
Fig. 23
Fatigue crack propagation data over a wide spectrum of growth rates
tional bending, resonance, and servomechanical

systems are discussed in this article and are compared in Table 3. Other specialized electromechanical systems are available to perform specific tasks.
Forced-Displacement Systems
Forced-displacement motor-driven systems
are the simplest type of electromechanical fatigue testers. They effectively reproduce service
environments that impart fixed, reciprocating

displacements to a component or test piece. An
electric motor-driven flywheel is used to carry a
loading arm at a variable distance from the center of rotation, much in the same manner as a
connecting rod in an automotive engine. This rotational displacement is transformed into a
guided, vertical displacement and is used to fatigue the specimen.
Although load can be monitored in such sys-
21
Rotational bending
Resonance
Servomechanical
No
No
Yes
Yes
0-10000 rpm
Yes
Yes
Yes
Yes
40-300 Hz
Up to 180 kN
(400001bf)
Yes
Yes
Yes
Yes
0-1 Hz
Up to 90 kN
(200001bf)
Open-loop
Rotation/ bending
Closed-loop
Load
fficient, durable,
simple
.otational bending
only, limited applications
l.0 mm (0.040 in.)

Fully closed-loop,
extremely efficient
Operating frequency
directly proportional
to sample stiffness
Closed-loop
Load, displacement,
strain
100 mm (4 in.)
Fully closed-loop,
high precision
Low frequency only
tems, the fixed displacement precludes the ability

to control load, which is a function of specimen
characteristics. Therefore, the load generally
drops as failure progresses. These systems typically are custom-built, inexpensive fatigue machines, used primarily for bend tests on soft samples in which load control, high frequencies, and
large loads are not required.
Forced-Vibration Systems
Forced-vibration motor-driven systems were
the first production fatigue testers in commercial
use. The centrifugal forces of an imbalanced rotor is used to impart a cyclic load to the test piece.
In operation, an electric motor is used to rotate an eccentric mass via flexible couplings. The
rotating mass is mounted in a frame that is
guided by flexure plates to restrict movement to
vertical motion only. The centrifugal force produced by the rotating eccentric mass (m) is
transmitted through the vertically guided frame
to the test piece. The horizontal component of
the centrifugal force is absorbed by the restraining flexure plates.
Because the centrifugal force usually is totally
absorbed by the mounting frame (of mass M),
the inertial reaction is separated from the centrifugal force in such a way as to transmit only
the centrifugal forces to the specimen. This technique involves the use of frame-support compensator springs; the natural frequency of the spring
(K)/mass (M) system is tuned to the revolutions
per minute of the motor. Thus, neither the specimen nor the rotating eccentric mass (m) "sees"
an inertial reaction from the frame, because the

inertial effects of the frame are totally compensated for by the frame support springs (not the
specimen).
This technique has two requirements: the rotating frequency (w) must be kept constant and
the mass of mounting frame (M) must be kept
constant. Consequently, the loading frequency
of the device is fixed at 1800 rpm, and masses
must be added or removed from the frame to
compensate for fixturing to keep M constant.
The magnitude of the dynamic load is determined by placing the rotating mass at a known
distance from the axis ofrotation (r). Because w,
m, M, and K are known, the force on the specimen, F, is calibrated directly as a function of r as
follows:
F
= Mw 2r (centrifugal) -
Ma, (inertial)
+ Kz (spring compensated)
where a, is the acceleration of the frame in the z

direction, and Kz is the spring-compensated displacement in the z direction. Because Ma, is
tuned to equal Kz, F = Mw 2r.
Thus, the forced-vibration rotating eccentric
mass system is an open-loop, load-controlled
system with the ability to accommodate up to 25
mm (1.0 in.) of total sample deflection at loads
up to 220 kN (50000 lbf) using special fixtures.
The mean or static load, onto which the dynamic
load is superimposed, is achieved by preloading
the inertia compensator spring, K.
Through special fixturing, forced-vibration
devices are capable of testing in tension, com-
22
Fatigue Testing
pression, bending, torsion, or reverse st~ess.
Although servo-controlled, mean-load-maintenance systems are available, the open-loop nature of the system prevents direct load measurement or control, which is characteristic of
closed-loop systems. The load applied to the
specimen is assumed to be a function of r, and a
graduated scale is provided to permit reasonably
accurate setup.
Rotational Bending Systems
Rotational bending systems effectively apply
reversed loading to the outer surface of rods or
shafts. The basic operating principle ofthe rotating beam consists of the use of a motor to rotate a
shaft of known dimensions around its longitudinal axis. By applying a known static force at the
end of the shaft, a bending moment can be applied to the test section, the outer surface of
which oscillates between tension and compression during each rotation.
The cantilevered specimen, however, is subjected to a nonuniform bending moment, which
is large at the supported end of the specimen ~nd
zero at the free end. To produce a more meanmgful uniform bending moment throughout the
tes~ piece, a specially designed tapered specimen
should be used or bending moments should be
applied to each end of the specimen. Figure 24
illustrates the rotating-beam operating mechanism and the resulting stress distribution in the
specimen.
Gage
area
Drive motor
I Tension
'----.. )
Hotation
0~:f:
Bending
mom,",
Compression
Fig.24 Schematic ofthe rotating-beam
operating mechanism and the resulting
stress distribution in the specimen
Resonance Systems
A high-speed fatigue testing system was developed by Amsler that operated at 40 to 300 Hz,
achieved high loads (up to 90 kN, or 20000 lbf),
and consumed minimal energy. It is based on a
resonant spring/mass system, in which the specimen is used, like a spring, as an integral part of
the oscillating mechanism.
The fatigue load, in the form of a sine wave, is
achieved by preloading the sample in the frame
via a complex optomechanical procedure and
dynamically loading the sample at the natural
oscillating frequency of the spring/ mass system.
The preload is maintained automatically during
the test. The dynamic load is achieved by pulsing
an electromagnet at the natural frequency of the
spring/ mass system. During resonance, the electromagnet restores any hysteresis energy lost
during the previous cycle, thereby maintaining a
constant, controllable dynamic load. Capable of
tension, compression, bending, torsional, and
reverse-stress fatigue tests, the Amsler resonant
fatigue testers were instrumental in obtaining the
vast amount of fatigue data currently available.
The resonant system is based on a similar principle, but incorporates solid-state technology to
achieve fully closed-loop control of mean and
dynamic loads. This system uses dual opposing
masses (unlike the single oscillating mass/ seismic
base of earlier systems), linked by the specimen
to achieve vibration-free resonance. A strain
gage load cell, in series with the specimen, senses
the load and automatically triggers the electromagnet to achieve self-tuning capability.
The mean load is achieved by physically moving the upper mass up or down to achieve tension
or compression, respectively; the dynamic load is
achieved by varying the width of the pulse to the
magnet beneath the lower mass. The dynamic
load, like the mean load, is electronically ma~n
tained at a preset command level through solidstate closed-loop circuitry. The remainder of the
controls and mechanisms associated with the resonant fatigue system maintain a preset air gap
between the magnet and the oscillating lower
mass, maintain preset loading conditions (S~1Ut
ting down at preset load levels or frequencies),
and power the electromagnet.
The high efficiency of resonant systems makes
them well suited to high-cycle fatigue tests, in
which closed-loop load control, high loads (up to
180 kN, or 40000 lbf), low power consumption
(around 750 W maximum for closed-loop sys-
23

Electronic
demand
signal
1-----,---- c
(bl
(a) Typical components. (b) Transfer functions. See text for details and explanation of symbols.
Fig. 25
Simplified block diagram for a negative-feedback closed-loop testing machine
tems), and high throughput are required. These

systems tolerate minimal hysteresis and produce
optimum testing results when used with stiff metallic samples.
Closed-Loop Servomechanical Systems
The most recent development in electromechanical fatigue testers is based on an electric actuator/load frame assembly. The system closely
resembles its servohydraulic counterpart in that
it consists of an actuator, a load frame, a load
cell, a power supply, and a solid-state closedloop electronic control console. Closed-loop systems compare live feedback signals to an input
command signal to maintain accurate control of
preset conditions. The closed-loop servomechanical system is, by virtue of its design, primarily
intended for low-cycle and creep-fatigue studies.
Servohydraulic Fatigue
Testing Systems
Servohydraulic testing machines are particularly well suited for providing the control capabilities required for fatigue testing. Extreme demands for sensitivity, resolution, stability, and
reliability are imposed by fatigue evaluations.
Displacements may have to be controlled (often
for many days) to within a few microns, and
forces can range from 100 kN to just a few newtons. This wide range of performance can be obtained with servomechanisms in general and, in
particular, with the modular concept of servohydraulic systems.
Usually, the problem of selecting the appropriate system is simply a matter of optimizing the
various components to form a system best suited
to the given testing application. In this section,
the principles underlying closed-loop servo systems are discussed briefly. In addition, the interaction between system components is illustrated,
and a brief description of their operating principles and characteristics is provided.
With any type of control system, the objective
is to obtain an output that relates as closely as
possible to the programmed input. In a fatigue
testing system, it may be desired to vary the force
on a specimen in a sinusoidal manner, at a frequency of 1 Hz over a force range of 0 to 100 kN
(0 to 22000 lbf). The only practical means to accomplish this with precision is through the use of
a negative-feedback closed-loop system. An overview of the basic principles of operation of
negative-feedback systems is provided in Fig. 25.
The blocks shown in Fig. 25(a) represent a group
oftypical components of a testing machine. The
transfer functions of each of these blocks can be
combined to produce the more simplified diagram shown in Fig. 25(b).
Placement of the switch, S I' has been added to
the diagram to permit analysis of the system
when it is open (no feedback, or an open-loop
condition) and when it is closed (providing feedback to the system). The equation governing this
simplified open-loop system is:
C= KoD
where Crepresents the controlled output, K; rep-
24
Fatigue Testing
resents the open-loop transfer function, and D
represents the electronic demand signal. Therefore, the output is simply proportional to the system demand if K is a constant. Unfortunately,
K is seldom a constant, because it can be influen"ced by several common system variations. The
electronic components may drift slightly, or the~r
gain may vary. The behavior of the hydraulic
components may change with tempera~ure, contamination, or wear, and the mechamcal components may vary because of thermal effects or
friction.
Servohydraulic System Components
Many commercially manufactured units are
available for each component in a typical servohydraulic testing system.
.
The programmer supplies the command signal
to the system, which is generally an analog of the
desired behavior of the controlled parameter.
For example, assume the same test conditions as
previously discussed (control the force on the
specimen in a sinusoidal manner at a frequency
of I Hz and a force range of 0 to 100 kN). In this
instance, the programmer might be set to produce an electronic signal with a sinusoidal waveform that has a frequency of I Hz and a voltage
output of 0 to 10 V. The analog is: 1 V represents
1000 N. The system can then be adjusted to produce the correct output. Any change in the programmer signal will result in a corresponding
change in the controlled parameter.
The servo-controller makes most of the adjustments necessary to optimize system performance. For example, it compares the command
signal with a signal produced by the controlled
parameter (stress or strain, for example) and relays a correction signal, if needed, to the control
device in the system (usually a flow-control
servo-valve). A servo-controller incorporates
numerous other compensatory features, such as:
Means to adjust the gain or proportional

band of the system
Controls to modify the feedback or correction signals for improved stability
Controls to adjust the mean level and amplitude of the command signal(s)
Controls to enhance and adjust servo-valve
response
.
Means to monitor the system error signal (a
measure of how well the command and
feedback signals agree)
Capability to select various command and
feedback signals
Auxiliary functions such as recorder signal

conditioning, calibration, and system startup and shutdown
The servo-valve controls the volume and direction of flow of hydraulic fluid between the hydraulic power supply and the hydrauI~c ram.
Within the control loop, it is the intermediary between the low-power servo-controller and the
hydraulic ram, which can supply large force~ a~d
displacements to the specimen. Characte~ISt1CS
of the device are such that the output flow IS approximately proportional to the input current
when the output pressure is constant. Also, the
output pressure is approximately proportlOn~1
to the square of the input current when the flow IS
constant.
Hydraulic rams, or actuators or cylinders,
furnish the forces and displacements required by
the testing system. These rams usually are double
ended to provide the greatest lateral rigidity and
to produce the balanced flow and f~rce characteristics desirable for push-pull testmg. The effective area of the piston is therefore equal to the
cross-sectional area of the piston minus the
cross-sectional area of the piston rod. Under
static conditions (very little flow), the maximum
force capability of the ram will approach the hydraulic supply pressures multiplied by the effective area.
The force available during dynamic operation
depends on the pressure drop and flow characteristics of the servo-valve. Reference should be
made to the load/flow/pressure characteristics
supplied by the servo-valve manufacturer.
Load Cells. The strain gage load cell is the
most widely used force-measuring and feedback
device in closed-loop fatigue machines. An external applied force causes the elastic deformation of an internal member to which a strain gage
bridge has been attached. An electronic sign~l
that is proportional to the resistance change m
the bridge and to the applied force can thus .b.e
produced. Some load cells are designed specifically for fatigue evaluations. Variable features
include sensitivity, natural resonant frequency,
temperature stability, fatigue rating, linearity,
hysteresis, deflection constant, load capacity,
overload rating, resistance to extraneous loading, and compatibility with t?e testing machine
and fixtures. Most commercially available cells
are very competitive with respect to these
features.
Load Frames. In a fatigue machine, the reaction forces to the specimen and to the housing of

the ram are supplied by the load frame. Many
styles ofload frames are available, but for fatigue
purposes the frames should be customized. The
requirements of good high-frequency response
demand that there be high axial stiffness in the
load frame. When a deflection occurs in the load
frame, additional flow is required from the
servo-valve. Therefore, this deflection should be
minimal in comparison with the deflection imparted to the specimen.
In addition, because fatigue specimens must
be subjected to fully reversed loading (i.e., compressive as well as tensile forces), lateral rigidity
must be increased to resist bending. This is generally considered necessary in the design of fatigue machines. The extra rigidity can be obtained by increasing the diameter of the support
columns or by utilizing three- or four-column
configurations.
Exceptional alignment is required of load
frames used in fatigue evaluations to minimize
undesirable bending forces. In addition, some
means is usually provided to refine the alignment
with manual adjustments when necessary. A
strain-gaged specimen can be used to make this
evaluation.
25
l-1.
S-N Curves Typical
27
for Steel
Schematic
S-.V curves for a material
at various
stress ratios. l!TS and IS indicate ultimate tensile
strength and yield strength, respectiveI),
in uniaxial tensile testing.
The results of fatigue tests are usually plotted as maximum

stress or stress amplitude
to
number of cycles. .V. to fracture using a logarithmic
scale for the number of cycles. Stress is
plotted on either a linear or a logarithmic
scale. The resulting tune of data points is called an
S-,Vcur\e.
A family of S-,Zcurves for a material tested at various stress ratios is shown schematically in the above curves. Stress ratio is the algebraic ratio oftwo specified stress Lalues in a stress
cycle. Twocommonl~
used stress ratiosare the ratio. A.ofthealternatingstressamplitudeto
the
mean stress (A = Sa,Sm) and the ratio. R. of the minimum stress to the maximum stress (R=
S,,,/ S,,,). If the stresses are fully reversed. the stress ratio R becomes -I: ii the stresses are
partially
re\,ersed. R becomes a negatke number less than I. If the stress is cycled between a
maximum stressand no load. the stress ratio R becomes zero. lithestress
isqcled
bewssn two
tensile stresses. the stress ratio R becomes a positive number less than I. .A stress ratio R oi I
indicates no variation
in stress. and the test \\ould becomea sustained-load
creep test rather than
a iatigue test. For carbon and lo\{-allo!
steels. S-.Vcur\es typically halea fairI> straight slanting
portion at low cycles changing into a straight. horizontal
line at higher cycles. with a sharp
transition
between the two.
An S-.Icur!e
usually represents the median life for a given stress-the
life that half the specimens attain. Scatter of fatigue li\es can cover a \ery ibide range.
Source Metals Handbook.%h

Park OH. 1978. p 667
Edann.
Volume I. Proprrr~rrnnd
Srlei~~on. lronsand Sreelr. .Amcrwan So&t!
ior hlctals. hlcr&
28
1-2. S-N Curves Typical for Medium-Strength Steels
100
Fracture region
(all specimens fractured)
80
s:
rn
Fatigue - Iraclure band
lii rz:
CIl
~
70
ViS:
i!
'"
CIl
0;-
:: (;
0_
.,,g
g>:;
c~
~-'
Q;
60
Finile-Iife region
(no specimens fractured)
50
II
..
40
30
0..
Infinile-lile region
Fatigue limit
20
10
O~
10
(1)
_ _-L-_ _-L-_ _- L_ _- ' - - L - _ - L_ _- L_ _- L

10'
(10)
103
(1,000)
10'
(10.000)
10'
10 6
107
(100,000) (1.000.000) (10,000.000)
Number of cycles to Iracture
SN curves that typify fatigue test results for testing of

medium-strength steels.
As an explanation, if the single-load fracture strength of the specimens is considered

to be 100 percent, for purposes of illustration this is the starting place, for the specimens can sustain no higher load without fracture. If ten specimens are fractured, the
results are placed as points at the top of the left axis at one load application.
Intuitively it is known that if the maximum load (or stress) is lowered to 90 percent of
the tensile strength, it will require more than one load application to fracture the specimens. The ten points shown in the diagram at 90 percent represent the possible life to
fracture of each of the ten specimens. Because the scale is logarithmic, the points appear to be relatively close, but in fact the scatter in life from longest to shortest is on the
order of more than 2 to I. At this high stress, plastic deformation of the test specimen is
likely to be great, such as in bending a paper clip or wire coat hanger to make it
fracture. Actual parts are not intentionally designed to operate in this regime, and
normal fatigue fractures have no obvious plastic deformation.
If the load is dropped to 80 percent of the single-load fracture strength and ten more
specimens are tested, they will run longer with a fatigue life scatter of perhaps 3 to I,
which is not unusual, even for theoretically identical specimens (which, of course, they
are not). When the load is dropped to 70 percent, the lives get longer and the scatter in
fatigue life increases to perhaps5 to I. Again dropping the load, now to 60 percent of
the single-load fracture strength, the fatigue lives again increase, as does the scatter
from longest to shortest life. Invariably, in actual fatigue testing, there is at least one
specimen that inexplicably fractures far earlier than any of the others in the same
group. One such specimen is shown at the 60 percent level fracturing at about 150
cycles, while the other supposedly "identical" specimens or parts had lives of from
about 1,000 to 10,000 cycles. The cause of such an early "anomaly" is often sought in
vain, although it is possible that some metallurgical reason, such as a large inclusion on
the surface, might be found. Frequently, this lone early fracture specimen is simply
ignored.
Dropping now to 50 percent of the single-load fracture strength, the fatigue lives
increase dramatically, as the S-N curve starts to flatten out. This flattening out is
characteristic of ferrous metals oflow and moderate hardness; many nonferrous metals and some very high-hardness ferrous metals tend to continue their downward path
at very large numbers of cycles. Now, the problem is when to stop the tests. The test
1-2. S-N Curves Typical for Medium-Strength Steels (continued)

machine will be needed for another test specimen after a very long test time, depending
upon the rate ofloading, or cycles per minute. If ten million is selected as the end point,
the test must be stopped at that figure even if a specimen is unbroken, and the point
shown with an arrow pointing to higher values, for it did not actually fracture. Frequently, five million, or even one million, cycles is selected as the end point, depending
upon the metal, purpose, and urgency of the tests. For example, five hundred million
cycles is sometimes used in the aluminum industry.
The region below the lowest portion of the S-N curve is called the infinite-life region,
because specimens that are tested at stresses below the curve should run indefinitely;
that is, they should have infinite life. The leveling of the S-N curve is the fatigue limit,
characteristic of ferrous metals but not of most nonferrous metals. However, the region to the left of the sloping part ofthe S-N curve is called the finite-life region, for at
the higher stress levels the test specimens or parts will eventually fracture in fatigue.
This is typical of certain structural parts in aircraft which have their histories carefully
recorded so that they may be inspected and/ or replaced as their fatigue lives are used
up in service. Also, growing fatigue cracks must not be permitted to exceed the critical
flaw size characteristic of the metal and the stress state.
Source: Donald J. Wulpi, Understanding How Components Fail, American Society for Metals, Metals Park OH, 1985,P 135-137
29
30
1-3. S-N Diagrams Comparing Endurance Limit for

Seven Alloys
100,000
...........
90,000
1.1
ao'tLerMcarOon.sttel,
Oil quenchedan 1 drawn'
~~
80,000
K"'!>
~,
~~e>/
10,000
l'
'~6QOOO
n
fir
J II, I,
~.
r;
....
<I)
r-,
c-J.-
<J'~
40,000
RC'-1,
'x
<Q-/'r.t>e:
" 30POO
~
~ '! ts as rolled
:l,fio..L....!. I
C
20,000 0..
~...Q'
.1
104
rr
III
.I II. J.I III . I I,. I II.
Not:
<:0,"",;." ~
I v.".,c ~
i?o;C:P~
l'fo~"ea-lea
~aJfn
r- ~~n
10,000
dJ.S.Jpe,.l.
cen/-carbon 'steel
O'i''7ve",
("o$e<tand drawn
0.
.t50.000
~
'wI. 0,7quenchedolnd -
(nd(cO/fis flf!f,mfn ~/c.;"lot rUf(furr
10 5
106
10'
Number of c~c1e5 for rupture.Ioq scale
108
Typical S-N diagrams for determining endurance

limit of metals under reversed flexural stress.
To determine the endurance limit of a metal, it is necessary to prepare a number of

similar specimens that are metals tested, and for most nonferrous metals, the S-N
diagrams become horizontal, as nearly as can be determined, for values of N ranging
from 1,000,000 to 50,000,000 cycles, thus indicating a well-defined endurance limit.
The S- N diagrams for duralumin and monel metal do not indicate well-defined endurance limits. The first specimen is tested at a relatively high stress so that failure will
occur at a small number of applications of stress. Succeeding specimens are then
tested, each one at a lower stress. The number of repetitions required to produce failure
increases as the stress decreases. Specimens stressed below the endurance limit will not
rupture. The results of fatigue tests are commonly plotted on diagrams in which values
of stress are plotted as ordinates and values of number of cycles of stress for fracture
are plotted as abscissas. Such diagrams are called S-N diagrams (S for stress, N for
number of cycles). In general, the S- N diagrams are drawn using semilogarithmic plotting as shown in the above diagram, which presents the results for various typical
materials.
Source: Fatigue and Creep Tests of Metals, P 220
1-4. Steel: Effect of Microstructure

I. 0 r--
- - - - - - - - - - - - - - - - ----,
" 0.9
"i
:--:.. 0.8
'<::
"i 0.7
o
'e 0.6
ec 0.5
e 0.4
:>
-g
w
0.3
0.2
0.1
o
Effect of steel microstructure on endurance ratio.
One of the more extensive investigations on influence of microstructure was conducted by Cazaud. The results of some of his work are summarized in the above bar
chart. His data confirm that 0.5 is a conservative number; he found ratios varying from
0.55 to 0.62 for highly tempered martensites. These data were also for steels in the
0.40% carbon range. When untempered martensite is included, the total ratio range is
from 0.26 to 0.62. Untempered 0.40% carbon martensite is about 55 HRC. Above 40
HRC, factors other than microstructure become more significant, especially nonmetallic content and residual stress.
Many believe that tempered martensite gives optimum fatigue properties. However,
much of the early work was with medium-carbon steels with intermediate hardnesses.
Only limited data are available for other structures, including low-carbon martensites.
Borik and Chapman determined the endurance limit of bainite and martensite in the
range 36 to 61 HRC. They used 5ll00, a 1.00% carbon steel. They concluded that
above 40 HRC, bainite had better fatigue properties at the same hardness than did
martensite, whereas below 40 HRC the reverse was true. They explained the results in
terms of carbide morphology and distribution. Below 40 HRC, the carbides in the
martensite are spheroidal. Above 40 HRC, the carbide associated with the bainite was
very fine and well-distributed, but below 40 HRC the carbides had a "pearlitic mode,"
which was less favorable in resisting fatigue.
Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure.
American Society for Metals, Metals Park OH, 1979,P 77
31
32
l-5.
Steel: Influence
of Derating
Factors on Fatigue
Characteristics
Derating factors for influence of surface condition

on fatigue.
The graph above gibes C, factors for various surface conditions.
It should be obvious that these factors are approximate.
since it is impossible to represent such variable conditions
by a single cur\e. C,. the size factor. is significant.
Earl) work b>
Horger firmly established that large-diameter
samples of the same metallurg)
were not
asgood in bending fati@e as weresmall samples. In the presence ofa stress gradient. as
in bending. a larger volume of metal is subject to high stress in a large part than in a
small-diameter
part. Since a large volume is subject to maximum
stress. there is a
higher probability
of a critical-size
nonmetallic
inclusion to be in that volume. The
fatigue properties established
by testing large specimens are thought to represent the
lower bound for a large number of small samples. Sinceaxial
tests. b> their nature. test
fairly large volumes at maximum stress. they also gibe lower-bound
results. C,, is usuall) taken at I.0 for diameters less than O.Jinches and 0.9 for diameters between 0.1
and 3.0 inches. It must be borne in mind that this is a \er) rough estimateand
that the
cur\es shown in the above graph are thought to be touard the conser\ati\e
side of
scatter bands. The 0.5 relation for S,and S,, is onl! reasonably accurate in the low and
intermediate
hardness
ranges because of limitations
related to microstructure.
nonmetallic-inclusion
content. and carbon content 31 higher hardnesses.
Source: 0. H. Brcen and E. hf. bene. Fatigue in hlashines and S~ruc~urcs~Ground

American So&t) for Metals. hlerals Park OH. 1979. p 72
Vshlslcs. . in Far~gue and hlcroaucture.
l-6.
Steel: Correction
Factors for Various
Surface
Conditions
Value for loading
in
Factor
Bending
K, .
1.0
0.58
0.9(a)
1.0
1.0
1.0
Kd, where:
d c 0.4 in.
0.4 in. < d
5 2 in. . . .
K. . . . . . . . . .
Torsion
33
Tension
0.9
0.9
1.0
From chart above
(aI A lower value 10.06 to 0.85) may be used to
account for known or suspected undetermined
bending because of load eccentricity.
Correction
factors for surface roughness
(kJ,
type of loading (14). and poti diameter
(KJ, for
fatigue life of steel ports.
Comparative
effects of various surface
various levels of tensile strength.
Source.
Metals Handbook.9rh
Parh OH. 19% p 61
Edlrlon.
Volume
I. Properr~erand
conditions
Selcc[ia,n.
on fatigue
Ironsand
Srscls.
limit
of steels at
American
SO~ISI)
ior hlclalr. hleralr
34
l-7.
Fatigue
Behavior:
Ferrous vs Nonferrous
Metals
S (stress)--N
(cycles to failure) curves. 4. ferrous
metals; B. nonferrous
metals. S, is the endurance
limit.
Traditionally.
the behavior
of a material
under conditions
of fatigue has been
studied by obtaining
the S-.lcur\es
(see above), where S is the stress and .E- is the
number of cycles to failure. For steels. in general. one obsenes a fatigue limit or endurance limit (curve A above) which represents a stress level below which the material does
not fail and can be cycled infinitely.
Such an endurance limit does not exist for nonferrous metals (curve B above). The relation between Sand ,V. it must be pointed out. is
not a single-ralue
function but serves to indicate a statistical tendency.
Up until the 196Os, almost all fatigue failures. and consequently
all the research in
the field, was confined
to moving mechanical
components
(e.g.. axles. gears, etc.).
Starting in the late 1950s. entire structures or very large structural elements (e.g.. pressure Lessels. rockets, airplane fuselages, etc.) have been studied and tested for fatigue.
This can beattributed
to the use of materials such as high-strength
alloys. together with
the advances in the fabrication
technolom.
resulting in monolithic
structures meant to
undergo high cyclic stresses in service. It is this class of materials which has shown
catastrophic
failures in fatigue. and it is for this kind of material that fracture mechanics is being applied. with considerable
success. to fatigue problems.
Source: Marc Andrt

Meyersand
Krishan
Engleuood
Chfk
NJ. 198-l. p 689
KumarChauls.
hlechanlcal
hle~allurg):
Prmaplesand
.Appllcaionr.
Prcm~ce-Hall.
Inc..
l-8.
Comparison
of Fatigue Characteristics:
Aluminum
Alloy
Typical
bending (R = -1) fatigue
rous and nonferrous
metals.
curves
Mild Steel vs
for fer-
Here it is noted the lack of the knee for the aluminum allo)
compared
with steel: that is. the point on the cune where the
number of cycles to failure becomes a straight line-essential11
infinity.
35
36
1-9. Carbon Steel: Effect of Lead as an Additive

120~-...,...--r----r--r---,---,---.-----,---.--...,.-----,
IOo/-----1I--+----1--+-----+~~
V>
n.
o BOI-------t---t----r-----:
o
o
Fatigue limit of leaded and nonleaded alloy steels as a

function of ultimate tensile strength.
Lead is often added to steels to improve machinability, although usually at the cost of a minor (usually) loss in mechanical
properties,. The interrelationship of lead additions with tensile
strength and fatigue limit is summarized in the above graph.
Source: George M. Sinclair, "Some Metallurgical Aspects of Fatigue."in Fatigue-An Interdisciplinary Approach. John J. Burke,
Norman L. Reed and Volker Weiss, Eds., Syracuse University Press, Syracuse NY, 1964. p 68
1-10. Corrosion Fatigue: General Effect on Behavior
Fatigue without corrosion

Fatigue limit
Logarithm of number of cycles needed for fracture _
Effect of alternating stresses with and without corrosion.
If a specimen is subjected to alternating stress (tension and compression in turn) over a range insufficient
to cause immediate fracture, gliding may occur within
some of the grains, but when the disloca tions reach a
grain-boundary they are halted, retracing their movement along the gliding-plane when the stress is reversed. If the material were ideal, it might be hoped
that the dislocations would merely move to and fro
along the plane, and that no damage would result. In
practice a large number of cycles can be withstood
without apparent damage, but in material as we know
it, slight irregularities will prevent smooth gliding indefinitely, and roughening along the original glidingplane will make movement difficult, so that gliding
will then start on another parallel plane. In the end,
bands of material will have become disorganized, and
ultimately one of two things must happen: (1) if the
stress range is low, gliding will cease altogether, the
only changes still produced by the alternating stress
being elastic, (2) if it exceeds a certain level (the fatigue
limit) the gliding will become so irregular, as to cause
separation between the moving surfaces, first locally,
producing gaps, which later will join up into cracks.
Thus above the fatigue limit (after a time which is
shorter at high stress ranges), there will be failure;
below the fatigue limit, the life, in absence of corrosion, should be indefinitely long as shown above.
In the presence of a corrosive environment the situation will be different. Disorganized atoms along a
gliding-plane may require less activation energy to
pass into a liquid than more perfectly arrayed atoms
elsewhere; certainly, while the atoms are in motion
along a gliding-plane, preferential attack may reasonably be expected even below the fatigue limit. This
means that there is no "safe stress range" within which
the life should be infinite. It is, however, convenient to
determine an endurance limit-namely, the stress
range below which the material will endure some specified number of cycles (the number must be stated).
It should be noted that, although stress-corrosion
cracking is often intergranular, corrosion-fatigue
cracks are usually transgranular, following glidingplanes inclined at such an angle as to provide high resolved shear stress. There are exceptions to both rules.
Whitwham, studying corrosion-fatigue cracks on steel,
found that, although mainly transgranular, they followed grain-boundaries for short distances, where
such boundaries chanced to run in a convenient
direction.
Source: Ulick R. Evans, An Introduction to Metallic Corrosion, 3d Edition, Edward Arnold (Publishers) Ltd and American Society
for Metals, Metals Park OH, 1982, P 160
37
38
1-11. Effect of Corrosion on Fatigue Characteristics of Several Steels

l00r------------,
'Copper' steel
l' 09')(, C Steel
80
:c
u
O
104
c.
100
,e
Ii;
40
'0
:>
0
:E
"'"e
e
'E
"
(/)
:c
u
105
106
107
.5
g-
lOB
ill
80
ill
40
20
104
105
106
107
lOB
:E
"'"<:
e
~
"R"."~
:>
0
60
l!!
"
104
105
106
.1
Chrome-
"
vanadium
20 steel
(/)
O'BB')(, Cr,014% Vo,0'46')(, C

lOB
.,
0
104
105
106
25
104
105
106
107
\4
50
,,~
chromium
i:
107
lOB
30
104
105
106
107
N = Cyclesto fracture 1I0g scale)
---Denotes testsmadein air

6------
"
"stream of fresh water

II
lOB
Tensile: 81000 Ib.lsQ,ln.
Tensile: 150600 Ib.lsq,in.
107
35
steel
27')(, Cr.0'2')(, C
(hardened and temperedl
(hardened and tempered I

Tensile: 65700 lb.Zsq.in,
0
103
40
'E
(/)
:E
"'e"
.
~
66
c.
'0
i:1<:
:>
0
0
103
129')(, Cr,011')(, C
Tensile: 89600 Ib./sq.in
45
Ii;
,e
i:1<:
20
w
"0
e
ill
(hardened and
tempered)
014')(, C,O'9B')(, Cu
Tensile:61500 Ib./sQ.in.
g-
ii.5
,e
20
.5
60
g"c.
40
(annealed)
Tensile: 103500 Ib.lsq.in.
saline riverwater
Typical curves showing the number of cycles needed to produce fracture at different
stress ranges in absence and presence of corrosion.
Two main procedures are available for corrosion-fatigue

tests:
One-stage tests. Here the corrosion fatigue is continued until
breakage. The logarithm of the number of cycles needed to produce breakage is generally plotted against the stress range, as in
the above curves selected by Gough from McAdam's experimental data.
Two-stage tests. Here the corrosion fatigue is interrupted after a definite number of cycles, and the residual strength is estimated by measuring either (a) the endurance limit in the absence
of corrosive influences (i.e., the stress which can be withstood
for some definite number of cycles, (b) the number of cycles
needed to produce fracture in the absence of corrosive influences at some definite stress, (c) the tensile strength, or (d) the
shock resistance (Izod number).
Source: Ulick R. Evans, An Introduction to Metallic Corrosion, 3d Edition. Edward Arnold (Publishers) Ltd and American Society
lOB
1-12. Steel: Effect of Hydrogen on Fatigue Crack Propagation
PR PAGAIION LIfE
I
-,---_.
I
i
1+--.,--,---,--..c--,...-,-,--.-.+---.--.---.--.,..-,-,....,.......,r+--...,.----,.---..
1
6 7 0 9 0
S-N type of fatigue curve.
In the majority of all cases, the external load changing with

time, whereby low frequencies 10- 2 Hz) have the highest practical importance. Under these circumstances a structural component can be subject to fatigue which is conventionally described by an S-N curve relating the cycle life, N, to applied
stress, S, as in the above chart. In non-aggressive environments
an endurance limit can be defined below which no fatigue failure
occurs. A disadvantage of this approach is that S-N curves do
not differentiate between crack initiation and crack propagation. The number of the cycles corresponding to the endurance
limit presents initiation life primarily, whereas the number of
cycles for crack initiation at a high value of applied stress is
negligible. Consequently S-Ntype data do not necessarily provide information regarding safe-life predictions in structural
components. Particularly, if the structure contains surface irregularities different from those of the test specimens, these are
likely to reduce or even eliminate the crack initiation portion of
the fatigue life.
Source: M. Kesten and K.-F. Windgassen, "Design of Equipment to Resist Hydrogen Fatigue Service," in Current Solutions to
Hydrogen Problems in Steels, C. G. Interrante and G. M. Pressouyre, Eds., American Society for Metals, Metals Park OH, 1982,
P 390
39
40
l-1 3. Relationship
of Stress Amplitude
(a) Finite life ASR diagram. showing

S-.Y diagram,
showing life prediction
stress.
and Cycles to Failure
R = -I equivalent stress for R = 0.6 loading. (b)

for R = 0.6 loading using R = -I equivalent
The ASR diagrams normally use theendurance-limit

fatiguestrength value. but substitution
of the fatigue strength at specific
finite li\es can also be used (see chart a abole).
Life estimations from the diagram can be done using such information
as is
shottn in charts a and b abo\,e. Here the hno\\n stress range at
some Rvalue is conierted
to an equivalent completely
reversed
(R= - I) stress. and thisequkalent
stress is applied to the matefor the life estimate.
rials R = -I S-.Vcur\e
Designers
have the ability
to calculate
the components
stresses using classical formulas or the computer-based
finiteelement-analysis
(FEA) techniques.
Both of these methods examine the elements for the maximum stresses that are normall>
in the areas of a discontinuity.
or stress concentration.
Source. D. H Brscn and E hl Hens. Faugue m hlachlner and StrucrtmpGround

Metals
Park OH. IYTY. p 6
Amencan So&l! ior hkrals.
\ ehlcles. in Fatigue and hlwrosrruc~ure.
1-14. Strain- Life and Stress- Life Curves
2Nf
Reversals to failure (lag scale)
Strain-life and stress-life curves,
Fatigue damage is caused by cyclic plastic strain, and consequently, the fatigue life
should be related to the plastic-strain amplitude. Coffin and Manson independently
proposed a relationship between the plastic-strain amplitude and the cycles to failure
of the form:
AEp/ 2 =
EJ (2NJ),
where Ej is the fatigue-ductility coefficient, 2NJ is the number of reversals to failure, and
c is the fatigue-ductility exponent. Their equation is very similar to the Basquin equation relating the elastic-strain or true-stress amplitude to the number of load reversals
to failure:
I::.E,E/2
= aa = aJ (2NJ)b
where I::.E,/2 is the elastic-strain amplitude, E is the modulus of elasticity, his the
fatigue-strength exponent, and
is the fatigue-strength coefficient. A schematic representation of these relationships and their superposition is shown in the above diagram. The summation curve is analogous to the stress-life, Wohler diagram, if the
strain amplitudes are replaced by their respective stress amplitudes. The intersection of
the Basquin and Coffin-Manson plots is normally defined as the transition between
high- and low-cycle fatigue. Consequently, the regime of low-cycle fatigue depends on
the properties (for example, the ductility) of a particular material.
a;
Source: Edgar A. Starke, Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979, p 211
41
42
1-15. Fatigue Plot for Steel: Ultrasonic Attenuation vs

Number of Cycles
0.264
0.249
CD
0.232
:5!
t:I
z 0.216
0
;::: 0.200
ADDITIONAL PULSE
OBSERVED
:::>
Z
0.184
0.168
0.4 dBATIENUATION
CHANGE OBSERVED
0.152
5
10.8xl0
Typical plot of ultrasonic attenuation versus number of fatigue cycles for steel.
Joshi and Green determined the attenuation coefficient IX for longitudinal bulk
waves in aluminum and steel at 10 and 5 MHz, respectively. The measurements have
been performed in a pulse-echo mode, with the acoustic pulse reflected at the back
surface of the material. The above chart shows their results obtained on cold rolled
steel bars. The attenuation started to increase at about 6 X 105 fatigue cycles (65% of
fatigue life). At roughly 7.5 X 105 cycles (85%), an additional pulse was observed,
arriving earlier than the one reflected from the back surface.
Results are interpreted in terms of a series of microcracks being formed, probably at
the surface. As soon as the microcracks are sufficiently deep, they will change the bulk
attenuation. As soon as a macrocrack has been formed (by coalescence of microcracks), it will reflect part ofthe pulse. After that, the attenuation is primarily determined by the transmission coefficient of this single crack. Thus, the attenuation curve
(versus fatigue cycles) becomes discontinuous, as may be noticed in the above chart.
Source: O. Buck and G. A. Alers, "New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstructure, American Society for Metals. Metals Park OB, 1979. p 135
2-1. Typical S-N Curve for Low-Carbon Steel Under Axial Tension
28
26
~
TYPICAL FATIGUE CURVE

FOR M.S. UNDER
REPEATED AXIAL "TENSION
(f MIN. = 0)
1.
~ 24
I
rc)(
~ 22
't-
'-'
III
III
uJ
0::
J-
i
20
III
0
W
.J
0.
0.
18
<t:
16
14
I~
--
~
--fATIG-U-E-LlMi.:r-(REPEATEO- TEtlSION)
10
\2.
NUMBER OF STRESS APPLlC",.\ONS- MILLIONS
The term "fatigue" refers to the failure of metals from repetitions of stress rather than from a
single application, as occurs for example in a simple tensile test or with a brittle failure. The value
of the stress necessary to cause failure of a material from fatigue is lower than its nominal tensile
strength. For example, a sample of mild steel may have a maximum stress of 27 t.p.s.i. when
subjected to a single application ofload as in an ordinary tensile test. If, however, a stress of say
25 t.p.s.i. is applied repeatedly to the same material, failure will not take place until this has been
done a certain number of times, while at a lower stress still, the number ofload cycles required to
cause failure will be even greater. If testing is continued in this manner, a stress value will ultimately be found at which fracture will not occur, no matter how many stress repetitions are
applied. This value is known as the fatigue limit of the material. If the results from such a series of
tests are plotted, a graph such as the one above will be obtained, the curve tending to run parallel
to the abscissa after approximately IOmillion cycles (for steel), the corresponding value ofthe
stress being known as the fatigue limit. Under conditions of repeated tension the value of the
fatigue limit for the above mild steel which has a tensile strength of approximately 27 t.p.s.i.
would be of the order of 16 t.p.s.i. If the same steel was tested under conditions of reversed
bending stresses a value of the order of 12 t.p.s.i. may be found.
It must also be pointed out that where corrosive conditions operate in addition to fluctuating
stresses, failure from "corrosion-fatigue"may occur and, in these circumstances, the concept of a
fatigue limit does not apply, since if the stress applications are continued for a sufficient number
of times, ultimate failure will occur. Further, most nonferrous metals and alloys do not possess
a fatigue limit.
Source: F. R. Hutchings, "Fatigue Failure of Components of Lifting Machinery," in Failure Analysis: The British Engine Technical
Reports, F. R. Hutchings and Paul Unterweiser, Eds., American Society for Metals, Metals Park OH, 1981, P 344
43
44
2-2. AISI 1006: Effects of Biaxial Stretching and Cold Rolling
..
C'l
Eoff
cP
.. 0.2
)(
..r
00.4
a 0.6
"j
:::>
f-
:::i
Do
::l!:
oct
z
a:
lii
..J
oct
f-
f-
10 3
104
loS
10 6
REVERSALSTO FAILURE, 2N,
..
C'l
)(
Eeff
.. 0.2
00.4
..r
a 0.6
:::>
f-
:::i
Do
::l!:
oct
z
a:
f-
Vl
..J
oct
f-
f-
Runouts
r
103
104
loS
10 6
REVERSALSTO FAILURE, 2N,
Strain-life plots for two modes of deformation for 1006 steel.
Plots in the top chart are for biaxial stretching; those in the bottom chart are for cold
rolling. Included is the data band for the undeformed material. The effect of balanced
biaxial stretching on fatigue life was as follows: at large strain amplitudes (/:;.EI/2~
~ 2.5 X 10-3 ) , the fatigue life remained approximately the same or decreased slightly
when compared to that ofthe undeformed material; in contrast, at small strain amplitudes the fatigue life increased as a result of the prior deformation.
After cold rolling, the fatigue life was approximately the same as in the undeformed
material at large strain amplitudes (short lives) but it was longer at small strain amplitudes (long lives). Thus, unlike BBS, CR appeared to cause no reduction in fatigue life
at short lives. Another difference between the two deformation modes was that the
scatter ofthe data was larger after BBS than after CR. Thus, BBS was somewhat more
detrimental to the fatigue life than CR.
Source: John M. Holt and Philippe L. Charpentier. "Effect of Cold Formingon the Strain-Controlled Fatigue Properties ofHSLA
Steel Sheets," in H'Sl.A'Stccls-c-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 217
2-3. AISI 1006: Weldment; FCAW, TIG Dressed
SAE-1006
R' 0.1. t' 0.13" 13.3mml
--~--
Smooth Specimen
--<>-- TIG-Dressed
_.--fr-.- As-Welded
00 0 - - - - 0
0 roo 6_ --rrtr-lS" i:J_ _
tr
6
~D;!.~_
0
0
0
-o 0
.-._
oro
~
'-.0
0-.
o DO ~._._
._....g 0
'-''''lJ..._
'--0
10 5
..............
10 6
NT' CYCLES TO FAILURE

Fatigue strengths of FCA W/TlG- dressed joints compared to those without TlG
dressing for AISI I006steel (unwelded). The improvement in fatigue provided by TlG
dressing tbe welds is obvious.
Source: Kon-Mei Ewing, Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens. "Weld Fatigue ofTlG-Dressed
SAE-98QX HSLA Steel,"in HSLA Steels-Technology & Applications. American Society for Metals, Metals Park OH. 1984,p 557
45
46
2-4. AISI 1006: Weldment; Shear Joints
~.
or
CJ)
<J
\0 1
(a,.-32KSI
\;
ur= 32 KSI
SAE 1006 LAP-SHEAR WELDS

Kfmax = 2.77 I R = 0.1
EXPERIMENT
PREDICTION
10 5
NT
106
I
CYCLES
Total fatigue life predictions and experimental results for FCA W, AISII0061ap-shear
joints. Note that the results and predictions compare closely.
Source: Kon-Mei Ewing. Pei-Chung Wang, Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue ofTIG-Dressed
SAE-980X HSLA Steel," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984,p 562
47
2-5. AISI 1006: Weldment; Lap-Shear Joints
(f)
<I
10 1
~"-'2KSI
;;;
CTr =32
SAE 1006 LAP-SHEAR WELDS
a.
~
10
KSI
Kfmax = 2.77. R = 0.1

EXPERIMENT
PREDICTION
101
NT,
CYCLES
Total fatigue life predictions and experimental results for FCAW, AISI 1006 lap-shear
joints. Here, the prediction and actual results are very close.
Source: Kon-Mei Ewing. Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue ofTIG-Dressed
SAE-980X HSLA Steel, "in HSLA Steels-Technology &Applicalions, American Society for Metals, Metals Park OH, 1984,p 562
vi
<l
48
2-6. AISI 1015: Effect of Cold Working
0.20
5.0 .
E
~
E
.x:
4.0
0.15
<,
M
E
E
<II
0-
'"
3.0
I-
:E
ui
-O.I0:J
I<l:
a:
a: 2.0
uJ
::J
'"
<l:
lLJ
::
0.05
I<l:
u,
1.0
oL--!------~---_:::_---_::::__'O
20
40
DEGREE OF COLD-WORKING,
60
Comparison of effects of cold working on wear rate and fatigue limit of fully
annealed 0.15%C mild steel. Wear was determined in sliding between the end
surfaces of cylinders at a speed of 0.56 mjs under the loads .:82 N, ():124 N
and 0:147 N in machine oil. Fatigue limit (.) was determined by reversed
bending fatigue tests of notched plate specimens 25 mm wide and 4 mm thick
having a central hole 1.5 mm in diameter.
Attempts have been made to determine effects of cold-working on the

resistance to wear and fatigue of a O.15%C mild steel. Fully annealed material
was then cold-worked to different degrees and the specimens were machined
from it. Wear experiments were conducted in a rotating cylinder machine as
described above with a machine oil as the lubricant. Care had been taken to
avoid the effects of work hardening during machining by electrolytically polishing the sliding surface. Reversed bending fatigue tests were carried out by
using notched test pieces of the same material. The wear rate and the fatigue
limit are compared with the degree of cold-working in the above chart, which
shows a definite correlation.
Source: Yoshitsugu Kimura, "The Role of Fatigue in Sliding Wear," in Fundamentals of Friction and Wear of Materials, David A.
Rigney, Ed., American Society for Metals, Metals Park OH. 1981, P 215
2-7. A533 Steel Plate: Fatigue Crack Growth Rate

Stross-intensity factor rango, .0. K, ksi in. 1/ 2
50
10- 2
10- 4
a
<9':!
10- 3
~
E
E
~
co
...
01
00 I
1
Region 1:
slow crack
growth
I Region 3:
I rapid
I unstable
I crack
growth
I
I
I
I
I
I
10- 4
",'
1!
ie
at
eu
...e
10- 6
Region 2: power-law behavior
'"
~
.5
Z
~
co
...
ti
10- 6
l!
~
e
...u'"
l!u
...l!!
'"
:iE
10- 6
10-7
'"
:iE
10- 6
Stross-intensity factor rango, .0. K, MPa . m 1/2
Fatigue crack growth behavior of AS33 steel. The material was ASTM
AS33 B-1 steel, with a yield strength of 470 MPa (70 ksl), Test conditions: R= 0.10; ambient room air; 24C (75 OF).
The general nature of fatigue crack growth and its description using fracture mechanics can be
briefly summarized by the example data shown in the above chart. This figure, based on the
work of Paris et al, shows a logarithmic plot of the crack growth per cycle, daj dN, versus the
stress-intensity factor range, t::.K, corresponding to the load cycle applied to a sample. The
da] dN versus t::.K plot shown is from five specimens of ASTM A533 B-1 steel tested at 24C
(75 OF). A plot of similar shape is expected with most structural alloys; the absolute values of
daj d N and t::.K are dependent on the material. Results of fatigue crack growth rate tests for
nearly all metallic structural materials have shown that the da I dN versus t::.K curves have the
following characteristics: (a) a region at low values of dald N and t::.K in which fatigue cracks
grow extremely slowly or not at all below a lower limit of t::.K called the threshold of t::.K, t::.K,,,;
(b) an intermediate region of power-law behavior described by the Paris equation:
~=C(t::.KJ'
dN
Source: J. H. Underwood and W. W. Gerberich, "Concepts of Fracture Mechanics." in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH. 1982. P 18
49
50
2-8. A514F Steel Plate: Fatigue Crack Growth Rates
20
MPavm
100
100
20
10'
A514F
CON Quality
'(J '
I
/;
/,.
.,
100
10
z!.
~ 10"
<
A514F
CaT Quality
~
E
E
lS----1
u-
-::'O
15-----<
TL-- ---- .. - --- -_ :
~I=-:-:=:':::- i
l'.K. ksl\ ii1.
'00
Plots of fatigue crack growth rate versus range of stress

intensity factor (best fit lines) for A514F plates.
The increased isotropy in the CaT over the CON steels is evident with the through thickness (ST, SL) orientation having the
fastest growth rate in the CON steel and showing the greatest
improvement by CaT.
Source: Alexander D. Wilson. 'The Effect of Inclusions on the Properties of Constructional Steels." in Wear and Fracture Prevention. American Society for Metals. Metals Park OH, 1981. p 196
2-9. A514F and A633C: Variation in Fatigue Crack Growth Rate

With Orientation
5.10-'
~Kof 50 kst/iil (55MPa.frii)
CON~
A633e
CaT
A633C
CON~
A514F
caTD
1110"' 3.10-'
~~~
ICaT I
4.10'
A514F
8x10"
12110"
16.10"
20110-'
daldN, inches/cycle
Comparison of fatigue crack growth rate variation with
orientation for A633C and AS14F plates at two tJJ( levels.
These data show that the CaT improvement in FCP growth rate
takes place only at higher L1Klevels. Additionally, this figure indicates
that there is a more substantial enhancement in FCP behavior for
ASI4F. Also there generally appears to be more anisotropy in the
ASl4F steels of both quality levels. It has previously been shown that
higher strength level steels tend to be more adversely affected by inclusions associated in groups, such as present in CON steels.
Source: Alexander D. Wilson, "The Effect ofInclusions on the Properties of Constructional Steels, "in Wear and Fracture Prevention, American Society for Metals, Metals Park OH, 1981, P 197
51
52
2-10. A514F: Scatterbands of Fatigue Crack Growth Rate

MPavm
20
:------ .....,
10. 0
I
I
100
,
,:
,
,,,
,
10"
I
I
I
I
I
~
A514F
E
E
2'/.ln(57mm)Gage
CjCON
ClCaT
10"
10' L-_ _-----''--_-'------'_'-'-'--'---'-L-_ _-----'

10
100
6 K , ksi\/fil.
Summary scatterbands of fatigue crack growth

rate versus range of stress intensity factor encompassing all data points in 6-orientation testing
comparing CON and CaT quality A514F plates.
Source: Alexander D. Wilson, "The Effect oflnclusions on the Properties of Constructiona ISteels,"in Wear and Fracture Prevention, American Society for Metals, Metals Park OH, 1981,p 197
2-11. A633C Steel Plate: Scatterbands of Fatigue Crack Growth

Rates
20
MPavm
100
10-'
10-'
GI
GI
'fi 10"'
,5
4In(102mm)Gage
C=JCON
DCaT
10-'
A633C
10-'
L-_ _---'-_---''---'------'-----'----'--'--'---'--_ _------"-'
10
,c.. K kslv'ln:
100
Summary scatterbands of fatigue crack growth

rate versus range of stress intensity factor encompassing all data points in 6-orientation resting
comparing CON and CaT quality A633C plates.
In this presentation the generally faster FCP growth rates for

the CON steels at higher 6.K levels are displayed, as well as the
improved isotropy of the CaT steels.
Source: Alexander D. Wilson,"The Effect of Inclusions of the Properties of Constructional Steels," in Wear and Fracture Prevention, American Society for Metals, Metals Park OR, 1981, P 196
53
54
2-12. Low-Carbon Steel Weldment: Effects of Various Weld Defects

100
80
60
r-
... 1"-1-
40
-,~
- --
---
...
uv
o VV
-in
.0/.
'."C
-
8
f-
6 fI
10'
"
data bank
.,
'flo
9-
LC?cation of failure:
o Plate or weld edge
Porosity in weld
... Slag near surface
b. Slag at midthickness
V Lack of penetration
10
r----... r-9 0
Data bank
;--I-- r-- mean
f..' curve
IA
Low-carbon steel
reinforcement off
~
ch
I,t:
. --
~-
_C
ri
if.. ...ll ...
20
on
+ 1 standard deviat ion
....c
.,
OJ"
poli:he~ Plai~ pl~te
111
11
V
11
I~
1<:;1-
B 10'
Fatigue life, cycles
10
6 8
10
S-N curves showing effect of various weld defects on fatigue life of a

low-carbon steel weldment, presented as a comparison with fatigue life
of the plate.
Source: Metals Handbook, 9th Edition, Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 848
2-13. Low-Carbon Steel Weldment: Effect of Weld Reinforcement

and.Lack of Inclusions
-e
Reinforcement intact
Reinforcement removed
;j!.
"0'
0;
;:
"0
C
::J
(;
-----
-S
'"
40 _---""
___-~......=c-I_---~I""'-_~=__iI_---___l
----
1;;
CI>
.,
::J
.'"
'"
u,
201_----1_----1--
Ol.--
.l.-
..l-
...J......
...l-
0.2
0.4
0.6
0.8
-J
1.0
Slag inclusion length, in.
Fatigue strength of a weldment containing slag inclusions as a

percentage of the mean fatigue strength of a sound low-carbon
steel weld.
Source; Metals Handbook, 9th Edition, Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 850
55
56
2-14. Low-Carbon Steel Weldment: Effect of Weld Reinforcement

and Lack of Penetration
100
\\
80
60
\\
r-....
-,~'" <,
"
1'-,
.....
...........
.....
R . fl.
I
em orcement mtact
~I
<, ~
t--.
100
oob cycles
--.- r--r-.r------<.
.................. ~ r--
40
Reinforcement remived
.i
\/ -:
-.
-.-
2 000 000 cycles

20
o
o
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Lack of penetration half depth, in.
Fatigue strength ofa weldment containing lack of penetration as

a percentage ofthe mean fatigue strength ofa sound low-carbon
steel weld.
Source: Metals Handbook, 9th Edition. Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 849
2-15. Low-Carbon Steel Weldment: Computed Fatigue Strength;

Weldment Contained Lack of Fusion
100
;fi.
80
,,'
Qj
\ ~
1\
\'\
'\
III
"
<,
-g 60
'0
-S
'"
~ 40
r-,
<,
....
Reinforcement intact
I
I
'
I
- . - Reinforcement rejOved
<,
<,
"""-
" - -::::--- -~
....... ..........""'" ~ ~-- ~

b..
r-._ r--
Ii:
1 100 000 cycles
Ql
".,'"5.
LL
--
~.- ::-- ....t.
2 000 000 cycles

20
0.02
0.04
0.06
0.08
0.10
0.12
0.14
0.16
Lack of fusion half depth, in.

Computed fatigue strength of a weldment containing lack offusion as a percentage ofthe mean fatigue strength oCa sound lowcarbon steel weld.
Source: Metals Handbook, 9th Edition, Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
QH, 1983, P 849
57
58
2-16. Low-Carbon Steel Weldment: Effect of Reinforcement and

Undercutting
100
80
*'
'0'
'0
5
g
60
...o
fia>
40
~~
~ \~
\ [\
<,
~ .......
I" '-,......r-
10-. ~
""'-
r-; :::::-
Cl>
:::l
t::-I-- i'---
'
r-- 1-.
a>
co
--.
"k
0.Q2
0.04
0.06
""'--.
~ t:- t--
20
a
a
;00 000 cycles
r::: t:-- r--.r-- .r:;'/

to-.
';;
u..
. - Reinforcement removed
r': t::'- r---
t:
. fl.
""""- ern orcement Intact
0.08
0.10
t--
00 000 cycles
0.12
0.14
0.16
Undercut depth, in.
Effect of depth of undercut in terms of percentage of fatigue

strength of a sound low-carbon steel weld.
Source: Metals Handbook, 9th Edition. Volume 6. Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH. 1983. P 848
2-17. Low-Carbon Steel: Transverse Butt Welds; Effect of

Reinforcement
'I
........
600
r ,
'1-0---
, - - - , - - r ,-,.-.-,--""--"~J.------,---r---"I---y---r-T"""T""
"-
~"""""O<UNWELDED
.0
500
ro
300
E
(/)
200
100
0 0
~<, O~
~REINFORCEMENT OFF
.~.,-.-3
~ Q....
o
__
~_
1.5
2.3 - .
* .
-~-
REINFORCEMENT
Q..
't:..
ON
3'-............
~--.
t.:
..~
j
----t--~
.........
...
h =3.8mm
r h
Tr==r?J
<,
<,
.'6~-------~ __ C
~.
~
-.r:~
CYCLES
Influence of weld reinforcement on fatigue strength (R=O) of transverse butt welds of
quenched and tempered carbonsteels, From these data it is evident that removal ofthe
reinforcement (weld dressing) improves fatigue strength and fatigue life.
Source: Drew V. Nelson, "Fatigue Considerations in Welded Structure," in Proceedings of the SAE Fatigue Conference P-109,
Society of Automotive Engineers, Inc., Warrendale PA, 1982, P 206
59
60
2-18. A36/E60S-3 Steel Plate: Butt Welds

IOOr:---,----,.--..-r-.-rn"----r--r--r-OT,,""TT--..--r-r-,T"T,,,
80
600
6Or---====:=--_
400
300
40
CT, =0
200
20
10
100~
7O:e
5O<l
CT,' +35 ksl
vi
/!l361 E60S-3 Double-V Bull Welds
~ma.'2.5I,
30
R=0, I = 5/8 in.
OCT, = + 35 ksl
20
2
10
7
10
N1 ' Cycles
Fatigue crack initiation life predictions and experimental results for a-in. (16-mm) A36/E60S-3
butt welds.
Source: F. V. Lawrence, "The Predicted Influence of Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981,P 114
2-19. A514F/E110 Steel: Bead on Plate Weldment
~14F/EII(j Begd On plgte
Kfmal
31~1 DJ 0.01In.
R -0. t -112 ln.
- - . "".120kl'
---0 "" -120kli
200
200
100
-:---_-
vi
<I
---
00 0 - 0 - - - _
l.
10 ~
vi
so <I
"'----
30
20
20
10
Total fatigue life predictions and experimental results for A514F /EllO weldments with tensile and
compressive residual stresses.
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 113
61
62
2-20. A36 and A514 Steel Plates: Butt Welded
514F / E 110 Bull Wold

10
8
6
100 ~
70 ::f.
K,...3.15, 01- D.OIIn. (0254mm)
.-90, , -60, I -1/2 In.. (I2.7mml
50
--.,.s,
- - - ' , -0
..
s~s
K,.....
r
'"
30
20
10
NT' Cy(lu
Predicted effect of stress relief and stress ratio on

A514/EllO butt weld fatigue life.
1~'r--~--r-r""""~"'--~-~~""""~"'--~-'-"""'~~"'6oo
400
60
300
40
R'O
20
A36/E60S-3 Bull Weld

K,,"_ 2.35, Ot' 0.01 In (O.254mmJ
10
8
6
.-90,' -60, 1-I/Zin,,(I27mm)
--.,-s),
- - - w,-O
~-.
s~s
K r_ .
30
20
'
10
10<
10'
NT' Cycle,
Predicted effect of stress relief and stress relief and

stress ratio on A36/E60S-3 butt weld fatigue life.
The results for the high-strength, quenched-and-tempered steels (upper chart), indicate that such
materials can sustain high residual stresses which do not relax. The total fatigue life of such materials is
strongly influenced by both residual stress (a,) and stress ratio (R). Stress relief or mechanically induced compressive residuals should be highly effective. An intermediate case is mild steel as shown in
the lower chart. Mild steels can have appreciable residual stresses; but, since the transition fatigue life
(N,,) is often very long (= 500,000 cycles), there are large amounts of plasticity at the notch root even at
long lives (106 cycles); this notch-root plasticity tends to relax rapidly the notch-root residual and mean
stresses with the result that N[is little affected forlives less than 106 cycles. The observed dependence of
N on stress ratio does, however, result in a predicted variation oftotal fatigue life with stress ratio R.
63
2-21. A36 Plate Steel: Butt Welded

100
80
600
60
Zero Mean Slress
40
200
20
...;;
VI
<I
10
8
--\---====::::::::d
Mean Slress Effects With No Rela.atian
~~
100
70 2:
30
A36 Butt Weld (HAl)
KImox =3 I R-O
a r =+35 ksi
20
10
IIO~
10'
N
Cycles
Mean stress relaxation behavior influence on fatigue

crack initiation life (A36 HAZ material, K f = 3, R = 0,
a r = +35 ksi (242 MPa) ).
Materials such as high-strength steels exhibit very little notchroot plasticity; consequently, a os may be larger than a r- The results obtained using the model agree with the experimentally
observed behavior. The above chart shows the qualitative behavior of N[ predictions.
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, pIli
64
2-22. Low-Carbon Steel Tubes: Effect of Welding Technique

.,.--'
10
-----
08
'--'.-,
""
~
" '"
.
- - - - f----- _._ .. _ -
1-. - ' ,,""' ~,~'-. "

I\. ....
"
"
i'..'..
'-'0
OB
"
"-
, .,~:,
','~
' ~~
,~
~/
..,~ ~'/~
-~/
.,~
.-:~~ /.--~-~
I~'"
V'~
%-. ~/.
..,~
'~r::-, '/8
~.,
04
f--------.--
~c
b-.,
-----_.-
A_
~.ij'/W
,-
~.
'~
~D
02
I I
I I
I I
I I
10
Fatigue strength of welded tubes: A - unwelded or welded without filler metal;
B - helical welding (700 angle); C - longitudinal or helical welding (550 or 600
angle); D - helical welding (500 angle).
Source: R. V. Salkin, "Low Cycle Fatigue of Welded Structural Steels: A Material Manufacture and Design Approach," in Proceedings of the Conference of Fatigue of Welded Structures, Vol 2, The Welding Institute, Abington Cambridge, 1971, P 193
65
2-23. Low Carbon Steel: Effect of Applied

Anodic Currents in 3% NaCI
.. 120
E
~
~
100
>-t--
w
o
80
~ 60
w
IX:
IX:
040
20
OL-_ _- '_ _-'--------'-_'--.l-L....L--L..J'--_ _- '_ _-'--------'-_'--L-JL..l...-'LIlk<
10'
10'
CYCLES
10'
TO FAILURE
Effect of applied anodic currents on the fatigue lives of low-carbon steel in

deaerated 3% NaCI solution. The corrosion rate ofthe steel in this solution is
virtually zero in the absence of applied currents. Note the independence of fatigue life at currents greater than-: 40 /-LA/cm 2, the absence of an applied stress
effectand the reappearance of a fatigue limit at currents less than ~ 0.2 /-LA/cm-,
The effects of salt concentration and temperature on the fatigue behavior of steels
have been studied. Experiments performed on mild steel specimens in distilled water
and in various concentrations of potassium chloride have shown that solutions ranging
from 2 molal to 1/40 molal have virtually identical effects on corrosion-fatigue lives,
but that at concentrations below 1/40 molal, the effect approaches that of distilled
water, although corrosion rates increase in an almost linear manner with solution ion
concentration. A similar result has been reported for deaerated 3% NaCI solution in
which corrosion rates were controlled by applied anodic currents (see above chart).
These observations indicate that a critical corrosion rate is a necessity to initiate
corrosion-fatigue failures. Additionally, increasing over-all corrosion rates over a long
range of rates has little effect on corrosion-fatigue resistance.
Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys." in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 344
66
2-24. Low-Carbon Steel: Effect of pH in NaCI and NaOH

60
55
/
I NORMAL SOLUTION
50
45
.,~ 40
Q
"'.H
=12.1
~ 35
UJ
II:
t-
'" 30
25
pH :10
20
.5
10
The effect of pH on the fatigue behavior of low-carbon steel in

NaCI+NaOH.
The effect of stress frequency on corrosion fatigue has been studied by a number of investigators but is still not completely understood. For example, an early review of corrosion fatigue
noted that it is difficult to compare the corrosion-fatigue properties of metals exposed to like
environments because data reported are usually taken at different frequencies. In general, a
given time was found to produce more damage at a higher frequency, but a given number of
cycles was found to produce greater damage at low frequencies. For low-alloy steels in fresh
water, a frequency of 1450 cycles/min produced failure in 106 cycles or II Y2 hours, but at a
frequency of 5 cycles/min, failure occurred in 0.11 X 106 cycles, or 400 hours.
To date, the effect of pH of aqueous solutions on corrosion-fatigue behavior has not received
extensive study. A study of the effect of 0.1 N HCl on the fatigue life of steels showed greater
damage in this medium than in neutral potassium chloride solutions. Tests conducted in alkaline
media, at a pH above 12.1, showed that a fatigue limit is regained, this limit improving at still
higher pH values (above chart). These investigators suggested that corrosion fatigue is a result of
differential aeration cells, which produce pits in the metal surface, and that a high pH provides
diffusion barriers (ferrous hydroxide) to oxygen on the surface. Higher fatigue limits at high pH
are explained in terms of a "better and more perfect film barrier."
Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys." in
2-25. Low-Carbon Steel: Effect of Carburization

and Decarburization
CJ)A
CJ)
NON-DECARBED MATERIAL
~ B ...............~
t)
~
INTRINSIC'
FOR DECARB
MATERIAL
. , ...DECARBED
-COMPOSITE
LIFE
Influence on fatigue SoN curve of soft surface
caused by decarburization.
Parts that were made from low-carbon steel, but have high-carbon surfaces resulting from carburizing, have special microstructural factors that
must be considered. From the carburizing process an intergranular oxide
network may develop. This oxide may be an alloy oxide which causes alloy
depletion in grain-boundary areas. As a rule, this condition is thought to
detract from fatigue properties. The two exceptions may be in combination
rolling and sliding contact fatigue, where the oxide network may enhance
low-cycle bending fatigue-somewhat the same as does decarburization. The
effect on high-cycle bending fatigue is deleterious, as is decarburization.
These concepts are shown schematically in the above chart.
Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979,P 80
67
68
2-26. A514B Steel: Effect of Various Gaseous Environments

on Fatigue Crack Propagation
r------;;;:;:;;;::::::j""". 18
16
0.4
14
2
AK = 29.7 MN/m 3/
0.3
0.2
0.1
4
2
The fatigue-crack propagation of ASTM A514B

steel in various gaseous environments.
The origin of the element (such as sulfur) Onthe surface could result from
its presence in the gas phase (such as for hydrogen sulfide). It could also
originate as an enriched sulfur layer associated with a propagating crack, as
would be the case for sulfur segregated to a grain boundary. Oxygen alone on
the surface tends to drive the hydrogen-dissociation reaction rates in the opposite direction from the sulfur. The above bar chart shows how a mixture of
environments can influence the fatigue-crack growth of an alloy when all the
loading factors are kept constant.
The main influence ofthe environment is to supply the active atoms to the
vicinity of the crack tip. Subsequent interaction with the crack allows the
degradation mechanism to take place. The next step in the environmental
interaction is the transport of the active species to the location in the vicinity
of the crack tip where the degradation mechanism takes place.
Source: H. L. Marcus. "Environmental Effects 11:Fatigue-Crack Growth in Metals and Alloys," in Fatigue and Microstructure,
American Society for Metals, Metals Park OH, 1979. P 371
2-27. Cast 1522 and 1541 Steels: Effect of Various

Surface Conditions
NONE
0.006
0.012
NONE
0.006
0.012
SHOT PEENING INTENSITY - C, ALMEN

The effect of shot peening, carburization, and decarburization on the endurance ratio of normalized and tempered cast steel with cast surfaces. Plate bending fatigue
.specimens were used to secure these data.
Decarburization of the surface lowers fatigue resistance. This

effect, along with the beneficial effects of carburization and shot
peening on the endurance ratio of cast low alloy plate specimen
in bending, is shown in the above diagram. The nominally 1.2%
Mn steels with 0.22% C and 0.4 I% C, respectively, were normalized and tempered to 78 and 95 ksi (538 and 65~ MPa) ultimate
tensile strength, respectively. The depth of decarburization
(0.05% C at the surface) was 0.06 in. (1.5 mm); that of carburization (1.15% C at the surface) was 0.08 in. (2 mm).
Source: Steel Castings Handbook. 5th Edition. Peter F. Weiser. Ed.. Steel Founders' Society of America. Rocky River OH. 1980.
P 15-29
69
70
2-28. Cast A216 (Grade WCC) Steel: Fatigue Crack Growth Rate
20
40
60 80 100
200
3r---.---.--__,_~__,_,...,.,_r_--"
lTyS : 48 ksi
2
(331 MPo)
TEST TEMP: 7soF ( 24C)

TEST FREQUENCY: 600 cpm
WOL TYPE SPECI MENS
- 4
UPPER SCATTER BAND

( SLOPE n : 3 ) """
2
-.I
U
-,
E
10- 6
- 8
"0
<,
0::
10- 5
0::
8
I
0::
{.?
0::
~.
'<l
I~
0
10-7{.?
0::
I<l:
"0
I-
"0
<,
"0
o.
0
<l:
0::
U
10- 6 '--_ _-'-_-'----'_L.-I-L.........--'-_ _---'

10
20
40
60
100
200
STRESS INTENSITY FACTOR

~K,
RANGE,
ksi ~
Fatigue crack growth rate as a function of ilK for

A216 (grade weC) cast steel.
The equation do I dN= CoI1K" is sometimes referred to as the

Paris law and predicts a linear plot of log dol dN versus log 11K
with slope n. This is observed for a wide variety of materials and
is illustrated in the above diagram for an ASTM-A2l6, Grade
WCC cast steel. Some materials show a significant influence of
the mean load or Klevel on fatigue crack growth rates. The ratio
ofKmin to K max is used to express the mean load conditions.
Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
p4--16
3-1. AISI 1030 (Cast) Compared With AISI 1020 (Wrought)

(\J
<,
IIII
<I
--r--...,
I.O....--r--..,.....--r---r--.......
'f "''0.1
ILl
C
/WROUGHT SAE 1020

....
I-
:J
Q.
0.01
<t
Z
0:
/CAST SAE 1030
/O,W
O\' '
0,'
'00"::0."
1)'0'
<i
PLASTIC
",
~O
0.001
"
t;
0.0001 ~_L--..l....::-~-..l...:-....l..:~....L.:=--~
7
100
10
102
106
10
REVERSALS TO FAILURE - 2N f
Low-cycle strain-control fatigue behavior of carbon steel.
A number of techniques are available for computing the lowcycle fatigue life, although a straightforward approach is simply
to compute the fatigue life from the expected cyclic plastic strain
amplitude in service. Errors in computing or estimating 6.Ep
produce a smaller change in the computed cyclic life than similar
errors in the elastic strain range. Note that there is a large difference in slopes "c" and "b" in the above diagram. Plastic strain
ranges may be computed using sophisticated finite element
techniques, estimated from simple approximations such as
Neuber's rule or experimentally measured in component or
model tests.
Source: Steel Castings Handbook, 51hEdition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,P
4-13
71
72
3-2. AISI 1035: Effect of Gas and Salt Bath Nitriding
50
A1MQSPHRE
NITRIDED
40
SALT BATH PRQCT_'S

0
lC
'0
QUENCIU:D AND TEl1PERID
AT l050 F (565 C)
~
HUKBER OF CYCLES
Torsional fatigue strength of AISI 1035 steel-stress vs number of cycles for

completely reversing torsional fatigue, featuring the effects of gaseous atmosphere and salt bath nitriding on fatigue strength.
Source: J. A. Riopelle. "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals. Metals
Park OH, 1977, P 286
73
3-3. AISI 1040: Cast vs Wrought
TENSILE
YIELD
STRENGTH STRENGTH ELONG. HARDNESS
ksi MPo
ksi MPo
%
BHN
CAST
WROUGHT
94 (6481
90(62~
56 (386)
56 (3B6)
25
27
187
170
350
en 50
.><:
en
J, 45
f3
g:
WROUGHT
NO
300
f3
0:
40
CAST
NOTCH
en
250 ~
~ 35
::::>
::::>
5(30
200
NOTCHED
<t
<t
~
~ 25
150
5
10
6
10
7
10
CYCLES TO FAILURE
Fatigue characteristics (S-N curves) for cast and wrought
1040 steel in the normalized and tempered condition,
both notched and unnotched. R. R. Moore rotating beam
tests, K, = 2.2.
Cast steel suffers less degradation offatigue properties due to

notches than equivalent wrought steel. When the ideallaboratory test conditions are replaced with more realistic service conditions, the cast steel shows much less notch sensitivity to variations in the values of the test parameters than wrought steel.
Under the ideal laboratory test conditions and test preparation
(uniform section size, polished and honed surfaces, etc.), the endurance limit of wrought steel is higher. The same fatigue characteristics as those of cast steel, however, are obtained when a
notch is introduced, or when standard lathe-turned surfaces are
employed in the rotating beam bending fatigue test. These effects are illustrated above.
P 15-10
74
3-4. AISI 1045: Relationship of Hardness and Strain-Life Behavior
.
<l
C\l
.....
0.1
SAE 1045
or
"0
.~
Q.
0.01
C3:
c:
>-;~-
_ _ _ 1l911HB
'.:::--=----_410
..... __ - - - 3 3 0
1:1
+0
(/)
0.001
- - - - 280
2211
10
102
103
104
lOll
Reversals la Failure, 2N
10
107
Strain-life behavior of medium-carbon steel as a

function of hardness.
Strain-life curves at various hardnesses are presented in the

diagram above to demonstrate the range of properties attainable by tempering. Such information, used in conjunction with
life-prediction models, provides guidelines for optimizing material processing for specific situations.
Source: R. W. Landgraf. "Control of Fatigue Resistance Through Microstructure-Ferrous Alloys," in Fatigue and Microstructure, American Society for Metals, Metals Park OH. 1979.P 458
3-5. AISI 1141: Effect of Gas Nitriding
ATHOSI'HERE
NIUIDED
ATHOSI'HERE NITRIDED niDI
QU!NCBED AND TDO'UlD
GROUND to IlDlOVE COHPOUND

I.\YEII
AT 1050 r (565 C)
40
NlHIEI OF C'l'CLlS
S-N curves for 1141 steel-gaseous-atmosphere nitrided vs not nitrided

(quenched and tempered only)-showing stress vs number of cycles for
completely reversing torsional fatigue.
Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 287
75
76
3-6. Medium-Carbon Steels: Interrelationship of Hardness,

Strain Life and Fatigue Life
1.0~
\.
\.
1\ \.
\\ \
0.1
1\\'\\
600
~500~00
en
"
c;
"
t:
S
'0
200
.'\. \
Hardness, HB
<; ~ t\
.~
,,30~
-, "'-.."
0.01
n;
600
s:
~ ~ 400
<,
300<,
-200-
c:
0
0.001
0.0001
1
10
1M
1~
1~
Stress reversals to failure
1~
Predicted plots of strain versus fatigue life for typical mediumcarbon steels at the hardness levels indicated above.
Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH. 1978, P 673
77
3-7. Medium-Carbon Steel: Effect of Fillet Radii

,---
60
<-,
t--
--~~p~~lf:o
~
~/lf:
1ft'! d
~O'47"
on
r--,
gj 50
o
x
lI
oil
oil
III
0:
li;
I~
~40
I"""
~
o
~
o
30
300,000
~(l'O'I7
d. r 'Y-. I
Q ...
r- u
~S-
d.2/3"~
(iQIf;
DIcJ.:2
~
{)r
D/d: 2
O/cl.ys
0+
----
I II
1.000,000
10.000,000
50,000.000
I'lUMBlR OF C'{CLES ,0 FAILURE (LOG. SCALE)
~A1
UNMODIFIED
EXTERNAL 5TRE55- REl.IEVER
~&j
RE-EI1TRAN, FILLET
~4
5EPARATE COLLAR
The fillet radius at a change in diameter should be made as

great as possible. This cannot always be done; e.g., if the inner
race of a rolling bearing must abut against a shoulder formed by
the change in diameter. In such cases the stress-raising effect can
be moderated very considerably by adopting one of the expedients illustrated above.
Source: G. A. Cottell, "Some Common Stress Raisers in Engineering Parts," in Failure Analysis: The British Engine Technical
Reports, F. R. Hutchings and Paul Unterweiser, Eds., American Society for Metals, Metals Park OH, 1981, P 108
78
3-8. Medium-Carbon Steel: Effect of Keyway Design
5LEO-RUNNl:.R KEYWAY
PROFILED Kf..YWAY
70,000
PROFILED ...EV.......'(
MEDIUM-CARBON S,.EEL
(NORMALISE-D)
200.000
I ~ I
1.000.000
10,000.000
NUMBER. DF C'(CLES,.O FI'.\LURE (LOG.StALE)
Keyways are severe stress raisers from which fatigue cracks

are very liable to develop. Where bending stresses are predominant the cracks usually run transversely in the region of the keyway end, but where torsional stresses predominate they originate at the root at one side and may cause a portion of the shaft
to peel off or they may lie diagonally across the bottom. Effects
of various keyway designs on fatigue life are shown above.
Source: G. A. Cottell, "Some Common Stress Raisers in Engineering Parts," in Failure Analysis: The British Engine Technical
Reports. F. R. Hutchings and Paul Unterweiser. Eds., American Society for Metals, Metals Park OH. 1981, p 109
3-9. Medium-Carbon Steel: Effect of Residual Stresses

+600
+400
+200
NEAR-SURFACE
RESIDUAL
STRESS
O~-----------~-------
MPa
-200
(=
+3000 po IN/IN
TESTS STOPPED AT
10 7 CYCLES MAX.
-400
-600L---------....,...L,.-----------,J.
1.0
10
0.1
UFE CYCLES
Ie
10 6
Fatigue life relationship to near-surface residual stress.
Fully reversed fatigue tests on smooth bar specimens in medium carbon steels fully hardened show, as expected, that fatigue life increases directly with surface and near-surface residual compressive stress (see above chart). Residual stress measurements are usually made in the direction of the applied stress.
The achievement of high residual compressive stress in a part
requires a careful balance of the factors which affect this property and often involves a number of trade-offs which vary with
the application.
Source: J. Alan Burnett. "Prediction of Stresses Generated During the Heat Treating of Case Carburized Parts," in Residual Stress
for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 44
79
80
3-10. Medium-Carbon Cast Steel: Effect of Changes in

Residual Stress
25
211
.>:
NDR"""L.IZE ~N~
15
.4'~
:a---~
<,
-, -, ~HOT BLASi'
e;
III
i-
lJl
lil....
'4
-'
A '
Ul
ILl
NORH"LIZE -
<,
TEMP~
'"
"a...
~
~
ti
II
-5
111
4
CYCLES
Residual stress at completion of testing,
III
-411 .1.--+-
+--_ _--<
--+
--+-
-+-
+--'
Ill"
CYCLES
Change in residual stress with cycles at constant applied

stress,
The upper chart shows residual stresses existing on the completion of individual tests. The
similarity to S-N curves is apparent with the exception that the curve for normalized bars
(R 1 =+22 ksi) is inverted. Since the initial residual stress was known, there was a question on the
manner in which the residual stress changed during the progress oftesting. To explore this point,
two shot blasted bars were tested with applied stress levels of 40 and 55 ksi. The test on each
specimen was interrupted periodically to measure the residual stress at that time. The results are
shown in the lower chart. It is apparent that the change in residual stress is proportional to the
number of cycles when the latter is represented on a logarithmic basis. The lower chart also
points to the fact that the rate of change in residual stress is dependent on the level of applied
stress. Since the initial and final residual stress values were known for all bars, the slope for each
line could be determined.
Source: P. J. Neff. "A Quantitative Evaluation of Surface Residual Stress and Its Relation to Fatigue Performance," in Residual
Stress for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, /981, pp 127-128
83
4-1. Medium-Carbon Alloy Steels, Five Grades:

Effect of Martensite Content
100
100
650
c,
::;;;;
600
"'"
.~ 550
u.
~ <,
500
6.
01340
.4042
_4340
05140
6. 80840
I--.I'---
'"I'--. ---
90
Vi
~
~.
:~
6.
6.
n
I--.
70
All specimens 36 HRC

450
400
60
100
80
60
40
20
Martensite. %
For specimens having comparable strength levels, resistance

to fatigue depends somewhat on the microstructure. A tempered
martensite structure provides the highest fatigue limit. However, if the structure as-quenched is not fully martensitic, the
fatigue limit will be lower (see graph above). Pearlitic structures,
particularly those with coarse pearlite, have poor resistance to
fatigue. S- N curves for pearlitic and spheroidized structures in a
eutectoid steel are provided in chart 4-40 (p 122).
Park OH, 1978, P 676
84
4-2. Medium-Carbon Alloy Steels, Six Grades:

Hardness vs Endurance Limit
H-ll
1----1I---t---I----1t-"'1 Aus t ernpered
160
150
140
Vi
a.
0
0
0
130
E
--'
100
Q)
120
110
90
c
~
- SAE
SAE
.a._ SAE
0 - SAE
G - SAE
Q - SAE
:J
"0
6-
c
w
50
20
30
Rockwell
40
"c"
4063
5150
4052
4140
4340
2340
50
60
Hordness
Relation of hardness and fatigue strength for several steels.
The above chart and other data can be used to show the importance of
limiting the system to low and intermediate hardnesses as well as to point out
the importance of residual stress in fatigue. These data from Garwood, Zurburg and Erickson show a very tight linear relation up to about 40 HRC.
Above that hardness, the relation deviates from linearity, seemingly depending on carbon content. Carbon, however, is in an intermediate role here,
because it affects temperability. Because response to tempering is dependent
on carbon and alloy levels, it was necessary for samples of different grades to
be tempered at different temperatures to achieve the same hardness; consequently, a variety of residual-stress conditions resulted. The tempering
temperatures were necessarily sufficiently high to obtain 40 HRC; the residual stresses were reduced to a very low level, making all samples similar in
that usually the tensile strength for small sections decreases with increasing
section size and I or decreasing hardenability to compressive values. The sequence of transformation from surface to center, together with the temperature gradients, governs the outcome.
American Society for Metals, Metals Park OH, 1979, P 73
4-3. Medium-Carbon Alloy Steels: Effect of Specimen Orientation
Steel
BOO
100
s:
::;:
.~
.>!
600
BO
5,
.~
5,
60 "g
400
u,
u,
40
200
4140
Hardness, HRC 30
X4340
32
4027
44
4063
46
4032
4B
Avg
tensile
No. of strength Hardness,
tests(a) MPa ksl HRC
Longitudinal
4027
4063
4032
Tests
11 1179 171
12 1682244
11 1627236
Transverse Tests
4027
10
4063
9
4032
10
37 to 39
47 to 48
46 to 48
1130 164 34 to 39.5

1682 244 47 to 48.5
1254 182 47.5 to 48.5
(a) Number of fatigue specimens. For 4140 steel, 50

longitudinal and 50 transverse specimens were
tested; for 4340 steel, 10 longitudinal and 10 transverse.
It must always be considered that in rolled steels fatigue behavior is affected significantly by specimen orientation. Shown
above is the effect of orientation relative to fiber axis resulting
from hot working on the fatigue limit of low-alloy steels.
Through hardened and tempered specimens, 6.3 mm (0.250 in.)
in diameter, were taken from production billets. Specimens for
each grade were from the same heat of steel, but the tensile and
fatigue specimens were heat trated separately, thus accounting
for one discrepancy in hardness readings between the chart and
the tabulation above. Fatigue limit is for 100 million cycles.
Park OH, 1978, P 677
85
86
4-4. 4027 Steel: Carburized vs Uncarburized
0.02
(\J
.....
<l'"
,,;
0.01
Bending Fatigue
4027
Carburlzed
b. O.OOS"Ca..
D O.OIS" Ca.e
o 0.03S"Ca..
"0
0.005
ji
.~
.. 0.002
Ul
0.001 L...-_--'-_----'--;.----'-....-_'-;;-_........_--'-;;_--'
102
103
104
lOS
10
1
Reversals 10 Follure,2N f
Bending-fatigue results for uncarburized and carburized 4027 steel.
Bending-fatigue results supporting the validity of the effect of

carburizing are presented in the above curves. An uncarburized
baseline curve is compared with curves for three case thicknesses. As predicted, all carburized specimens show inferior
low-cycle resistance. At longer lives, the thinnest case offers
some improvement but tends toward the baseline as a result of
subsurface failure initiation. The thickest case, which shows the
greatest life improvement and has been found to exhibit surface
failure initiation, seems close to optimum.
Source: R. W. Landgraf, "Control of Fatigue Resistance Through Microstructure-Ferrous Alloys," in Fatigue and Microstructure, American Society for Metals, Metals Park OR, 1979, P 463
4-5. 4120 Steel: Effect of Surface Treatment in

Hydrogen Environment
2006
1600
600
25 Cr '"'A 4 - annealed
.4~.= 1.2%. H1 (10MPo)
=aOO3 sec:'
Cl bose metal - mech. pot.
m bose metot - chem. pol.
prior to surface treatmMt
surtoce treatments
Effect of surface treatment on fatigue life in hydrogen environment for a O.23C-O.98Cr-O.22Mo

steel.
In the above bar chart, the effect of surface treatment on fatigue life is summarized. The base metal was mechanically polished before surface treatment. For comparison, pot galvanizing and Ni-plating have been performed after mechanical as well
as after chemical polishing of the base metal. The results after
chemical polishing are given above in the form of dashed
columns.
The galvanizing such as Ni- and ZN-plating is by no means an
appropriate method to increase the fatigue life in hydrogen in
spite of the reduced surface roughness and protecting effect.
This is because the galvanizing produces a relatively high tensile
residual stress and the deposits possess generally poor ductility.
Source: Kyong-Tschong Rie and Werner Kohler, "Improvement of the Resistance of Metals to Cyclic Plastic Loading in High
Pressure Hydrogen Environment," in Current Solutions to Hydrogen Problems in Steels, C. G. Interrante and G. M. Presouyre,
Eds., American Society for Metals, Metals Park OH. 1982, P 380
87
88
4-6. 4120 Steel: Effect of Surface Treatment in

Hydrogen Environment
500'.-----~---~-----,-------,
hammered
I
/ alvanized
I
e-
300
III
I
Ia/hed
III
200
Ni- la/ed
-1 '25 erMa, -annealed

<a=!o.6%
t =0.003 s-'
H2 (10 MPaJ
number af cycles N
Effect of surface treatment on cyclic now curve in

hydrogen environment.
The above graph shows the cyclic strain hardening and softening curves for different surface treatments. It can be seen that the
fatigue behavior in hydrogen environment can be improved by
some surface treatments.
Pressure Hydrogen Environment," in Current Solutions to Hydrogen Problems in Steels, C. G. Interrante and G. M. Pressouyre,
Eds., American Society for Metals, Metals Park OH, 1982, P 379
4-7. 4120 Steel: Effect of Various Surface Treatments on Fatigue

Characteristics in Air vs Hydrogen
..e
..
2.0
..
0.
''';'.
"'l
1.0
.....
,:~
25 erMa' -annealed -+~

i: =0.003 sec-'
-H1,IOMPo)
S
--air
<II
. "," chem. polished .........-+-.J0.'
0,_ pol galvanized
g
~
0 Ni-ptatea
0,2
60
0,+ mech. polished

100
200
500
1000 2000
cntxa! nurrtJer of cycles Na
5000
MXJ
Fatigue life curve for various surface treatments of steel in

hydrogen environment and in air. Steel contained 0.23 C,
0.98 Cr and 0.22 Mo (4120).
Pressure Hydrogen Environment," in Current Solutions to Hydrogen Problems in Steels, C. G. Interrante and G. M. Pressouyre,
Eds., American Society for Metals, Metals Park OH, 1982, p 380
89
90
4-8. 4130 Steel: Fatigue Crack Growth Rate vs Temperature in

Hydrogen
10- 3
O"ty = 1330 MN m- 2 I O"tu = 1660 MN m- 2
O"ty = 1190 MN m- 2
O"tu =1310 MN m-
K =40 MN m- 3 / 2
.,
U
<Il
V>
3.4
3.8
4.2
I
4.6
5.0
liT, oK-I
Crack growth rate versus temperature in hydrogen

gas, for 4130 steel with yield strengths of 1330 and
o2
1190 MN m
The striking characteristic of hydrogen which sets it apart from other

causes of embrittlement is its large diffusivity. Although the diffusivity of
hydrogen does vary significantly among metals and alloys, it is nevertheless
always several orders of magnitude larger than the diffusivities of other species. Consequently, hydrogen transport is a prominent feature of discussions
of hydrogen-induced crack growth kinetics, and ofthe unique strain rate and
temperature dependence of hydrogen embrittlement.
Nelson and Williams reported the first complete investigation ofthe kinetics of crack growth in high strength steel exposed to hydrogen gas (see graph
above).
Source: Herbert H. Johnson, "Keynote Lecture: Overview on Hydrogen Degradation Phenomena," in Hydrogen Embrittlement
and Stress Corrosion Cracking, R. Gibala and R. F. Hehemann, Eds., American Society for Metals, Metals Park OH, 1984, P 17
4-9. 4135 and 4140 Steels: Cast vs Wrought

85
80
UNNOTCHED NOTCHED
75
65
<Il
.>t:
I
(f)
(f)
60
I-
D~-&
o~
............... o
"<, 80--0~
~"
06..,
....
0--
6,6..
100
80
0
.....
o,
""8<,6 . . - 6 .__
55
O-B~
UJ
a:
4140 (LONGITUDIAL) }NO.32

4140 (TRANSVERSE)
70
6.
4135 CAST STEEL
(f)
(f)
50
\~\A
(f)
45
40
UJ
a:
I60
\\.\.
,,~& .........
35
~~
30
25
104
(f)
-A-A_
~.--
'.~-.-
105
106
40
107
CYCLES TO FAILURE
SoN curves of a normalized and tempered AISI 4140
wrought steel in the longitudinal and transverse direction

and cast 4135 steel normalized and tempered. Tensile
strength for wrought steel: longitudinal, 110.0 ksi (758
MPa); transverse, 110.7 ksi (763 MPa); cast steel: 112.7
ksi (770 MPa).
In general, if the longitudinal and transverse ductility, impact,

or fatigue property values of rolled steel are averaged, they will
be about the same as properties of cast steel. One example of this
is shown in the S-N curves presented above. For these, a 4140
rolled steel was tested in fatigue in the longitudinal and transverse position and compared with a similar Cr-Mo cast steel.
Source: Steel Castings Handbook, 5th Edition. Peter F. Weiser, Ed.. Steel Founders' Society of America, Rocky River OH, 1980.
P 3-16
91
92
4-10. 4135 and 4140 Steels: Cast vs Wrought

STRENGTH
YIELD
TENSILE
ELONG
%
113 (779) B7 (560)
43
110 (758) 80 (552)
61
III (7651 81 (558)
30
ksi
CAST
4135
WROUGHT 4140-L
-T
80
WROUGHT
75
CAST
(MPo)
ksi (MPol
BHN
223
217
217
_
500
70
65
<l
W
o
~
60
...........
' ......
fh . . . . . . . . .
UNNOTCHED'
-,
LONGITUDINAL
------~
<, <,
400 11.
---
en
en
en
en
w
lr
I-
en
\
\
45
30
lr
300 I-
en
""
I ',
40
35
,1__ -.
55
::i
11. 50
,~RANSVERSE
NOTCHED
LONGITUDINAL
''-......
AND TRANSVERSE
- ---NO FAILURE
=*
200
106
CYCLES
TO FAI LURE
Fatigue characteristics (S-N curves) for cast and

wrought 4100 series steels, quenched and tempered to the same hardness, both notched and
unnotched.
The number of cycles to failure ofa structure subjected to the

above stress history can be expressed in terms of the SoN curve
shown above. The fatigue life increases as the cyclic stress amplitude decreases. For ferrous alloys a true endurance or fatigue
limit is reached below which fatigue failure is not observed.
The data presented in the above S-N curves illustrate several
important points. First, a fatigue limit is evident. That is, below
a certain cyclic stress amplitude, fatigue failure will not occur for
any arbitrarily large number of cycles. Secondly, while the fatigue properties of cast steel are lower than those obtained with
the wrought steel, it has less anisotropy. And, finally, the presence of a notch equalizes the fatigue properties of cast and
wrought steels.
The above data also illustrate that the fatigue limit of notched
test specimens is substantially below that of unnotched samples
when the fatigue limit is computed on the basis of nominal
stress.
P 4-8
4-11. 4140, 4053 and 4063 Steels: Effect of Carbon Content and
Hardness
1000,-------------,--------,--------.--------;
140
o to 2 micro-in.finish
900.1--------/-------t--------t
130
120
8001--------/-------t--,
in
Q.
sc
::;;
110
~.
E
~
.~
7001-------j-----;
100
:J
.~
.
:J
.~
u,
U.
90
6001-------:
80
70
60
30
40
50
60
70
Hardness, HAC
Effect of hardness and carbon level on fatigue

limit of alloy steels.
As shown above, when steels are hardened to 45 HRC or

higher an increase in carbon content can increase fatigue limit.
Although other alloying elements may be required in order to
attain desired hardenability, they have little effect on fatigue
behavior.
Park OH, 1978, p 676
93
94
4-12. 4140 Steel: Effect of Direction on Fatigue Crack Propagation

10-'
10'
.....
Q)
.....
Q)
10-'
>U
10:'
>U
.....
.....
E
E
....
....E
C121(j'
cI2 Ill"' -
'tJ'tJ
'tJ'tJ
10
1-1-_ _' - - _ - - - ' - _ - ' - - ' - - ' - - _ - - - - ' - '
10
20
304050
AK (MPaV'ffi)
a
100
10
L.L_ _-'--_---J'---'--'--'-_--,-lJ
100
AK(MPaV'ffi)
b
Fatigue crack propagation in an AISI 4140 steel: (a) longitudinal direction (parallel to rolling direction); (b) transverse direction (perpendicular to rolling direction).
The Paris power law, which describes the crack propagation rate in stage II
for a series of metals, is very useful because of its extreme simplicity. For
example, it has been observed experimentally that data points in the form of
log (do / dN) versus log ~K for a given material (constant metallurgical structure) from three different samples-edge crack in a compact tension sample,
through-thickness central crack in a plate, and plate containing a partially
through-thickness crack -all fall on the same line. Also, there is experimental
evidence that shows that the stress level by itself does not influence the fatigue
crack growth rate for stress levels below the general yielding. Thus, it can be
considered that the parameter ~K describes uniquely the crack growth rates
for many engineering applications. However, the structure of material can
influence fatigue crack growth rates drastically; the value of m can change a
lot. The above charts illustrate the directionality in the fatigue crack propagation rate in an AISI 4140 steel. The exponent m has a much higher value in
the transverse direction than in the longitudinal (rolling) direction, due to the
presence of elongated inclusions.
Source: Marc Andre Meyers and Krishan Kumar Chawla, Mechanical Metallurgy: Principles and Applications, Prentice-Hall, Inc.,
Englewood Cliffs NJ, 1984, p 714
4-13. 4140 Steel: Effect of Cathodic Polarization
Carras.....
LLi-O.4
I
Potential . - . _ " - - -
.~
(fj
(/) -0.6
~
>-0.8
.r
<{
~ -1.0
I-
0...
-1.2
10 5
106
CYCLES TO FAILURE
The effect of cathodic polarization on the fatigue behavior of
4140 steel (heat treated to HRC 52) in 3% NaCI solution at a
stress level below the fatigue limit in air is shown above. The use
of cathodic protection to prevent corrosion fatigue of steels depends sensitively on the hardness of the steel. For example, cathodic protection of a 4140 steel was shown to be feasible for
hardness values of Rockwell C 40. At higher hardness values, an
improvement in fatigue resistance is observed for moderate cathodic potentials, but complete protection is not possible. At potentials large enough to inhibit corrosion fatigue for softer
steels, a decrease in fatigue resistance is observed, presumably
due to hydrogen embrittlement (note above chart).
Source: D. J. Duquette, "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OR, 1979, P 360
95
96
4-14. Cast 4330 Steel: Effects of Various Surface Conditions

SURFACE ROUGHNESS- RMS - rnrn- 10- 2
.5
2.5
1.5
FUL L Y MACHINED
f===:J~
,
.... 0.2
<l:
0::
L,\
~
w
SILICA PRIME INV.

ZIRCON INV.
SILICA ALUMINA
C 6 C 16 WASH
ZIRCON GR.S.
INV
CRYOLITE
HIGH
WASH
""""';;:::::====~~:.:;:=~:::..CHROM ITE
GR. S.
PROPR. WASH
_ _ _ _==--========,JS~IL~ICAGR.
COPE
DRAG
200
MACHINED
400
600
800
1000
SURFACE ROUGHNESS - RMS- MICROINCHES

Relationship between surface roughness and endurance ratio (endurance limit divided by tensile strength) of quenched and tempered cast
4330 steel rUTS = 165-185 ksi (1138-1276 MPa)]. Fully reversed plate
bending tests.
.
Plate bending tests for quenched and tempered low alloy cast
4330 steel indicate that investment cast surfaces, or conventional castings produced with special mold washes, performed
equal to, or better than, fully machined and polished plate specimens. The data also suggest a tapering off of the surface effects
on the endurance ratio at 600 or more RMS surface roughness
as indicated in the above diagram.
Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America. Rocky River OH, 1980,
P 15-29
4-15. 4340 Steel: Scatter of Fatigue Limit Data

Tensile strength, ksi
12
800 iir=0-
160
---'-T-'------
200
----,rr--
240
rr-r --
280
.,---,
110
7001-------1---------11--------1--------:::J 100
50% survival
e:;;
90
Ii 600 f-------t--------::;;;Io--"""'-----t------=90%
~
c
'"
.~
.""'"
1;:
99%
500
1----.....~l=____=,....",~~~k:;;o>'-=::::::=:1f===-----l
80 c
'"
.~
70
400 f - - - - - - - - ' - - - t - - - - - + - - - - - - t - - - - - - - i 60
Approximately 1000 specimens. 1 heat
50
300 '--
800
L-
1100
-----''-----
~~
_:_:_'
2000
Interrelationships of alternating stress, tensile strength and expected percent survival for heat treated 4340 steel.
These data show survival after I a million cycles of AISI-SAE

4340 steel with tensile strengths of 995, 1320, and 1840 MPa
(144, 191, and 267 ksi). Rotating-beam fatigue specimens tested
at 10 000 to II 000 rpm. Coefficients of variation range from
0.17 to 0.20. From these data it is evident that scatter increases
as strength level is increased.
Park OH, 1978, P 678
97
98
4-16. 4340 Steel: Strength vs Fatigue Life
'.
0', = 0, = 174 ksi

/ ' F..tigue strength eOr'ie;ent
0. - 0',{2N,)b - 17412N,ro.o9
I
Fatigue strength
exponent
slope
/
=
-0.09
Reversals to failure. 2N f
Typical data for strength versus fatigue life for annealed 4340 steel.
Park OH, 1978, P 672
4-17. 4340 Steel: Total Strain vs Fatigue Life

100, - - - -....- - - - - - , , - - - -....------,-----r-------,------,
10-11----+---"O:-"'.--I-----r- 2
= l1p +
~
E
0 5Bl2N r O.57 + 0 0062(2N r O.09

.
,
.
,
10-2 1 - - - - - + - - - - 1 - - - - - - - 1 " - " ".......:----1----+-----+------1
10-4l::-
--'-:-
---lL::-
-'-::
-'--:--
-'-::--
-..L:--~"___J
100
Typical data for total strain versus fatigue life for annealed 4340 steel.
Park OH, 1978, P 672
99
100
4-18. 4340 Steel: Stress Amplitude vs Number of Reversals
4
10
( V = Uf =1200 MPa
C
Q.
3
10
~
C
b
~
~IN
10
10
Stress amplitude (!:J.aj2) versus number of reversals (2 N J) for AISI 4340 steel.
It is convenient to consider separately the elastic and the plastic components of strain. The elastic component can be readily described by means of a relation between the true stress amplitude and the number of
reversals (i.e., twice the number of cycles):
~E
a (a'J)
E
_e_=_o_= _ _ (2N)b
where ~Ee12 is the elastic strain amplitude, a o the true stress amplitude, aj-the fatigue strength coefficient
(equal to stress intercept at 2NJ = 1), NJthe number of cycles to failure, and b the fatigue strength exponent.
This relation is an empirical representation of the S- N curve above the fatigue limit. The above chart shows
an application of this relation to SAE 4340 steel. It was observed that fatigue life increased with decreasing
b. Morrow, based on energy considerations, showed that the fatigue strength exponent is given by:
n'
b=--1 + 5n'
where n' is the cyclic hardening coefficient.
Thus, the fatigue life under elastic cyclic conditions (whether stress- or strain-controlled) increases with a
reduction in n'. Of course, the higher the material coefficient aj, the better it is for fatigue. There is evidence
that ajis approximately equal to aJ' the monotonic fracture strength.
The plastic strain component is better described by the Manson-Coffin relation:
~E
-p-=
Ej( 2NJ
where ~Ep12 is the plastic strain amplitude, Ej is the ductility coefficient in fatigue and is equal to strain
intercept at 2NJ = 1, 2NJ is the number of reversals to failure, and c is the ductility exponent in fatigue.
Source: Marc Andre Meyers and Krishan Kumar Chawla, Mechanical Metallurgy: Principles and Applications. Prentice-Hall, Inc.,
Englewood Cliffs NJ, 1984. p 697
101
4-19. 4340 Steel: Effect of Periodic Overstrain

2000
250
1500
200
::;;
"c
'"
~
~
.>l
1200
1J,
c
1000
Vi
o No overstrain or single over-
strain at beginning of test
150
0
~
~
Vi
Periodicoverstrain
800
100
600
2
10
3
10
104
105
106
7
10
Number of cycles to failure
Overstrain superimposed on constant strain may have a significant effect on fatigue life. Shown above is the effect of periodic large strain cycles on fatigue life of AISI-SAE 4340 steel
hardened and tempered to a yield strength of 1100 MPa (160
ksi).
Source: Metals Handbook. 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH. 1978, P 681
102
4-20. 4340 Steel: Estimation of Constant Life

Minimum SIren, ksi
600
--400
~200
200
.00
600
800
1000
1200
1'00
Minimum Sites'. MPa
Potter has described a method for approximating a constant-lifetime fatigue diagram for unnotched specimens. Using this method, a series of points corresponding to
different lifetimes are calculated and plotted along the diagonal line on the left side
(R = -1). Each of these points is connected by a straight line to the point of the other
diagonal (R = 1.0) that corresponds to the ultimate tensile strength. A comparison
between the estimated constant-lifetime diagram and the experimentally determined
diagram is given in the above illustration. Here is presented a comparison between a
calculated constant-life fatigue diagram (solid lines) and experimentally determined
data (dashed lines). The calculated lines correspond well with the experimental lines.
Generally, the predicted lines represent lower stresses than the actual data. Estimating
fatigue parameters from the Brinell hardness number provides more conservative estimates. These results are only approximations, and the methods may not apply for
every material.
Park OH, 1978, P 681
103
4-21. 4340 Steel: Effect of Strength Level on Constant- Life Behavior

Minimum stress. ksi
1.0
200
c,
:;;;
~
E
150
1000
"'"
~
800
"E
'x
"1""""')
600
100
O't;..;
:9~
600
"""~~'1;1.()
..
400
10 6 cycles lifetime
50
200
o'----_---'-__---'--__--'--__-'---__ _ __''"____
~
-1200
-1000
"'1100
-tiOO
-400
-200
____<_ ____L_ ____'___ _ ___'____ _ _~
200
400
600
800
1000
L __
1200
_ _ '_ _____'
1400
1600
Minimumstress, MPa
Constant-lifetime fatigue diagram for AISI-SAE 4340 alloy steel (bar), hardened and
tempered to tensile strength levels of 860 MPa (125 ksi), 1035 MPa (150 ksi), 1380 MPa
(200 ksi) and 1790 MPa (260 ksi), All lines represent fatigue lifetimes of one million
cycles.
It may be noted that lives of the specimens at the three higher

strength levels are about the same; the scatter in data is at least as
great as any real differences in fatigue life among specimens.
Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Jronsand Steels, American Society for Metals, Metals
Park OH, 1978, P 669
'x
104
4-22. 4340 Steel: Notched vs Unnotched Specimens

Minimum stress,ksl
160
~ 800
~
,
1;;
E
E
"~
100
~
"w
600
:!E
400
60
200
O'--_ _....L-1000
-een
' -_ _---'-
-il00
-400
...L-_ _- ' " ' " -_ _--'---_ _- - ' ' - -_ _....L-
-200
200
400
600
-'--_ _- ' -
800
1000
-'0
- ' -_ _
1200
1400
Minimum stress.MPa
Constant-lifetime fatigue diagram for AISI-SAE 4340alloy steel (bar), hardened and
tempered to a tensile strength of 1035 MPa (150 ksi), Solid lines represent data obtained from unnotched specimens; dashed lines represent data from specimens having
notches with K, = 3.3.
Source: Metals Handbook, 9th Edition. Volume I, Properties and Selection: Irons and Steels, American Society [or Metals, Metals
Park OH. 1978, p 667
4-23. 4340 Steel: Effect of Decarburization

1800
250
1500
200
...
Ie.
If. 1200
:;:
~
;
900
.~
~ 600
0
0
o Not decarburized
Decarburized
0
0
150""
::i
e'
:;;
0>
C
.~
0
0
100 ~
r.
<t
.....
~
300
50
>-
Decarburization is the removal of carbon from the surface of

a steel part; as indicated in the above S-N curve, it significantly
reduces the fatigue limits of steel. Decarburization offrom 0.08
to 0.76mm (0.003 to 0.030 in.) on AISI-SAE 4340 notched specimens that were heat treated to a strength level of 1860 MPa (270
ksi) reduces the fatigue limit almost as much as a notch with
K,=3.
When subjected to the same heat treatment as the core of the
part, the decarburized surface layer is weaker and therefore less
resistant to fatigue than the core. Hardening a part with a decarburized surface can also introduce residual tensile stresses,
which reduce the fatigue limit of the material. Results of research studies have indicated that fatigue properties lost through
decarburization can be at least partially regained by recarburization (carbon restoration in the surfaces).
Source: Metals Handbook, 9th Edition. Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 674
105
106
4-24. 4340H Steel: Effect of Inclusion Size

1100
1000
--.r--..
s::;;
~ 900
1;:
en
.~
E 800
;;:
<,
o Small inclusions
Large inclusions
<,
140
~ ~.
700
100
Fatigue life of two lots of AISI-SAE 4340H steel; one lot (lower curve) contained
abnormally large inclusions; the other lot (upper curve) contained small inclusions.
Points on the lower curve represent the cycles to failure for a

few specimens from one bar selected from a lot consisting of
several bars of 4340H steel. Large spherical inclusions, about
0.13 mm (0.005 in.) in diameter, were observed in the fracture
surfaces of these specimens; the inclusions were identified as corundum or silicate particles. No spherical inclusions larger than
0.02 mm (0.00075 in.) were detected in the other specimens.
Large nonmetallic inclusions can often be detected by nondestructive inspection; steels can be selected on the basis of such
inspection. Vacuum melting, which reduces the number and size
of nonmetallic inclusions, increases the fatigue limit of 4340.
Source: Metals Handbook, 9th Edition, Volume 1, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 673
4-25. 4340 Steel: Influence of Inclusion Size

Xl0
200
190
180
STRESS
RANGE.
Ib/on 2
170
tal
160
150
140
130
i20
103
" " ,;: :I
SURFACE INCLUSIONS
a SUBSURFACE INCLUSIONS
DIAMETER,
on
0001-
I--~~
CYCLES
SoN curve and dependence oflife on inclusion size

for AISI 4340 steel.
Typical initiated crack sizes are l-lO Mm. As this is an order-of-magnitude greater
than dislocation substructure sizes, such an initiated crack will behave as in a continuum. For materials with lower stacking fault energy cross-slip and PSB formation is
difficult, thus inhibiting initiation. For such materials crack initiation can occupy a
significant fraction of life. Other microstructural sites for initiation are discontinuities
such as grain and twin boundaries, the latter being particularly operative in hcp metals.
Usually, however, at ambient temperatures it is the dislocation substructures which
dominate initiation.
For stronger, more complex alloys planar slip behavior dominates, making localized slip bands the initiation sites cfthe random notch-peak topography generated by
shearing a pack of cards. The interaction of slip bands with second-phase particles
(inclusions, precipitates) can produce a local stress concentration which cracks the
interface, producing a surface crack. The above SoN curves show the results of this
process for a low alloy steel. Variations in fatigue life relate to variations in inclusion
size. As well as debonding, oxide or carbide particles can crack under concentrated
localized stresses.
Source: B. Tomkins, "Fatigue: Mechanisms," in Creep and Fatigue in High Temperature Alloys, J. Bressers, Ed., Applied Science
Publishers Ltd, London, England, 1981, p 115
107
108
4-26. 4340 Steel: Effect of Hydrogenation; Static Fatigue

Normal notch strength
Ai>
.~,
275
250
'~
225
-.
"\\
\
'00
.x:
ul
(/)
~
200
(;)
"0
.~
a.
C
175
e\
125
\
\
100
:\
150
--
"..
Uncharged
++-
Bake 24 hr -
'---
0\ \
"\
\
\.
<,
Bake 18 hr
\ t-,
75
I'\.. '"
300 ksi
Bake 12 hr
Bake 7 hr
--
Bake 3 hr
Bake 0.5 hr
-.-..
50
0.01
0.1
1
10
Fracture lime, hours
100
1000
Static fatigue curves for quenched and tempered 4340 notched

specimens charged with hydrogen and baked at 150C (300 OF)
for the times shown.
There are many embrittling effects of hydrogen on steels: the ultimate strength of a steel may be
reduced; ductility as measured by total elongation to fracture or reduction of area may be decreased;
and crack growth may be significantly accelerated. The hydrogen responsible for these effects may be
present in the environment external to the steel or may be present internally as a reslt of steelmaking or
processing operations such as pickling or electroplating. Hydrogen may promote a transition from a
ductile to brittle fracture mode or it may reduce ductility without a change in fracture mode.
The graph above shows the effects of baking at 150C (300 OF) on the static fatigue (sustained loading) of the hydrogen-charged specimens. Increasing baking time effectively lowers hydrogen content
even in the plated specimens, and sufficient baking eventually restores the strength of charged specimens to that of uncharged specimens. The horizontal portions of the curves in the graph above are
designated as static fatigue or endurance limits, i.e., the stress level below which failure would not occur
no matter what the duration of stress application. As hydrogen content is decreased by baking, the
static fatigue limit increases.
The specimens used to obtain the above data were notched and therefore the static fatigue limits hold
for that particular notch geometry. In general, the sharper the notch, the lower the static fatigue limits,
an indication that a critical combination of hydrogen concentration and triaxial stress state is required
for crack initiation.
Source: George Krauss, Principles of Heat Treatment of Steel, American Society for Metals, Metals Park OH, 1980, p 223
4-27. 4340 Steel: Effect of Hydrogen
400
4340
250,000 PSI
350
300
250
co
P-t
0
0
0
200
......
.....
......
r-I
Cf)
Cf)
co
150
100
RATE
UNEMB. EMBRIT.
PLATED IN LAB
1000 RPM
COMMERCIALLY PLATED
50 250 RPM
0
PLATED IN LAB
200 RPM
'V
PLATED IN LAB
.33 RPM
0
10,000
1000
10
100
CYCLES TO FAILURE
<>
Schematic representation ofthe effect of cycling rate on theS-N curve of

hydrogen-containing 4340 steel, heat treated to a strength level of
250,000 psi.
Source: George Sachs, "Test Methods for Evaluating Hydrogen Ernbrittlement," in Materials Evaluation in Relation to Component Behavior (Proceedings of the Third Sagamore Ordnance Materials Research Conference). Syracuse University, Syracuse NY,
1956, P 508
109
110
4-28. 4340 Steel: Effect of Nitriding
----- -- ---
ATMOSPHERE
HITfUDED
------------
ATMOSPHERE NITRJDED TJfEN
GROUND TO RDlOVE COMPOUND
lAYER
5'>
QUENCHED AND TD1PERED

AT 1050 f (565 C)
HUHBER OF CYCLES
10 ~
10 5
S-N curves for 4340 steel, gaseous atmosphere nitrided versus

not nitrided (quenched and tempered only), showing stress versus number of cycles for completely reversing torsional fatigue.
Park OH, 1977, P 287
4-29. 4340 Steel: Effect of Nitriding and Shot Peening

1200 r-----,------,-----,,---,---,,------,------,-----------,
160
Nitridedcrankshafts
120
800
100
:Ii
~ 6001----+----'
Normally
heat treated
500 I----+-crankshafts --+--+--+----+----+-----,---------'5,....---1
350 L -_ _--'-
- ' - _ - - - ' _ - - ' - - - - ' ' - -_ _---...
----'_.L.>...
----'
105
Cvcles to failure
Comparison between fatigue limits of crankshafts (S-Nbands) and fatigue limits for separate test bars, which are indicated by plotted points
at right.
Mechanical working of the surface of a steel part effectively increases the

resistance to fatigue. Shot peening and skin rolling are two methods for developing compressive residual stresses at the surface of the part. The improvement in fatigue life of a crankshaft that results from shot peening is
illustrated in the above curves. Shot peening is useful in recovering the fatigue resistance lost through decarburization of the surface; decarburized
specimens were shot peened, raising the fatigue limit from 275 MPa (40 ksi)
after decarburizing to 655 MPa (95 ksi) after shot peening.
Park OH, 1978, P 674
111
112
4-30. 4340 Steel: Effect of Induction Hardening and Nitriding

160
._ 150 \
a..
'"
:5 140
C!.
~-130
e
u;
120
110
'r-.
As ni ide
\
\.
\ "I'
ilrid ~
"1\"'r--.
nducti n har en d
DU
Indu tio a dened
10 5
10 6
10 7
Cycles to Failure
As demonstrated in the above S-N curves, fatigue tests of
AISI 4340 steel in various surface hardened conditions show
that combined treatments produce endurance limits between
those developed by separate treatments.
Source: Sander A. Levy, Kenneth E. Barnes and Joseph F. Libsch, "Combining Nitriding With Induction Heating Pays a Bonus," in
Source Book on Nitriding, American Society for Metals, Metals Park OH, 1977, P 241
4-31. 4340 Steel: Effect of Surface Coatings

190r---------------------------,
0---
180
----
O~---------<>6
170
Cr + SFl
<>-
Cd+
O .. ~
~Chromate
~1f8are
~160
j
ISO
o
R 0.8
140
2000 cpm
RT
Cr +
~TUngsten
Carbide
SFl
we - Shot Peened
lJOL-----,---,-.w..u.uJ.,-----,---,-.w..u.u.l.;-----'---'-.L..LlllJJ.,-----'---'-.L..Lu.u-,;---'----'-L.LJL..LUJ
SoN curves (axial tension) of bare and coated 4340

steel in air environment.
Air
Stress
3.5% NaC1
Test
Condi tion
ksi
MN/m 2
Change,
%
Rotating Bending
R = -1
Bare
Cd + Chromate
Cr + Ory Film'
WC + Dry Film'
WC + Cori cone +
Dry Film'
105
105
95
90
so
724
724
655
621
621
0
-9.5
-14.3
-14.3
Bare
Cd + Chromate
Cr + SFL'
160
165
175
1103
1138
1207
WC + SFL'
140
WC + Cori cone +
SFL'
140
Axial Tension
R = 0.8
Stress
MN/m 2
Change,
%
90
13B
552
621
621
621
-81
-24
-14.3
-14.3
-14.3
+3.1
+9.4
110
165
90
758
1138
621
965
-12.5
60
414
965
-12.5
60
4.4
ks;
20
80
so
so
-31. 2
0
-43.8t
-4B. 6~
-62.5t
-57~
-62.5 t
-57f
'Shot peened
tCompared to bare alloy air value
~Compared to coated alloy air value
The above SoN curves, in conjunction with the table, contain the data obtained in air for 4340 steel, bare
and coated. Fatigue data at 107 cycles showed that the cadmium and chromium electroplates, particularly the
chromium, improved the fatigue strength. They were similar in both rotating bending and axial tension
fatigue tests. But in NaCl solution, significantly greater reductions in axial fatigue strength of the coated
alloys were observed due to environmental effects, which remains to be elucidated. Since the Cr and WC hard
(brittle) coatings have a relatively low intrinsic fatigue strength in comparison with the steel, they will become
discontinuous at a relatively low stress level owing to the development of fatigue cracks. (The Cr normally
contains internal cracks.) These cracks will permit access ofthe corrosive NaCI solution to the steel base at the
root of the fatigue crack. In the case of the axial tension test (high steady tensile load), it may be easier for the
environment to reach the crack tip.
Source; M. Levy and C. E. Swindlehurst. Jr., "Corrosion Fatigue Behavior of Coated 4340 Steel for Blade Retention Bolts of the
AH-I Helicopter.Yin Risk and Failure Analysis for Improved Performance and Reliability.John J, Burke and Volker Weiss. Eds .
Plenum Press. New York NY. 1980. P 275
113
114
4-32. 4340 Steel: Effect of Temperature on

Constant-Lifetime Behavior
Minimumstress, ksi
150
'"
c,
:2
";;;
'"
800
e
1;;
E
E
"
"x
100
600
"1'/.
~~.,
E
:J
E
"x
"00
107 cycles lifetime
Q)lj...;
".9s
400
..
~~
J'J'
..
'1"..0
6'a
1:>
50
200
0"--_ _-'----_ _--'--_ _---'--_ _---'_ _--'''"---_ _--'---_ _---'--_ _----'

-800
600
-1000
-200
200
400
--1l00
-400
Minimum stress, MPa
'"-_ _--'---_ _---'--_ _----'0

800
1000
1200
1400
Constant-lifetime fatigue diagram for AISI-SAE 4340 alloy steel (bars) hardened and. tempered to a
tensile strength of 1035 MPa (150 ksi) and tested at the indicated temperatures. Solid lines represent data
obtained from unnotched specimens; dashed lines represent data from specimens having notches withK1
= 3.3. All lines represent lifetimes of ten million cycles.
Park OH, 1978, P 669
4-33. 4520H Steel: Effect of Type of Quench
z
....
I
p:\
....:l
0
0
0
40
30
20
......
....:l
l=l
~4",~...
DIRECf (OIL) QUENCH

(COLD OIL) /
~.;~~.'::......
~....
~1ARQUENCH /-.:::
10
8
6
(400 0F OIL)'
-........:.::-~
-----.....:.....
---
4
3
2
10
10 3
10 4
CYCLES - N
Effect of quench type on fatigue of carburized differential cross.
Since 4520H steel is relatively low in hardenability for the part

involved, the depth-hardening characteristics of the two groups
were significantly different. The marquenched group had shallower case depths, which resulted in fracture origins below the
surface at the case-core junctures. However, when the comparison was made with higher-hardenability steels, with sufficient
gradient strengths and thus all fracture origins at the surface, the
difference was very slight, though still in favor of direct quenching. This is consistent with what is known concerning the differences in residual stress, which in this case would have been the
only other contributing factor. In other instances-such as for
gears, where distortion could be a factor-the results might turn
out differently for the marquenching.
115
116
4-34. 4520H Steel: Effect of Shot Peening
40
~,
30
CQ
,...J
20
0
0
0
......,
,...J
c:
u..;
if;
l:>:
~':".':'.":
~.;.......
---~
10
SHOT PEE~D
:::.:.:
lJNPEE\'ED~ __
----...:....
--
6
5
4
3
10
10 2
CYCLES - r\
Effect of shot peening on fatigue of carburized differential cross.
Shot peening is known to increase fatigue strength; hence,

tests were run to determine the amount of increase to be expected. The above chart shows some of the results. Shot peening
was found to provide significant fatigue-strength improvements. Peening surfaces that had suffered grinding damage was
found to be very beneficial, although not recommended because
of the high risk of having grinding cracks. Peening parts that
had marginal strength gradients improved the strength at the
surface but moved the failure origin to a subsurface location.
The net gain was small. It was also determined that to gain significant improvement the hardness of the shot used was very
important. Since carburized surfaces are very hard, the shot
must also be hard to be effective.
117
-- ---- -- -- --
ATMOSPHERE NITRIDED THEN

GROUND TO RI1'IOV COMPOUND
---
r.~Y~p
Al110SPHERE
NITRIDED
QUENCHED AND TEMPERED

AT 1050 F (565 CJ
HUHBER OF CYCLES
10~
1') 5
S-N curves for 4620steel, nitrided versus not nitrided (quenched

and tempered only), showing stress versus number of cycles for
Park OH, 1977, P 286
118
4-36. 4620 Steel: P1M-Forged

ksi
100
Height strain
80
oE
60
r.-------'-----Time
LI-!
, - - I_ _- ' - -_ _' - - - _ - - - - - ' -
4.0
5.0
6.0
Log cycles
7.0
Axial fatigue of P 1M-forged 4620 steel as a function of height strain during forging.
In general, sensitive properties improve as the level of upsetting is increased during the forging process. The diagram above
shows the effect for fatigue resistance, although the cyclic stress
state also influences fatigue behavior. An interesting feature of
P / M-forged parts is the fact that deformation does not significantly affect through-thickness properties as it does detrimentally for wrought material. For re-pressed parts, throughthickness toughness is slightly lower than longitudinal toughness. Upsetting increases longitudinal toughness while toughness
in the through-thickness direction remains at a relatively constant level.
Source: B. Lynn Ferguson, "Part II: Fully Dense Parts and Their Applications." in Powder Metallurgy: Applications, Advantages
and Limitations, American Society for Metals, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 100
4-37. 4620 Steel: P/M-Forged at Different Levels

700 ,.------,-----....---------,r--------,-------,
100
600
co
a..
:2:
Height strain, HIH o

71%
.'+-----+
.
CI)
CI)
80
56 & 65%
(JrnaK
500
42%
1'------+ 30%
cil
60
400
o~
0
7.0
8.0
Log cycles
SoN curves for P 1M-forged 4620steel at various levels offorging

deformation. As shown, fatigue limit increases as deformation
(level of strain) increases.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, p416
119
120
4-38. 4625 Steel: P1M vs Ingot Forms

100
1/
50
Ingot material
L,o 192 h
10
1/
/
\
P/M material
L,o 5
o
100
200
500
163
1000
2000
5000
Life, h
Fatigue life characteristics of P 1M roller bearing

cups, as shown by a typical Weibull plot. Shown is
a 10% life (L IO ) of 563 hr for P/M material
compared with L IO life of 192 hr for ingot
material.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals. Metals Park OH, 1984, P 620
4-39. 4640 Steel: P1M-Forged

100
,-------.----.r------...-------.-----.------.
0.365-1
':3.
8
80 dir
3-1/4
r-
'"
'Scatter band of SAE 4340
',~,Wrought steel, tested in the

'/
"
longitudinal direction
'0
"
~
~ u~
c'i)
'0,
,,
9-7/8 R.
60
-,
,,
'------------
,,
'~
'------_Q:
21
32
Specimen
configuration
40 '--------'---------''------'-------'---'
105
10
10'
103
10'
Cycles of stress
R. R. Moore fatigue curve for P/M-forged 4640

steel hardened and tempered to 33 HRC and a
yield strength of 138,000 psi.
Water-atomized 4600 steel powder was blended with graphite

and compacted in the split punch tooling. Fatigue data for P/M
forged 4640 are shown above, and these data fall within the scatter band for 4340 steel. The most impressive statistic is that the
P / M-forged parts passed the Army ambient and low-temperature
firing endurance tests.
Source: B. Lynn Ferguson, "Part II: Fully Dense Parts and Their Applications,"in Powder Metallurgy: Applications, Advantages
and Limitations, American Society for Metals, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, p 103
121
122
4-40. High-Carbon Steel (Eutectoid Carbon): Pearlite vs Spheroidite

Property
350. - - - - - - - - , - - - - - - . - - - - - - - , 5 0
45 '"
~.
~
i------j
40
'f;
c:
35 ~
7
6
10
10
Spheroidite Pearlite
Tensile strength,
MPa (ksi)
641(93)
676(98)
Yield strength, MPa
(ksi)
490(71)(b) 248(36)(c)
Elongation in 2 in.,
% ..............
28.9
17.8
Reduction in area, %
57.7
25.8
Hardness, HB .....
92
89
(a) Composition 0.78 C, 0.27 Mn, 0.22 Si, 0.016 S
and 0.011 P. (b) Lower yield point. (c) 0.1% offset
yield strength.
Both pearlitic and spheroidized structures have notably lower

fatigue strength than martensitic structures (see 4-1, on p 83). As
is shown above, the fatigue properties of spheroidized structures
are superior to those of pearlitic structures for eutectoid steels.
Park OH, 1978, P 677
4-41. 52100 EF Steel: Surface Fatigue; Effect of Finish and Additives
600
...
12
10
500
'"
1
2
3
4
5
6
7
8
9
10
11
.:<
12
.;
..'" 450
....
MLNG
MU\ G
MLCG
SLD G
MLNH
MU\ H
MLCH
SLD H
MLN P
MU\ P
MLCP
SLD P
~
4.14
3.79
3.45
3.10
til
..
~ 400
2. 76
'"
E
.~x
350
2.41
300
2.07
Mean Predicted Cycles to Failure
Effect of surface finish and additive on mean predicted surface

fatigue life. 52100 EF steel, high slip, high speed.
The mean predicted fatigue life is highest with a polished surface and least with a ground finish (9 versus I, etc.). Polished
surface has about 6 times and honed surface about 3 times the
fatigue life of ground finish. No interaction effect between additives and surface finish is revealed.
Source: S. Bhallacharyya, F. C. Bock, M. A. H. Howes and N. M. Parikh, "Chemical Effects of Lubrication in Contact FatiguePart II; The Statistical Analysis, Summary, and Conclusions," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 277
123
124
4-42. 52100 EF Steel: Surface Fatigue; Effect of Surface

Finish and Speed
II
600
4.14
3.79
3.45
... 500
Ul
""
.
'"
rl
..
::: 450
:;
2
400
'"e
~
350
:2
300
3.10
/---e---i
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
5E
5E
5E
5E
5V
5V
5V
5V
8V
8V
8V
8V
5E
5V
8V
5E
5V
8V
GLl
GL2
PLl
PL2
GU
GL2
PLl
PL2
GLl
GL2
PLl
PL2
GH2
GH2
GH2
PH2
PH2
PH2
2. 76 ~
to
2.07
10 7
Mean Predicted Cycles to Failure
Effect of steel, surface finish and speed on mean predicted fatigue

life. Low-viscosity mineral oil, no additive.
Interaction effects of steel with speed and of surface finish

with slip and speed on fatigue life are shown in the above graph.
The direct steel effects are nonsignificant. The effect of surface
finish is shown in the difference between the two line groups 13,
14, 15 (ground) and 16, 17, 18 (polished). The difference in the
line groups 4, 8, 12 (low slip) and 16, 17, 18 (high slip) again
brings out the very large detrimental effect of high slip on life.
Higher speed decreases life with the maximum effect observable
on 8620 CV steel (compare lines II and 12), on polished specimens at low slip ratio.
Source: S. Bhattacharyya, F. C. Bock, M. A. H. Howes and N. M. Parikh, "Chemical Effects of Lubrication in Contact FatiguePart II: The Statistical Analysis, Summary, and Conclusions," in Source Book on Gear Design, Technology and Performance,
Maurice A. H. Howes, Ed .. American Society for Metals, Metals Park OH, 1980, P 277
4-43. 52100 EF Steel: Surface Fatigue; Effect of Lubricant Additives
HLN 1.2
HL\ 1.2
HLC L2
~1.0 1.2
HI.N 112
HL\ 112
HLC 112
~LO 112
600
....
.:<
'"
4.14
J.79
J.45
90% Con fidence

Band
'"
'"
"
J.IO
N
::l
400
2. 76
.~
x J50
2.41
"
"
:E
JOO
2.07
250 L...;--'----'--'L..L__'_'..u..L..;-_ _'____'L....L..L..L.J..u.~--'----'L....IL..L--'-'..u.....--'----'L....IL..L--'-'L..1..L! I. 72

IO
10
10 7
:ican l'r-ed i c t ed Cyc l e s to Fa i lure
SoN diagram for 52100 EF steelshowing the effect oflubricant additives

on surface fatigue. The specimens had a ground finish, and a lowviscosity oil was used. Additives were used for I, 2, 3 and 4; the favorable
effect of the additives is obvious.
125
126
4-44. 52100 EF Steel: Surface Fatigue; Effect of Lubricant Viscositv,

Slip Ratio and Speed
1
2
3
4
5
6
7
8
9
10
600
550
-;: 500
11
.Yo
12
MLN
MLN
MHN
MHN
SLD
SLD
MLN
MLN
MHN
MHN
SLD
SLD
Ll
L2
Ll
L2
Ll
L2
Hi
H2
Hi
H2
Hi
H2
4.14
j.79
) .45
'
:: 450
f---<T---i
.
.
3.10
l/}
-!:
:: 400
2.76
'"
E
~ 350
2.41
:>!
300
2.07
250 '-c--'-------'---L-...L.J...J....IJ.Lr--.1_L-LJ...L.I...l..L.L...,-_...L--..L----'-...L-L.LJL..U.,,--........J_-'----.LJ~L.LJ.J
10 5
10
10
Mean Predicted Cycles to failure
SoN diagram for 52100 EF steel.
The effects of lubricant viscosity, slip ratio, and speed on fatigue life are
shown in the diagram. The 12lines in the figure are separated in two distinct
groups, low slip (lines I to 6) and high slip (lines 7 to 12). In each group the
effects of viscosity and speed may be noted. Viscosity X speed interaction
produces complex effects on mean predicted lives which under low slip conditions are not statistically significant in their differences. Only under high
slip condition, lines 9 versus 10indicate a small statistically significant lowering in mean fatigue life in high-viscosity oil under higher speed. A comparison
of lines 11 and 12 shows that the lesser life in synthetic oil with additive is a
statistically borderline case, though the trend is similar to that with mineral
oil under the present operating conditions. The regression analysis shows
that in the present tests both speed and viscosity have nonsignificant direct
effect on life, and a few small interaction effects with steel, surface finish,
viscosity, and slip were observed.
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, p 276
4-45. 52100 EF Steel: Rolling Ball Fatigue; Effect of Oil Additives

SOO , - - - - - - - - - - - - - - - : : - - - - - - - - - - - - - - , 3 4 5
400
(OMe bl & 68'"
2.76
BASI 011 OA'"
300
2.07
vi
~
<,
D<
Co
"'4
::c
"
"
"
z
e
200
1.38
.}Is-,f....
:>
~o
o ()"''''
)(
ot
""'..
-,
-----
'HASiIllOAU.
'I(DI(1I0 BY MUllIPll UQRfUION ANAlYSIS
10
100
l,. LIFE. 10 6 CYClES
Comparison of stress/life data for the mineral oil with and without the
ZnDDTP additive in surface fatigue; 85 percent confidence bands for the LSD
life estimates are shown and compared with the stress/life relation predicted
from regression analysis.
The synthetic and mineral oil no-additive conditions had about the same
life. However, the life at all stress levels tested was significantly reduced for
the mineral oil with additive below that without additive, by almost a factor
of three at the L so level, further indicating a detrimental effect of the
ZnDDTP additive on life.
Both the synthetic and the mineral oil tests had lives almost two orders of
magnitude below the standard Lundberg-Palmgren calculated life. A life reduction factor is used with the Lundberg-Palmgren theory when applied to
rolling bearings having high contact angles and thus high slip; but rarely does
the slip at bearing contacts approach that level used in these tests, so it is not
surprising that the life reductions observed are much greater than the life
reduction factors normally used for bearings.
The stress/Iife plot shown above is particularly revealing. There is no
doubt that the stress / life slope for the additive oil is significantly steeper than
for the base stock, which seems to approach the Lundberg-Palmgren theory
in stress/fife slope except for the highest stresses where it is even shallower.
Source: w. E. Littmann, B. W. Kelley, W. J. Anderson, R. S. Fein, E. E. Klaus, L. B. Sibley and W. O. Winer, "Chemical Effects or
Lubrication in Contact Fatigue-Part III: Load-Life Exponent, Life Scatter, and Overall Analysis," in Source Book on Gear
Design, Technology and Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 285
127
128
4-46. 52100 Steel: Carburized vs Uncarburized

99.
0
ILl
...J
95.
90.
80.
70.
60.
50.
40.
30.
ex 20.
IL.
...
Z
ILl
10.
a:::
ILl
a.. 5.0
4.0
3.0
2.0
I. 0
L..-_-'------'-----'----'''--'--'l..J...J...L--_-'------'-----'----'''--'--'LJ...LJ
10
100
MILLIONS OF STRESS CYCLES
Rolling Contact Fatigue Tests

Bar specimens, 0.973 em (0.383 in.) in diameter, about 8 em
long, were machined from spheroidize-annealed 52100 steel.
Two pieces were copper plated to prevent carburizing, then,
along with two unplated pieces, were austenitized at 815C
(1500 OF) for two h in a carburizing atmosphere, oil-quenched
and tempered for I.5 h at 175C (350 OF). After finish grinding
to 0.953 mm (0.375 in.), pieces were fatigue tested using a Polymet Model RCF-I testing machine with a computed maximum
hertzian contact stress of 503 MPa (729 ksi).
A Weibull plot, shown above, of the 16 tests on each type of
specimen shows that pieces with a carburized surface had a fatigue life about 50% longer at all failure rate levels than pieces
which were subjected to the same thermal cycle, but not carburized. The nonparametic Walsh test for statistical significance
indicated at a 99.5% level of confidence, the two batches of fatigue test data came from different populations.
Source: C. A. Stickels and A. M. Janotik, "Controlling Residual Stresses in 52100 Bearing Steel by Heat Treatment," in Residual
Stress for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, 1981, p 34
4-47. 8620H Steel: Carburized; Results From Case and Core
CASE---CORE---CARBURIZED----
10'
10
10'
REVERSALS TO FAILURE,
Summary plots of total strain amplitude: reversals-tofailure data for simulated case, simulated core, and carburized materials.
Carburized material is seen to have low-cycle fatigue resistance intermediate between the simulated case and core material,
a common intersection with simulated core material at intermediate lives; and in the long-life regime, carburized material specimens are more fatigue resistant than either simulated core or
case material specimens.
Plotting the strain-life curves for both case and core simulated
materials on a common set of axes , as shown in the above chart,
reveals an interesting feature. It has been observed that curves of
these materials intersected at a life of approximately 2.NJ = 105
reversals. This is in agreement with the results of this investigation. Intersection of the life curves for simulated case and core
materials accounts for a shift of failure location in carburized
components.
Source: J. M. Waraniak and D. F. Socie, "Cyclic Deformation and Fatigue Behavior of Carburized Steel," in Wear and Fraclure
Prevention, American Society for Metals, Metals Park OH, 1981, P 249
129
130
4-48. 8620H Steel: Effect of Variation in Carburizing Treatments
300
Single --.........
reheat
.........
-------~rr"~
<lID
- /""":
Direct quench ---./
100
.0
L....---L--L....L..JL...U.UJ..._...L......l-JL...L.I..LU.l...----l.---l.....L...L.L.I........ _-'--'-.L..L.J..L.LJ.J---J
103
105
107
Cycles to failure
The above S- N curves show results of a study of the effect of

martensite morphology, including the effects of micro cracking
on fatigue resistance of a carburized 8620 steel. These specimens, which were directly quenched from the carburizing
temperature, had the coarsest structure and the highest density
of microcracks, some of which were directly exposed on the
specimen surfaces bychemical polishing. The single-reheat specimens had a finer austenite grain structure and therefore finer
martensite plates and a lower density of microcracks. Since the
retained austenite content and hardness profiles of the direct
and single-reheat specimens were identical, the improved fatigue
resistance of the single-reheat specimens is attributed to the
smaller size of the microcracks and their lower density in the
finer structure. The best fatigue resistance was shown by the
double-reheat specimens.
Source: George Krauss, Principles of Heat Treatment of Steel, American Society for Metals, Metals Park OH. 1980, P 264
131
50
....
_-- .... _--
~
~
--
-- .... --
.THO~l'\IrRr NITRIDw TNr.N

GROUND TO RJ>1OVr COMPO''''"
lAnR
AnlOSI'IlERE
NITRIDED
:;;
iO
QlJ[HCHED AHD TD4PERD
AT 1010 F (\6\ C)
tHJKBER OF CYCUS
10
S-N curves for 8620 steel; nitrided versus not nitrided (quenched
and tempered only), showing stress versus number of cycles for
Source. J. A. Riopelle, "Short Cycle Atmosphere Nitriding,"in Source Book on Nitriding, American Society for Metals, 1977,p 286
132
4-50. 8622 Steel: Effect of Grinding

EFFECT OF GRINDING BURN
C/)
40
~ 30
...... 20
10
o
o
o
.....
...
"':'.:.:-. - _
. ....
-.-.-.-.
8622 MATERIAL
............... .
...................
....
..... ............
........
GRINDING WITHam BURl, - - . - '"

SEVERE DAHAGE BY GRINDING
(REFER FIG. (~4))
10
10 2
10 3
10 4
10 5
CYCLES TO FAILURE
106
Influence of grinding quality on fatigue properties of carburized differential cross.
C/)
z
......
I
P-
o
0
0
.....
t::l
40
30
20
~~
"2'::.".::-
....-::-.:.......
10
8
6
8822 MATERIAL
.....::: ........
...... ........
........ ::: .
~
'"
4
2
GRIND AFfER H. T.
GRIND BEFORE H.T.
- . _.-
Z
......
ffi
~
10
10 2
10 3
10 4
lOS
CYCLES TO FAILURE
10 6
Influence of grinding sequence on fatigue of carburized

differential cross.
As shown in the above charts, grinding has an important influence on fatigue.

Elimination of grinding damage resulted in drastic improvement in fatigue performance (upper chart). However, it was also determined that a high-quality ground
part gave better fatigue performance than when the carburized surface was left
unground (lower chart).
American Society for Metals, Metals Park OH, 1979, pp 91-92
4-51. Cast 8630 Steel: Goodman Diagram for Bending Fatigue

LEGEND
-. - 0 -
1.0
0.9
OoB
X-
- a-
CAViTIES
CAST STEEL - SOUND
R.R. MOORE FATIGUE SPECIMEN I UNNOTCHED I
R. R. MOORE FATIGUE SPECIMEN (NOTCHED 0.001!! in. R I
CONTAINED WITHIN THE BAND ARE THE
FOLLOWING DISCONTINUiTIES
WELD -INCOMPLETE PENETRATION
WELD - UNDERCUT
WELD - SLAG
WELD - MACHINE - SOUND
AS WELDED - SOUND
SLAG INCLUSION
HOT TEARS
0.7
0.6
0.5
l:
l-
l:
I-
e> ~ 0.4
Z
III
III
Ul
:n
0:: 0::
I-
;.:.:.:=.::.::.....:.:.:=--=-""-=~
III
III
.J
e>
~O.2
=>
lL
RANGE
0.3
III
I-
0.1
o.s
MEAN
- 0.1
STRESS
TENSILE STRENGTH
- 0.2
- 0.3
- 0.4
Goodman diagram for bending fatigue for normalized and tempered cast 8630 steel.
Data here show that severe discontinuities lower the fatigue strength of
cast steel. However, it will be observed from the Goodman diagram above
that the results of the notched [0.0015 in. (0.0381 mm) radius] R. R. Moore
fatigue specimen fall below those of the other bending fatigue values. Goodman diagrams for torsion fatigue and for a quenched and tempered heat
treatment show similar conditions with the notched fatigue values below the
surface discontinuity values. In many cases, therefore, design, based upon
notched R. R. Moore fatigue data, introduces a safety factor.
It must be remembered that the discontinuities were very severe and exceeded all ASTM classes of nondestructive inspection standards. The allowable discontinuities described in the ASTM standards are therefore expected
to exert a somewhat less damaging effect on fatigue behavior.
Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed.. Steel Founders' Society of America, Rocky River OH, 1980,
P 15-32
133
134
4-52. Cast 8630 Steel: Effect of Shrinkage

0.6
,--------,----,--------,r------,
END.
UTS
0.5
(/)
(/)
!.!l
(~)
138
(951)
138
135
137
(951)
(931)
(945)
~
0.17
0.13
0.26
0.28
g:(/)1-::c 0.4
o
WZ
::::>W
00::
i= Ii;
It
0.3
W
0-.J
I:::=0CLASS 2 SHRIN K
6.
z_(/)
~ ~ 0.2
""
t.
SHRIN~~A~~
Zl-
CLASS 6
0::
( EXTENDS TO ~URFACE )".
I-.J
<3:
~.
(SUB SURFACE)
flLASS 2 SHRINK
6. (EXTENg~ SURF.)
--.-~
0.1
NO FAILURE
0'--------'-----...1.--_ _---'
104
10 5
106
10 7
CYCLES
----'
TO FAILURE
Effect of shrinkage on plate bending fatigue of quenched

and tempered cast 8630 Ni-Cr-Mo steel.
As shown in the chart here, plate bending tests (completely

reversed tension and compression with cast-to-size specimens)
oflow-alloy cast 8630 steel indicate only minor effects of Class 2
internal shrinkage.
Source: Steel Castings Handbook, 5th Edition. Peter F. Weiser. Ed., Steel Founders' Society of America. Rocky River OH, 1980,
P 15-30
4-53. Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue

0.5
...------,-------r-----r--------,
I
I
I
TENSILE STRENGTH
(f)
(f)
0.4 -
.84 - 91 ksi
(579 - 627 MPa)
OCr
t;;t;
oc Z
w
<1
w OC
0.3 -
rt;;
(f)
~....J
Z(f)
CLASS 6
zlOC
W
I....J
<1
NO FAILURE
....
NO. 15
NO. 13
CLASS 6 SHRINK
x-x
NO. 13
CLASS 2 SHRINK
0-0
0.1 r-
~~SHRINK
SHRINK-./
-1-..
......
0.2 I-
~~
rSOUND
x I
<.
X~
0
. . . . . . . . . ~- 9t,--:;, O'::CLASS
0 """"--
SOUND
CYCLES
TO FAI LURE
Effect of shrinkage on torsion fatigue properties of annealed cast 8630 steel.
Source: Steel Castings Handbook. 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 15-31
135
136
4-54. Cast 8630 Steel: Effect of Shrinkage on Torsion Fatigue

0.5
,....------r----~----...,__---___.
T ENSILE
en
en
0.4 I-
W
lr
. . XC
o ,~?...
o Xo ..... _
0
o 0 o~--o_
<t W 0.3 ~
W lr
II-
<.
en en
X_X_
o~ 0
00 - - - 0 _
~--1
in
(917-951 MPo)
....
lr
STRENGTH
133-138 ksi
I-I
en I-
0.2 I-
j::Z
<tW
NO FAILURE
ZI-
0-0
NO. 15
SOUND
lr
W
0--<>
NO. 12
CLASS 6
SHRINK
x-x
NO. 12
CLASS
SHRINK
~
<t
0.1 I-
10 5
10 6
10 7
CYCLES
TO FAILURE
Effect of shrinkage on torsion fatigue properties of water

quenched and tempered cast 8630 steel.
p 15-31
4-55. Cast 8630 Steel: Effect of Shrinkage on Plate Bending
en
en
W
a::
I-J:
en t;
6,
SHRINK -----
0.4
( SUB SURFACE 1
~...J
Z
_en
I-Z
<l:W
Z I- 0.3
a::
STEEL
I-
...J
<l:
AXSOUND
CLASS 2
~a::
Lt
0.5
WZ
:::>W
~tn
l:i.-,-------,-----,-------.
0.6
0.2
L..-
II
14
o_"'\..
6 "'0'0
6''00,
0
UTS
~ I MPo 1
83
B4
..J.......
(5721
(5791
----'
'--6_
0_
END.
RATIO
.35
.32
.........
.....J
104
CYCLES TO FAILURE
Effect of shrinkage on plate bending fatigue of normalized and tempered cast 8630 Ni-Cr-Mo steel.
P 15-31
137
138
4-56. Cast 8630 vs Wrought 8640

600,--------.----------,-----------,
Notched
Unnotched
Wrought 8640
0
80
Cast
8630
'"
'"
5001--------,,------=""""'ct----------t---------;
.0;
0:
60 "
::i 400 1-----------+-----2-"""=-+-~---------l
~
~
~
:;;
IJ)
300 I-----------+"""-~""'_=------t------------j
40
Normalized and
tempered to 220 HB
0.1
10
Millions of cycles to failure
600
Ouenched and
80
tempered to
286 HB
500
.0;
c,
sc
:;;
60 ~.
400
U5
U5
300
40
200L----------'----------'-0.01
0.1
-----J
10
The fatigue limit for smooth-machined specimens is generally

about one half the tensile strength, but is reduced considerably
by notches or a rough cast surface. The S- N curves in the graphs
above compare wrought 8640 and cast 8630 steel in two different
conditions of heat treatment. In both of these comparisons, the
wrought 8640 is superior, but the two steels are practically identical in the notched fatigue test. This is significant because most
articles fabricated from either wrought or cast steel contain
more than one notch and more than one type of notch.
Park OH, 1978, p 397
4-57. 8630 and 8640 Steels: Effect of Notches on

Cast and Wrought Specimens
Tensile strength. ksi
80
700
100
120
140
160
r---.-,-----..------.-.---.----.----.------y-----,
100
600
Cast steel
I---I-------Ir----t-----t-----..~!i'i~_+--___l
80
5001---1--
"'----+---t----j
60 ]
"
.~
u.
300
40
20
100
L -_ _-'---_ _-'----_ _--'---_ _--'---_ _-'---_ _-'---_ _---l
500
600
700
800
900
1000
1100
1200
Tensile strength, MPa
The effect of notches on fatigue limit is apparent when comparing similar wrought and cast steels with regard to fatigue
limit at selected static tensile strength levels; note curves above.
Park OH, 1978, P 397
139
140
4-58. Nitralloy 135 Steel: Effect of Nitriding
60
.....
..........
..... ......
.... .....
...... ......
ATHOSPHEAE NITAIDED THEN

CADUND TO RDlDVI:
CDHPOUND lAYER
---
AntDSPHERE
NITRIDED
50
40
QUENCIlED AND mtPERED
AT 1010
r (16' C)
MlHIER OF CYCLES
10
S-N curves for aluminum-bearing nitriding steel (Nitralloy 135),

gaseous atmosphere nitrided versus not nitrided (quenched and tempered only), showing stress versus number of cycles for completely reversing torsional fatigue.
Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals. 1977.p 287
141
4-59. AMS 6475: Effects of Welding
120
_110
V)
e,
':I. 100
If)
(/) 90
lJJ
~
AMS b 4-75 (CE.VM)f ~TI r;.UE

R:.R Me OR 1:.- CSMOO Ir-H i)Pec,
"
\ "",
----- ;OR~
'\.'~ "
WEWDE't>+ST~~~!l
<, "--
" '.,
"
80
- - - - ~~ \~EL.\:)ED ( E.B')
.......
~eL..
i
--+ -- ---- -
i"o .. _
~--
IO~
106
107
10 8
CYCLES TO FA'LUR~
Fatigue strengths for case-hardened materials as well as
through-hardened may be satisfactorily defined using the R. R.
Moore rotating specimen test. The smooth unnotched Moore
specimen is ideally suited for studying many of the effects of
manufacturing and processing variables upon fatigue endurance. An example of the use of this testing technique in the evaluation of electron beam welding and postwelding aging effects
upon the endurance limit of basic AMS 6475 material is shown
in the above S-N diagram.
Source: Charles W. Bowen, "Review of Gear Testing Methods," in Source Book on Gear Design, Technology and Performance,
142
4-60. Medium-Carbon, 1Cr-Mo-V Forging: Effect of

Cycling Frequency
600
500
'E,'! 0.75 %
400
300
w 500 .-..---r-.-n'TTTT"--'--'-rTT'1~---.-.--rTTTTrr-....,.-;".,."""""
a::
::>
--l400
g 300
(f)
w 200
--l
U
t>
100
L-..L......L--L.L..I.lllL----'--l-L..L1.JL.U.L...-..L.......L.L.L.U.Lll....---'---'L.J...J.J..WJ
~:] ~': ;,:,~:~ ::~ ::::' : : : :J

10- 2
10- 1
1.0
10.0
100.0
FREQUENCY - CYCLES PER MINUTE
Influence of cycling frequency on the fatigue

properties of forged lCr-Mo-V steel at 1049 OF
(566C); no dwell period.
Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH. 1980,
P 15-55
4-61. EM12 Steel: Effect of Temperature on Low-Cycle Fatigue
1= 1lef"rmal:ion ""'90-
I-
li'T
(%)
I-
II-
-~
I-
1550
O/ll---I---+-+-H--+---+-+-H--+---+----I---If-+--+---+-I-H--t---+---I-l--l
I-----I-++t--+--+--HH----I--+--t--t-+---t----t-+--HI--+---t--t-t--l
10
10~
Low-cycle fatigue ofEM12 at 20 and 550C (68 and 1020 OF).
As holds true for other ferritic steels, the effect of hold time in
compression is slightly detrimental to fatigue life.
Source: Philippe Berge, Jean-Roger Donati, Felix Pellicani and Michel Weisz, "Properties of EM 12." in Ferritic Steels for HighTemperature Applications, Ashok K. Khare, Ed., American Society for Metals, 1983, p 114
143
144
4-62. Cast O.5Cr-Mo-V Steel: Effects of Dwell Time in

Elevated- Temperature Testing
~
+,'
3.0 , - - - - - r - , . - , - - , - ,r-tr-'r-r-r-t-t-t-r-r-r-r-r-r-r-r-r-......-r-r-r-rr-r-r-r-r-r-r-r-r--r
o
"J
-l 1.0
REVERSE BENDING
MATERIAL A
MATERIAL B
MATERIAL C
PUSH - PULL
MATERIAL
LL
............CONTI NUOUS CYCLE
t:l
Z
<l
0::
LL
-l
<l
I 0.1
L-_----'-_..L-....l-..I......JL..LLJ...JL-_-'-_-'---...l......l.....J.-L-L..LJ--:-_-'
10 2
10 3
CYCLES TO FAILURE Nf
'f--
___
-6-----6_
.....
z
~ 0.5 % STRAIN
W
0::
-6----_6
::J
-l
100
f0-
--- A-=
~ 1.5 % STRAIN
g
(f)
.J
>U
I
1.0
10
DWELL PERIOD - h
Effect of dwell periods on fatigue characteristics of
low-alloy cast steel.
As the upper diagram shows, when a D.5Cr-Mo-V steel was

tested at 1022 F (550 0C), a 20% drop in fatigue life in reverse
bending resulted when a D.5-hour dwell was added to each cycle.
The lower diagram shows that extended dwell periods, up to 10
hours, have relatively little additional effect beyond that induced by the D.5-hour dwell.
pp 15-56 - 15-57
4-63. Cast 0.5Cr-Mo-V Steel: Effect of Environment

at 550C (1022 OF)
o"e
..,'
"
REVERSEO BEND TESTS

0.
IN AIR
+ IN STEAM
x IN VACUUM
-l 1 0
PUSH - PULL TESTS

IN AIR
~
u,
o
W
o
Z
a::
I '-::2--'-----'-...L..JL....L..L.L.L-'--;:---'-----'-...L..J'--'--'c...L.L'--;---'---'--'--'--'-'..........
10
CYCLES TO FAILURE - N
Fatigue endurance behavior of cast 0.5Cr-Mo-V steel at

1022 of (550C) in air, steam, and vacuum (no dwell
period).
Source: Steel Castings Handbook. 5th Edition. Peter F. Weiser. Ed . Steel Founders' Society of America, Rocky River OR 1980,
p 15-55
145
146
4-64. Cast C-O.5Mo Steel: Effect of Temperature and Dwell Period

on Cyclic Endurance at Various Strain Amplitudes
TEMPERATURE 10 4
100
200
300
c:---,-----r---,-----.--....--...----,
103
If)
0.6 %
W
...J
U
0.7 %
1.0 %
W
U
Z
<t
1.5 %
=>
0.5 %
0::
w
10 2
1.0 %
CONTINUOUS CYCLE
30 MIN. DWELL TIME
10 '--_ _. L -_ _- ' -_ _...L-_ _--'-_ _- ' -_ _- ' - - - '

400
600
800
1000
200
1200
TEMPERATURE - OF
Influence of temperature and dwell period on the cyclic
endurance of C-O.5Mosteel at various strain amplitudes.
Source: Steel Castings Handbook, 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980
pIS-55
147
5-1. HI-FORM 50 Steel vs 1006
~ 0.010
....::::;
::l
o HI-FORM 50
I STRAINED
t.1006
lAND AGED
...
::E
<I:
Z
<I:
....
~
VI
.....
u
>u 0.001 '-3
10
---'10
....L.-:-
105
REVERSALS TO FAILURE, 2Nf
Strain-life data for AISI 1006 and HI-FORM 50 (a columbium-bearing

HSLA steel) in the strained-and-aged condition.
Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention,
American Society for Metals, Metals Park OR, 1981, P 214
148
5-2. HI-FORM 50 Steel vs 1006: Stress Response
0500L
~400
~1.FORM50
o-o-~
~~
~
0.0065
"-"-~C)oO-OC>-_~0.002
gJOO
~
1006
STRAINED AND AGED
OL.-----.L..;-----...I..::------'--;:------L-;------'
2
3
1
10
10
CYCLE NUMBER, N
10~
Stress response of strained-and-aged AISI 1006 and HIFORM 50 steels.
The imposed constant total strain amplitudes are indicated on

the graph. The degree of softening of these two steels is less
compared with that of dual-phase steels, which simply reflects the
significantly lesser degree of strain hardening of the 1006 and
HI-FORM 50 compared with the dual-phase steels.
5-3. HI-FORM 50 Steel Compared With 1006, DP1 and DP2

500
3=
.......o
U
cyo-~
400
0
II>
~~ 200
~:;;
DP 2
~HI-FORM50
a~ 300
>Uv;
..:
~
_000-0 DP 1
/0
'-1006
100
STRAINED-AND-AG ED
II>
OL...-_...L-_--'-_----'L-_...J..-_--'-_ _
.002
.004
.006
.008
.010
CYCLIC STRAIN AMPLITUDE
Comparison offour steels: AISIl006, HI-FORM

50 (a columbium-bearing HSLA steel), a lean
phosphorus-bearing dual-phase HSLA steel
(DPl), and a carbon-manganese dual-phase HSLA
steel (DP2).
Here it can be seen that all three high-strength steels offer

substantial increase in load carrying capacity at the same gauge
when compared to the plain low-carbon steel. This confirms the
potential for gauge, and consequently weight, reduction that
can be realized from the use of higher-strength steels.
149
150
5-4. HSLA vs Mild Steel: Torsional Fatigue
300
\,
\\.
~ -.
~
I
0: 1'-.~"" " .
~~
I
spc(o,81)
spc(O.81)
-'-0-- IAI'I'C4011.111
..06-- I APFC45It.Ol ISPcIO.81
--0-- APFC5O(1.01 SPC(O,811
IWIK
SPC(1.21)
SPC : MLD STEEL

APFC: IIGH STllEHOTH STEEL
: THICKNESS(mm)
I
-.......:;.r......:.~ ~.
...
~ ~.'- _
D
50
10'
--~
10'
NUMBER OF CYCLES
SoN curves showing torsional fatigue of automobile frame steels.
To determine whether the foregoing basic test results apply to

the frame models, experiments were conducted. The above
chart presents the torsional fatigue behavior of the frame models fabricated with the mild steel (0.8 mm) and each ofthe three
high-strength steels. In the high-stress, low-cycle range, fatigue
strength differs with the class of high-strength steel but virtually
no differences of that nature are seen in the low-stress, highcycle range. The three high-strength steel combinations showed
virtually the same torsional fatigue strength values as those of
the mild steel (1.2 mm) combination, indicating the possibility
of gauge reduction.
Source: M. Takahashi. "Criteria of High Strength Steels for Applying to Automobile Frame Components," in HSLA SteelsTechnology & Applications, American Society for Metals, Melals Park OH, 1984, P 498
5-5. Proprietary HSLA Steel vs ASTM A440

0.03 .-----,----,----T""""--T""""----,
0.01
t---=~~f___t--_t--_+----l
.,
"tl
.e
Q.
E
'"
e
en'i!
Proprietary HSLA
690 MPa (100 ksi] min UTS
0.001
0.0004
t---t------jt---t---t--~
L..-_~L..-_~L..-
102
103
104
_ _L -_ _.L-_
_____I
105
Cycles to failure
Total strain versus fatigue life for ASTM A440

having a yield strength of about 345 MPa (50 ksi)
and for a proprietary quenched and tempered
HSLA steel having a yield strength of about 750
MPa (110 ksi),
Park OH, 1978, P 672
151
152
5-6. Comparison of HSLA Steel Grades BE, JF and KF for Plastic

Strain Amplitude vs Reversals to Failure
2.0
1.0
t
'ill-
~N
ui
0.1
/.Aj-pl= 1.871(2Nf)-o8396
::J
e,
R 2= 0.968
::iE
-e
:(
a:
Ien 0.Q1
i=
en
:5e,
BE(Cb-)
JF(Cb-V)
KF(Cb-V-SI)
0.001
10 2
10 3
10 4
10 5
10 6
REVERSALS TO FAILURE. 2Nf
10 7
Plastic strain amplitude vs reversals to failure for Cb (BE), Cb- V

(JF) and Cb-V-Si (KF) steels.
For plastic strain-life relationship the statistical analysis indicates that there are no significant differences between the three
steels (F-ratio is not significant). This is further illustrated in the
above chart, where all the plastic strain data are plotted as a
function of reversals to failure. It is clear that a single straight
line can adequately describe all the data. Such a regression line is
drawn as the solid line in this chart.
Source: Shrikant P. Bhat, "Influence of Composition Within a Grade on the Fatigue Properties of HSLA Steels," in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 588
5-7. Comparison of HSLA Steel Grades BE, JF and KF for Total Strain
Amplitude vs Reversals to Failure '
.,.
w
::::II-
::::i
e,
:::l:
<
<
a:
I-
en
...J
-e
I-
8
12 0.1'------'----'----'----'-----'
10 2
10 3
10 4
10 5
10 6
REVERSALS TO FAILURE. 2N f
Total strain amplitude vs reversals to failure for Cb (BE),
Cb-V (JF) and Cb-V-Si (KF) steels.
Strain-life behavior: The strain-life curves for the three steels

are compared in this graph. It is clear that when plotted as total
strain versus reversals to failure, the three steels behave similarly
and the differences between them are minor.
Source: Shrikant P. Bhat, "Influence of Composition Within a Grade on the Fatigue Properties of HSLA Steel," in HSLA SteelsTechnology & Applications, American Society for Metals, Metals Park OH, 1984, P 587
153
154
5-8. Comparison of a Dual-Phase HSLA Steel Grade With

HI-FORM 50: Total Strain Amplitude vs Reversals to Failure
~ 1.0
w-
...='::::;
DUAL PHASE 1
16
AS-RECEIVED
STRAINED AND AGED
HI-FORM SO - - -
".
:e
...'"
'"
...u
u
>- 0 . 1 ' - : : - - - - - - - - - - - ' - - : - 4

u
103
10
.1--:10
__
Total strain amplitude versus life data for DPI (a lean-phosphorus

HSLA steel) in the as-received and strained-and-aged conditions. Data
for HI-FORM 50 (a columbium-bearing HSLA steel) are included for
comparison.
American Society for Metals. Metals Park OH. 1981, P 213
5-9. AISI 50 XF Steel: Effects of Cold Deformation

8
'?
...
IJo
tJ>
Eeff
I(
0.2
0.4
00.6
.r
w
0
00
I-
::::i
e,
cta:
I-
CIl
...J
<l:
l-
e
I103
104
106
105
8
Eeff
'?
...
0
I(
.r
0.20
00.30
6
0.57
!:
...J
e,
<l:
ct
lii
a:
...J
<l:
2
0
I-
o
I103
104
105
106
Total strain amplitude versus reversals to failure for AISI 50 XF

HSLA steel. Upper chart: after balanced biaxial stretching;
lower chart: after cold rolling.
Although the effects of prior deformation by BBS or CR on the

strain-life behavior of 50 XF were generally similar to those in 1006,
some specific differences were apparent; for example, the effect of
prior deformation was stronger for 50 XF than for 1006 in that both
the decrease in life at large strain amplitudes and the increase in life at
small strain amplitudes were greater in 50 XF than in 1006.
Source: John M. Holt and Philippe L. Charpentier, "Effect of Cold Forming on the Strain-Controlled Fatigue Properties of HSLA
Steel Sheets," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 218
155
156
5-10. AISI 80 OF Steel: Effects of Cold Deformation
po
)(
-r
::l
I-
:::i
a-
'lI
Effective Str8in 8nd Mode

of Deformation
4
60.0
'90.06
a 0.08
00.16
1 Uniaxi81
T8nsion
881. Biaxial
Stretching
Solid Symbols-D8ta for

Specimens Tr8nsverse to
Hot Rolling Direction
II:
In
..J
~
Runoutl
I-
Strain-life curves after deformation for AISI 80

DF HSLA steel.
In this steel, the fatigue life appeared to remain unchanged or

to increase very slightly as a result of deformation, at least for
the effective strain levels investigated (see graph). Also, the fatigue life appeared to be unaffected by the mode of deformation
and the specimen orientation.
Source: John M. Holtand Philippe L. Charpentier, "Effect of Cold Forming on the Strain-Controlled Fatigue Properties ofHSLA
Steel Sheets," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, p 218
5-11. Comparison of Three HSLA Steel Grades, Cb, Cb-Vand

Cb-V-Si: Strain Life From Constant Amplitude
2,..,-----------------,
w
c
:::>
I-
::::i
0.
::l!i
2,-----------------,
0.1
:::>
0.1
I-
::::i
0.
::l!i
4(
TOTAL
.PLASTIC
AELASTIC
4(
4(
a:
:;:
I-
en
a:
I-
en
0.01
0.01
10 3
10 4
10 5
10 6
REVERSALS TO FAILURE, 2 N f -
10 7
Strain-life curves for the Cb steel.
0.001 '-----''---------'-----'-----'-----'
10 4
10 5
10 2
10 3
10 6
REVERSALS TO FAILURE, 2 N f -Strain-life curves for the Cb-V steel.
0.1
l!f
E
..J
~
4(
Z
:;:
a: 0.01
lii
.1OTAL
PLASTIC
AELASTIC
Strain-life curves: Strain-life data from constantamplitude tests for the three steels are plotted in the
three charts here respectively in the form of total
strain amplitude versus the number of complete
reversals to failure.
0.001'----'----'----'----L....:..---'
10 2
10 3
10 4
105
10 6
RI;VERSALS TO FAILURE, 2Nf
Strain-life curves for the Cb- V-Si steel.
Source: Shrikant P. Bhat, "Influence of Composition Within a Grade on the Fatigue Properties of HSLA Steels," in HSLA
157
158
5-12. Comparison of Stress Responses:

DP1 vs DP2 Dual-Phase HSLA Steels
------ =........... --
600
_11
1
_
'==... - ....
.............
:E
:::
...
:-..
"'- ...:=:.-~
.
.....
~_
--;::J"
..- 0.0065
-.:-----0.003
~ .. 009--o_-.>-""""_ _ ~-0.0065
0' 2
...~boo
~
v
0---
0.0025
~200
STRAINED AND AGED
0'-::10
'-;-
'-::-
'-::-
---'''--:-
--1
10'
Comparison of stress response of strained-and-aged DPl (a lean

phosphorus HSLA steel) with that ofDP2 (a carbon-manganese HSLA
steel) for the total strain amplitudes indicated.
Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention.
American Society for Metals, Metals Park OR, 1981, p 209
5-13. Dual-Phase HSLA Steel Grade: Stress Response for

As-Received vs Water-Quenched
..;
0.005
6
u
~ 400'r-o-OCo--o-----..,.--<-->--e..-"d""--ov-...o-O0.002
'"
~
o
~300
v
::::;
v
>-
v200
Ol.I
..L..,-
....L..:--
........-::-
--'-:-
--'
10 2
10 3
CYCLE NUMBER, N
Stress response of a water-quenched dual-phase steel in the as-received

condition for total strain amplitudes of 0.002 and 0.005.
Source: N. Lazaridis and S. P. Bhat, "Fatigue Behavior of Cold Rolled Dual Phase Steels," in Wear and Fracture Prevention.
American Society for Metals, Metals Park OH. 1981, P 208
159
160
5-14. Dual-Phase HSLA Steel Grade: Stress Response for

As-Received vs Gas-Jet-Cooled
.500
Go
:e
..__
E.. oo
=
~
'"
~-----o.--o_o(a
o
3 0 0 r - - o - - _....-
I~
....-
0.006
...._ -....._-"''-oA.-o''-_..o-_ _....._0.0025
~hoo
.
..>
>v
Cl00
C
OL-
----''-;-
----'-:;-
--l.-:;-
--l.-:;-
---'
CYCLE NUMBER, N
Stress response of the gas-jet-cooled dual-phase steel in the as-received

condition for total strain amplitudes of 0.0025 and 0.006.
American Society for Metals, Metals Park OH. 1981. P 208
162
5-16. Comparison of Dual-Phase HSLA Steel DP2 With HI-FORM 50

0
0.010
DUAL PHASE 2
1'"
AS-RECEIVED
STRAINED AND AGED
HI-FORM SO - - - -
0.001L..:---------"---=,.-10 4
10 3
...L..,,---
..J
10 5
Strain-life curves for DP2 (a carbon-manganese HSLA steel) in two

conditions compared with HI-FORM 50 (a columbium-bearing HSLA
steel).
American Society for Metals, Metals Park OH. 1981, p 214
164
5-18. Fatigue Crack Propagation Rate: Effect of Temperature for

Two HSLA Steel Grades
HSLA-1
G.S. =lOfLm
R =0.1
o 300K
t> 233K
Cl 173K
o 123K
Q)
!'/;f/:/.
10 ,
'/:
>.
o
"E
I:
g' i
t> 233K
Q)
I,
o~
',,
(PlI
0123K~I1!I:
, I
I
<,
01:
o 173K
o
>.
o
1/ /
1//,
o 300K
1/ t>A
I/,~
R=O.1
/?~ I
0,: t>.
HSLA- 2
G.S. = 10JLrn
Prof
, 6'
01 '
.,pI
,
I ,
/ t>,
I
Pi
/,/
P,
T(K)
300
233
173
123
3.6
5.1
6.5
10.8
T(K)
300
233
173
123
14
n
7.6
8.8
12,1
15.6
18 22 26 30 36
6K, MPa -m 1/2

The effect of test temperature on the fatigue crack propagation rates in the Paris law regime for two HSLA
steels in the as-received condition.
The only significant difference between HSLA-I and HSLA-2 is that HSLA-2 contains double the
amount of Nb that HSLA-I contains (see compositions on p 165).
The effect of temperature is seen to decrease the crack propagation rate with decreasing temperature
at low values of 6.K. However, as the stress intensity increases, a crossover occurs wherein higher
growth rates were observed, as shown in the above charts. This crossover is further reflected in the
increase in the Paris law exponent, n, where it ranged from 3.6at room temperature to 10.8at l23K for
HSLA-l. The large increase is a result ofthe change in the fracture mechanism from ductile transgranular fracture to cleavage. This behavior has also been seen in iron binary alloys where n increased from
3.5 at room temperature to 20.9 at l23K.
Source: Khlefa A. Esaklul, William W. Gerberich and James P. Lucas, "Near-Threshold Behavior of HSLA Steels." in HSLA
Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, p 569
165
5-19. Effect of R-Ratio and Test Temperature on Crack Propagation

of H SLA Steel Grade 1
HSLA-1
HSLA-1
G.S. 10pm
T-300K
G.S. 10pm
T -123K
R-O.IO
R-O.IO
R-0.35
R-0.70
10'
10
10 R-
0.35
R-0.70
e.
-,
c lO
"0
II
II
II 9 10
12 14 It III 20
2
24
II
II
II 910
1214161920
6K, MPa-m l12
6K, MPa-m"
The effect of R-ratio on fatigue crack propagation behavior of HSLA-l at test temperature of 300
and 123K in the as-received condition.
Compositions of HSLA-l and HSLA-2

Alloy
HSLA-1
HSLA-2
..
.
Mn
Nb
SI
0.07
0.06
0.51
0.35
0.014
0.03
0.03
0.03
<0.005
0.01
Al
0.005
0.01
0.01
Ni
Cr
Fe
0.01
0.01
Rem
Rem
Source: Khlefa A. Esaklul, William W. Gerberich and James P. Lucas, "Near-Threshold Behavior of HSLA Steels," in HSLA
Steels-Technology & Applications. American Society for Metals, Metals Park OH, 1984, p 571
24
166
5-20. Effect of Test Temperature on Fatigue Crack Propagation

Behavior for Two HSLA Steel Grades
HSLA-1
HSLA-2
G.S. IOlJm
R-O.I
G.S. 10JLm
R' 0.1
e300K
.233 K
.173 K
.123 K
II)
u>u
.....
o
o
300K
233K
173K
123K
1I
.6K,
8 9 10
12 14 16 18 20
MPa-m"2
24
7 8 9 10 12 14 16 1820 24
6K, MPa - m112
Fatigue crack propagation behavior of two HSLA steels tested at temperatures of 300,233, 173 and
123K in the as-received condition.
The only significant difference between HSLA-I and HSLA-2 is that HSLA-2 contains twice
as much Nb as HSLA-I (for compositions of the steels, see p 165).
Near-threshold crack growth and threshold stress intensities for both steels in the as-received
condition are depicted in the above charts for all test temperatures. Comparison of crack growth
rates and threshold stress intensities at room temperature indicate that HSLA-2 has a higher
resistance to fatigue crack propagation than HSLA-l. The stress intensities amplitude, 11K, for
constant growth rates of IO- s and 10-9 ta] cycle are 2.0-2.5 MPa-m 1/2 higher in HSLA-2 than in
HSLA-l. The threshold stress intensityl1K'h' is also higher for HSLA-2 (8.0 MPa-m 1/ 2) compared to HSLA-I (5.5 MPa-m 1/2). The 2.0-2.5 MPa-m 1/2 difference in threshold and for the two
growth rates clearly demonstrates that there is an inherent difference in the fatigue crack propagation behavior of these two HSLA steels. This difference is also reflected at low temperatures,
where HSLA-2 showed lower crack propagation rates and higher threshold stress intensities
than HSLA-I. Furthermore, by comparison of threshold stress intensities for these two steels in
relation to the effect of decreasing temperature on increasing 11K/I" it was found that the ratios
of I1K,h (D t 11K,,, (300K) are the same for both steels.
Source: Khlefa A. Esaklul, WilliarnW. Gerberich and James P. Lucas, "Near-Threshold Behavior of HSLA Steels," in HSLA
5-21. Stress-Cycle Curves for Weldments of Different

HSLA Steel Grades
'II'EI.~E
""'-'~"---dE]r
~ ""
~~
~
=Jl
.~:~,:,.,
(00
~~~
lotAlERrAl
oOJtrElDl!1
I~.~
USIW 22->.....
'..po'- '.
-'ed'" , "
IoU5
I
I~
lHlER OYIW.IC
lCW> - cotI:XTJ:)N:S
S"mESS - LMTS OF WB.DED SAt.flI.ES
OSIE 3eON
5' 52.
JO~
.=2C~
on
'<:
...... ~ ~--.....--.:.
....., ~
<,
10:
.....,~
....
~ - - - f=:-~
~-..::.
00
$~ ,,~
i
NJIIB[A
CYCLES
.q
~
I--
"
Stress-cycles curves of welded samples of different materials under tension load.
"._.-"---dEl
lHlER OYIW.IC
S"mESS - LMTS OF WB.DED SAt.flI.ES
:;).('..,100
,.-1
""
,
, " ~.
i~!
I~
'00
LOAD - 00NXT10N9
WATEAW.
I~%
USIW 22J.E"'B
,'0;;',
.. tiS
T~
...,.,M"A.E- wac,
'-PO -,
OStEJOOn-l
-,
5152.3
...::""
)00
"""
."" F--",,~
",
~~
.",
---
"- 1--.0-
.";" c-c
0=
"'''''':
.-
sc.J
I
"".,.BR ~
CYCLES
Stress-cycles curves of welded samples of different materials under alternating bending load.
Fatigue data were derived from testing welded samples under tension and bending loads. It was surprising that under both types of load the HSLA steel and the
soft unalloyed steel hardly differed in fatigue strength; thus it can be said that the
use of HSLA steels is not justified if a component has a weld in the highest-stressed
area. An explanation for this is the loss of the thermal-mechanical effect, which is
responsible for increased strength, by the heat influence during the welding operation; and it is thought that a higher-strength manganese-alloyed steel, such as St
52.3 (according to DIN), the strength of which results from the chemical composition, would be more favorable in this respect.
Source: Klaus E. Richter, "Cold and Hot-Rolled Microalloyed Steel Sheets in Opel Cars-Experience and Applications,"in HSLA
167
168
5-22. Weldments (FCAW): SAE 980 X Steel vs 1006
R=0.1. toO.13" 13.3mml
--6--
SAE-1006. Smooth
-'-0-'- SAE1006. FCAW
- - . - - SAE-980X. Smooth
- - - - SAE-980X. FCAW
- ..----.-........ -..
----.1;,.1---...- __
-----6-
-'"l:>"'&--C'IS"" 6-"tl.
10 5
106

Fatigue properties of smooth and FCA W SAE 1006 and SAE 980 X steels.
The fatigue strengths of the smooth HSLA steel were higher than that of the low-carbon steel.
The I06-cyclefatigue limit stress of the smooth SAE 980 X steel was 469 MPa (68 ksi) and that
for the SAE 1006 steel was 283 MPa (41 ksi).
However, after welding, SAE 1006 and SAE 980 X steels exhibited similar fatigue properties
over the 104-1 06-cyclelife range studied. The I06-cycle fatigue limit stresses for FCAW SAE 1006
and SAE 980 X steels were between 114 MPa (16.5 ksi) and 117 MPa (17 ksi).
Source: Kon-Mei W. Ewing. Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue of TIGDressed SAE-980X HSLA Steel." in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH,
1984, P 556
5-23. Weldments (TIG): DOMEX 640 XP Steel Welded Joints vs

Parent Metal
1000
000
"00
100
r---~
__
600
Parent IIletal
sou
400
TIC-treated bull
weld
JOO
TlGtreUed fillet
weld
~ '.00
Untreated butt weld
Untreated fillet
100
50 '--
.L-
.L-
-'-
105
106
.L..-........
CYClES TO FAILURE
Fatigue strength for DOMEX 640 XP. Standard-Wohlerdiagram (log-log scale) with pulsating load (R=min
stress/max stress=O). Sheet thickness 5 mm and ultimate
tensile strength 767 MPa.
.
For unwelded parent metal the fatigue strength of a steel is

improved with increasing static strength. For welded joints the
fatigue strength in the endurance range 105-2 X 106 is mainly
dependent upon the weld geometry and is therefore roughly the
same irrespective of the static strength of the steels. For making
full use of an increased static strength for a steel subjected to
severe fatigue, special attention must be paid to the configuration of the welds. After welding, grinding or TI G-treatment can
be used to improve the weld geometry. The notch effect at the
weld toe is decreased and the fatigue properties can be improved. Another solution is to place the welds in areas where the
stresses are low.
Source: Tony Nilsson, "Formable Hot-Rolled Steel With Increased Strength," in HSLA Steels->Technology & Applications,
169
170
5-24. Weldments (FCAW Dressed by TIG): Fatigue Life Estimates

Compared With Experimental Data for SAE 980 X Steel
10
103
--.::::::::::::
CJ)
~
...
-,
(U r =-87KSI
.e'!:
en
<I
10
TIG DRESSED SAE 980 LAP-SHEAR WELDS

Kf = 2.52
J!.
:E
..'=ur = 87 KSI
10 2
R = 0.1
EXPERIMENT
- - PREDICTION
I 3
10
101
105
NT,
106
108
CYCLES
Total fatigue life estimates compared to the experimental data for the FCA WITIG-dressed
SAE 980 X steel.
It should be emphasized that life estimates made on the FCAW{TlG-dressed welds were
based on geometry changes brought about by TlG-dressing. The other beneficial effects such as
removal of slag intrusions and inclusions were not considered. The close agreement between the
calculated and observed long-life fatigue properties suggested that the majority of fatigue improvement seen in TIG-dressed joints was attributable to the geometry change. The smaller
flank angle contributed significantly to the increased fatigue strengths of TIG-dressed
weldments.
Source: Kon-Mei W. Ewing, Pei-Chung Wang, Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue of T1GDressed SAE-980X HSLA Steel," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH,
1984, P 563
en
<I
5-25. SAE 980 X Steel Weldment (FCAW): Smooth Specimen vs

TIG-Dressed vs As-Welded
SAE-980X
R= 0.1, t= 0.13" 13.3mml
--06--- Smooth Specimen
TIG-Dressed
_.-(}-.- As-Welded
---0---
~-------4..II
~o
o '-'il... 0
__ ~
-----ZP-n---..I\
~00o
0
o
,-.~
-0....
o Jgt. '1I'o-."'"b-
'[jCD-.,_.
00
0
-c._
'-
''''"00'_
10 5
,-"",..It.0
o'-'n-
10 6

Fatigue properties of FCA WITI G-dressed SAE 980 X steel compared to the smooth specimen and as-welded data, From these data, a significant improvement in fatigue characteristics
can be obtained by TlG-dressing the welds,
Source: Kon-Mei W. Ewing. Pei-Chung Wang, Frederick V. Lawrence. Jr., and Albert F. Houchens, "Weld Fatigue of TIGDressed SAE-980X HSLA Steel," in HSLA Steels-Technology & Applications, American Society for Metals. Metals Park OH,
1984. P 558
171
172
5-26. SAE 980 X Steel Weldment (FCAW): Lap-Shear Joints
en
<]
10 1
SAE980 LAP-SHEAR WELDS
Kfmax =3.49 I R = 0.1
EXPERIMENT
PREDICTION
105
106
NT, CYCLES
Total fatigue life predictions and experimental results for the FCAW, SAE 980 X lap-shear joints.
Source: Kon-Mei W. Ewing, Pei-Chung Wang. Frederick V. Lawrence, Jr., and Albert F. Houchens, "Weld Fatigue of TIGDressed SAE-980X HSLA Steel,"in HSLA Steels-Technology and Applications, American Society for Metals, Metals Park OH.
1984, p 562
5-27. Microalloyed HSLA Steels: Properties of Fusion Welds
STRESS - LMTS
. v; ...~
:.~"",.TC'SS'O\T~ST
.~
. ':._':'0
a: WELDED SN.f'lES
"-'-dEl
10
)10.5
~~
::.
"
'0(>0'
-=
lC~
nn
'l;:
I~:
'~
oi~hOl
~~
-;.'
051E 38CN
, , ",
>l
"3
:.-........-.:..: ~
......,
'-.I,!~
~ ..... -...;:..-" ~~
.............
l:SIW n."E."~'"
-.
00
--
'---
......l[R'.l1.
",-
"00',
101"
'00
iec
COIClIJ1DN8
Lp"o -
I~""
I~
lHlER OYNAlolC
w(lO~O'""
II
lA.IM8ER(JFC'fC~t'S_
Stress-cycles curves of fusion welded samples of different

materials under tension load.
STRESS - LMTS
..
".~' C' lQl.Orh(i BfK>'MjY=ST
1.C.... '00
.. n
~~:
,co
lHlER OYNAlolC
a: WELDED SN.f'lES
"-"-dEi
10
Lp"o -
COIClIJ1DN8
~~~-"'S~~n
)IQ.S
I~%
'. o~
,.11'
' 10&1 0'"
'-~
IUTERlAl
~%
USIW 22."'.',
-.
OSIE380T'4
",
SI S23
'-="
'00
""""
'" f"........C'l>~
'00
~~
".,
--~
~--'=
=---=
r-''!!::='- ..
"=
--
-II
III
I
fUo'8(A OF CVClS-
Stress-cycles curves of fusion welded samples of different

materials under alternating bending load.
Due to the preferred crack location in the welded areas, it

seemed necessary to examine the dynamic strength of fusion
welded joints of HSLA steels in more detail, testing the steel
used for the crossmember at a minimum yield strength of 380
N /mm 2, in comparison to a soft unalloyed, hot rolled steel
sheet. Tensile load and alternating bending load were the selected types for dynamic test. The above charts show the respective stress-cycle curves.
Source: Klaus E. Richter, "Cold and Hot-Rolled Microalloyed Steel Sheets in Opel Cars-Experience and Applications,"in HSLA
Steels-s-Technology & Applications, American Society for Metals, Metals Park OH, 1984, P 487
173
174
5-28. Microalloyed HSLA Steels: Properties of Spot Welds
LOAD AMPLITUDE
6P, kips
AND NUGGET ROTATION
68 N , degree
5.0
0.2
1.0
0.5
GroupC
Fatigue test results for the 2.06 mm (0.081 inch) thick

sheet with various weld diameters.
LOAD AMPLITUDE
6P. kips
AND NUGGET ROTATION
68 N degree
5.0
2.0
1.0
0.5
LOAD
AMPLITUDE
6P, kN
5.0
2.0
1.0
0.5
Group K
LOG CYCLES TO FAILURE, N,
Fatigue test results for 1.02 mm (0.040 inch) thick sheet of

different stiffnesses.
Results of spot-weld fatigue tests are presented in the four plots (above and
on the facing page) for the stated conditions. Each curve shows the load
amplitude, I1P, and nugget rotation values, 116 N' for each test as a function of
cyclic life. Straight lines were fitted through the data.
1~oe""",roOlI"
LOAD AMPLITUDE
dP, kips
AND NUGGET ROTATION
d(~N' degree
5.0
175
~::S';=I~~:"
LOAD
AMPLITUDE
dP, kN
2.0
1.0
0.5
Group E
5.0
2.0
1.0
0.5
5.0
0.2
1.0
0.5
d0
dP
Fatigue test results for variations in specimen width and

thickness.
5.0
LOAD AMPLITUDE
dP, kips
AND NUGGET ROTATION
:>SN. degree
2.0
1.0
0.5
~
-- -
5.0
2.0
1.0
0.5
"W-IOI61M11~O"1
' 02 _ " " . '
0-66In"lf02fj1fl1
::~'
GroupN
5.0
2.0
1.0
0.5
de
dP
Group 0
5.0
2.0
1.0
0.5
1,.1 0211Yn1004..., ,
W .. l01611'WT1140n,
0 .. 66 ........,0416 ... ,
dP
de
LOG CYCLES TO FAILURE. Nf
Fatigue test results for 1.02 mm (0.040 inch) thick sheet

with single and multiple welds.
Source: James A. Davidson, "Design-Related Methodology to Determine the Fatigue Life and Related Failure Mode of SpotWelded Sheet Steels," in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984, p 542
176
6-1. HY-130 Steel: Effect of Notch Radii
~::;;
800
<l 600
b
g 400
I-
:J
Q.
9.5mm
~ 200
6.4
3.2
~
tii
100
80
~ 60
0:
~z
1.6
0.80
40
0.40
0.20
20
2
3 4 5 6 8 10
3 4 5 6 8 10
3 4 5 6 8 10
CYCLES TOFATIGUE-CRACK-INITIATION, NjXI03
Cycles to fatigue-crack initiation versus nominal stress amplitude,

for notched specimens with various radii of curvature.
~a,
~:::!:
4000
3000
2000
1000
~ 800
"
~
<l
p=0.20mm
600
500
400
300
200
2
3 4 5 6 8 10
3 4 56 8 I
3 4 56 8 I~
CYCLES TO FATIGUE-CRACK-INITIATION, NI xI0 3
Same data as in upper graph but plotted versus Ill(
/.JP rather than ~a.
Curvature ofthe notch and I:1Kis the stress-intensity amplitude computed for an imaginary crack whose length is the same as the notch depth, a. Barsom and McNicol used this
parameter to compare N j , the cycles to fatigue-crack initiation, in HY-130 steel for
notches of constant depth but various radii of curvature. The results are shown in the
above graphs. In the upper graph, N, is plotted versus l:1a, where N, is defined as the
number of cycles to give a 50-J..!m side notch. There is a wide spread in the curves. As
expected, the sharpest notch, lowest p, gave the most rapid initiation at a given stress. The
lower graph shows I:1K/ vp plotted versus N;. A narrow-spread family of curves results;
these converge as the value of I:1K/vp is decreased to a threshold value I:1K/ vp I'h' the
minimum value to initiate fatigue cracks in notches.
Source: M. E. Fine and R. O. Ritchie, "Fatigue-Crack Initiation and Near-Threshold Crack Growth," in Fatigue and Microstructure, American Society for Metals, Metals Park OR, 1979, pp 256-257
177
6-2. 300 M Steel: Effect of Notch Severity on

Constant-Lifetime Behavior
Minimum stress,ksi
300
s:::;:
1500
~.
~E
200 1;;
E
:>
E
:>
.~
.~
::;:
::;: 1000
100
Notch
500
severity
(107 cycle
IifeUmel
_~00'::0:----...,."..::-----,.L..._-_--L._------"L-_-_.L-_-_--L._-----,,------..,..L...------'25~0
-1500
-1000
--500
2000
1000
1500
Constant-lifetime fatigue diagram for 300 M alloy steel, hardened and tempered to a tensile strength of 1930
MPa (280 ksi). Solid lines represent lifetimes obtained from unnotched specimens. Dashed lines represent
lifetime of ten million cycles for specimens having the indicated notch severity.
Source: Metals Handbook. 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals. Metals
Park OH, 1978, P 670
178
6-3. TRIP Steels Compared With Other High-Strength Grades
IlXXl
1200
1400
MN/m2
1600
1800
2lXXl
2200
1400
200
R 0.1
TRIP
~175
1200
u
~150
.,.
SR 4340
IlXXl"/;
~125
v;
4ll
800
:J
.2'
.l!loo
600
Fatigue strength at 10' cycles (R = 0.1) vs ultimate tensile strength for

TRIP steels compared with other high-strength steels.
Studies on fatigue-crack propagation (FCP) conducted under controlled stressintensity amplitude (~K) conditions indicate that deformation-induced transformation retards crack growth in lower-strength metastable austenites, particularly at low
~K, and also exerts a beneficial influence in high-strength TRIP steels, although to a
much lesser extent. This growth retardation may be due to crack-closure effects arising
from the transformation volume change, which may be particularly effective in the
fatigue-threshold regime. Smooth-bar fatigue properties appear to be dominated by
transformation hardening, which is desirable under stress-control conditions (reducing strain amplitude) but generally undesirable under strain-control conditions (increasing stress amplitude). In lower-strength austenites, transformation reduces fatigue life under conditions of controlled plastic strain amplitude; under controlled
total strain amplitude, transformation is detrimental to low-cycle fatigue life, but a
small amount of transformation may be beneficial at high cycles. Similarly, the lowcycle fatigue properties of high-strength TRIP steels are found to be degraded by transformation under controlled total strain amplitude. Under stress control, the fatigue life
of lower-strength austenites is greatly enhanced by transformation; for a stress ratio
(R= amin/ a m x ) of 0, fatigue limits in excess of the yield strength are observed. Investigation of the smooth-bar fatigue properties of high-strength TRIP steels at R= 0.1, in
which thermodynamic stability was varied by heat treatment, also revealed transfermation enhancement of fatigue life. Such enhancement allows the achievement of exceptional fatigue strength at high ultimate strength levels, as illustrated by comparison
with other high-strength steels in the above graph.
Source: G. B. Olson, "Transformation Plasticity and the Stability of Plastic Flow," in Deformation, Processing, and Structure,
George Krauss, Ed.. American Society for Metals, Metals Park OH, 1984, p 419
6-4. Corrosion Fatigue: Special High-Strength Sucker-Rod Material

10.000
Q,
UI
UI
80,000
50,000
....
a:
lUI
40.000
....0
::i
lL
lL
ce
30.000
AB- IN AIR
C0 - IN ACID BRINE
EF- IN ACID BRINE WITH INHIBITOR
20,000
10.000
102
10~
t04
NUMBER OF CYCLES
10'
FAILURE
10'
to'
Effect of corrosion and corrosion inhibitors on the SoN curve for highstrength steel (sucker-rod material).
After the first brittle crack is initiated, No.2 is the slow step in the
process and electrochemical action is the slowest part of this step.
Thus, the effect of corrosion can be illustrated with curves of stress vs
logarithm-of-number-of-reversals-to-failure for sucker-rod steel. Corrosion accelerates cracks propagation, so the fatigue curve drops from
AB to CD, as shown in the graph. Deceleration of the slow stage with a
corrosion inhibitor will raise the S-Nfatigue curve from CD to EF.
Source: Joseph F. Chittum, "Corrosion Fatigue Cracking of Oil Well Sucker Rods," in Corrosion: Source Book, Seymour K.
Coburn, Ed., American Society for Metals, Metals Park, OH, 1984, P 380
179
180
6-5. Corrosion Fatigue Cracking of Sucker-Rod Material
.8
Ul
~.6
Ii
0
ct:
~.4
...
0
:J:
I-
I
I
~.2
UJ
II
oJ
I
I
II
0
FIRST STAGE
SECOND STAGE
!THIRD,
ISTAGEI
0
RELATIVE NUMBER OF CYCLES, PERCENT OF FAILURE
Corrosion fatigue cracking of sucker rods.
This graph shows typical progress of a crack at high stress plotted

against number of cycles, showing stages in the fatigue process. Observations of sucker-rod crack penetration as a function of reversal accumulation are possible using a bending apparatus and a magnetic
fluorescent powder technique. Penetration vs reversal curves resemble
the one shown above when the stress is well in excess ofthe endurance
limit.
During bending, no penetration is apparent in the first 40-60% of the
specimen's fatigue life, even though intrusions and extrusions may
form earlier. A crack eventually appears and progresses through the
specimen. When the penetration reaches a certain percentage of the
cross section, the cracking accelerates until catastrophic failure occurs.
Source: Joseph F. Chittum, "Corrosion Fatigue Cracking of Oil Well Sucker Rods," in Corrosion: Source Book, Seymour K.
Coburn, Ed., American Society of Metals, Metals Park OH, 1984, P 378
181
6-6. Hydrogenated Steel: Effect of Baking Time on

Hydrogen Concentration
300
Normal Notch Strength = 300,000 psi
o.. ~
275
i---::::::::::::
o~
250
200
(f)
(f)
Q)
...
175
( /)
"0
150
Q)
a.
a.
<t
i'...
-,
~
125
100
..1
-,.
75
\
<,
Bake
---
24 h r -
Bake 18 hr
Bake 12 hr
Bake 7 hr
Bake 3 hr
Uncharged
+-+-
,~
:---:-
0\ \
\
-\ ~
+
+
"
.-
-\
~.
~~
"\\'\
a. 225
(f)
0
0
0
i'-..
Bake 0.5 hr
--
<-
1-
50
0.01
0.1
1
10
Fracture Time, Hours
100
1000
Static fatigue curves for various hydrogen concentrations obtained by baking different times
at 150C (300 OF). Sharp-notch specimens. 230,000 psi strength level.
These are, in essence, static fatigue curves, and the lower critical stress may be considered a
static endurance limit-that is, a stress below which failure will not occur for an indefinite period
of time. This behavior is sensitive to hydrogen concentration as shown above, where it may be
seen that all delayed-failure parameters-notch strength, rupture time, and static fatigue limit
increase with decreasing hydrogen concentration. Also, even after 24 hours at 150C (300 OF),
there is still a substantial stress range, of the order of 60,000 psi, over which delayed failure will
occur. In an unnotched specimen, full recovery ofthe ductility as measured by the reduction of
area can be attained in less than 20 hours at 150C (300 OF), yet delayed failure will occur after
24 hours or longer of baking time at 150 C (300 OF).
Source: Alexander R. Troiano. "The Role of Hydrogen and Other Interstitials in the Mechanical BehaviorofMetals,"in Hydrogen
Damage Source Book, Cedric D. Beachem, Ed., American Society for Metals, Metals Park OH, 1977, P 154
182
6-7. Hydrogenated Steel: Effect of Notch Sharpness

300
275
250
en
0..
0
0
0
225
T~.
I~~ 1\
'\
200
en
en
Q)
L..
175
( J)
"0
Q)
0..
0..
150
-a_
""'"
0
01'\.
~\
125
~\\
Notch
Radiu~ = 0.010 in.
It-- - - -
>--
~.
75
; a1lus
J. = 2'In.
>--
>--
Notch Radiu's = 0.020 in.
-\..
100
Notch
\\
i'\
Notch Radius = 0.25 in.
Notch Radius = 0.001 in .
.....-
50
0.01
0.1
10
100
1000
Fracture Time, Hours

Static fatigue curves for specimens of different notch sharpness. Baked 0.5 hour at 150C
(300 oF).
The variation of lower critical stress with notch severity is shown in this diagram. It is evident
that the static fatigue limit rises as notch severity (radius) decreases for hydrogen-charged
high-strength steels (using the same baking time).
Source: Alexander R. Troiano, "The Role of Hydrogen and Other Interstitials in the Mechanical Behavior of Metals," in Hydrogen
Damage Source Book, Cedric D. Beachem, Ed., American Society for Metals, Metals Park OH, 1977, P 155
7-1. O.5%Mo Steel: Effect of Hold Time in Air and Vacuum at

Different Temperatures
+ 10 I
CYCliNG
KEY
CONTINUOUS
..J
<
e
l-
....
....... ' ........
lLL
loO
l.'
<
Z IX:
IX: I-
<
VI
VI
10 0
.. =:::::.:.~.........:.~ . ... :--;:

'
30 on HOLD
"
.',~""--............:..
AIR
VAC ...
..........,,:~:-:-: ....
" .""'"
".. . . ..:r.
. "=':..,:-:. . :--;; -
I::NV. TEMP.
AIR 275 K
VAC . ..
AIR 775 K
VAC. II
--.
__
.....::--........
._~.
CYCLES TO FAILURE
Effect of hold time in air and vacuum upon the fatigue endurances of a
O.5%Mo steel at 275 and 775 K.
Source: R. H. Cook and R. P. Skelton, "Environment-Dependence of the Mechanical Properties of Metal sal High Temperature,"
in Source Book on Materials for Elevated- Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals
Park OH, 1979, p 83
183
184
7.2 DIN 14 Steel (1.5 Cr, 0.90 Mo, 0.25 V): Effect of
Liquid Nitriding
1000
'\
900
- 140
:'\.
""-8
800
700
I'
~ 600
.s
13
IS
- 120
- 100
0A
- 80
';;;
-'"
.~
13
>-
500
400
300
- 60
\
~
105
106
- 40
107
Number of load cycles
Effect of nit riding on fatigue behavior of DIN 14 CrMoV

69 steel (0.14 C, 1.5 Cr, 0.90 Mo, 0.25 V). Curves A and C
are for hardened and tempered (not nitrided) specimens;
Band D are for liquid nitrided specimens. A and B are for
smooth specimens; C and D are for notched specimens
K,=2.
Nitriding introduces residual compressive stresses at the
surface of steel parts; these residual stresses, together with the
increased strength of the nitrided layer, increase the fatigue
resistance of the part. The increase in fatigue strength that results from nitriding is illustrated in these S-N curves.
Park OH, 1978, P 541
7-3. 2.25Cr-1.0Mo Steel: Influence of Cyclic Strain Range

on Endurance Limit in Various Environments
10.0
8.0
6.0
KEY
3.0
2.0
z
..:
a: 1.0
fVI
0.8
-.
~XX
ENVIRONMENT
Na (30 ppm OF 21
AIR
HELIUM
Na \300 ppm OF 02)
. ,
x~
'~
'U ...~~~.
........~
-~
.J
u 0.6
>u
....
11K
- .. ~
.........=-
0.4
0.3
0.2
C.I
10~
CYCLES TO FAILURE
Influence of cyclicstrain range upon fatigue endurance of 2.25Cr-l.OMo steel in

sodium, air, and helium at 865 K. (Cycle used was approximately up 5 s, hold
5 s, down 5 s, hold 5 s.)
Source: R. H. Cook and R. P. Skelton, "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
in Source Book on Materials for Elevated-Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals
Park OH, 1979, P 83
185
186
Testing temperature
~C
o 425
,
?f!.
OJ'
"Q
540
595
l1
FI
800,
1000
1100
.'=
a.
E
c
'iii
'"
J5
0.2
0.1 2
10
103
Cycles to failure
The results of strain-controlled fatigue tests of 2.25Cr-l.OMo

steel at 425, 540 and 595C (800, 1000 and 1100 OF) on specimens of annealed 2.25Cr-I.OMo steel are presented in these S-N
curves. Within this range, test temperature had relatively little
effect on number of cycles to failure.
Park OH, 1978, P 659

and Strain Rate
Temperature,Oc
104
:e
B
:G
>
u
103
500
450
400
350
550
Temperature,OC
I
600
700
800
900
1000
1100
Temperature, of
Effect of elevated temperature on strain-controlled fatigue behavior of annealed 2.25Cr-l.OMo steel.
Strain-controlled fatigue tests have also shown (note above) that reducing
carbon content to 0.03% results in a reduction in fatigue strength. Furthermore, because of variations in strain aging effect, specimens from one heat
with a higher carbon content ran longer at 427C (800 OF) than at 316 C
(600 OF).
Source: Metals Handbook, 91h Edition, Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, p 659
187
188
7-6. 2.25Cr-1.0Mo Steel: Effect of Temperature on Fatigue Crack

Growth Rate
6K, stress intensity factor, ksi..Jiil.
10
20
60
I--+-------+-------+--------+-----i
10.4
Testing temperature
10-31--+--oC 205
o 370
of - - - + - - - - - - - + - - - - - - - - - - - f + - - - l
400
700
1---+-6-455-850---+--------+--------,.~____J<-----,_F____I1____l
6595
1100
10-6
Q>
Q>
~
.
E 10-4
E
:i
\!
s:
~
...e
s:
~
0
l;,
.><
eu
10.6
:e'
l;l
1;
:e'
..
:!'!
:!'!
"D
"D
10-6
10
20
60
s, stress intensity factor, MPa vmm

Variations in fatigue crack growth rate with test temperature for specimens of 2.25Cr-l.OMo steel tested in air.
Specimens were subjected to cyclic loading at a constant maximum load.

Stress ratio was 0.05; cyclic frequency was 400 per minute. As shown, the
stress-intensity factor range increased as the crack length was increased.
Park OH, 1978. P 660
7-7. 2.25Cr-1.0Mo Steel: Effect of Cyclic Frequency

on Fatigue Crack Growth Rate
b.K, siren intemity laClor. ksi...rm:
2
I 0-4
4C"7f'/
/,r:/
/ ! vr:~400C~
/
40 CfKO
2
Frequency,
cvcles/mln
o 4
40
2
2
ilK. stressintensity laflOr, MPaym
(.1
IJJ.K. stress intensity factor, ksi
2
vrn:
2
I
1-//
~E
4C~
II;
t
III
1/
2
I
40c"m
400c:pm
2
Frequency.
cycles/min
o
4
40
2
2
5
4K. slreu intensily 'actor. MPa V;;
Ibl
Data shown above indicate that in elevated temperature tests at a given

stress-intensity factor range, crack growth rate increases as cyclic frequency
is decreased. These fracture mechanics data may be applied to the design of
structural components that may contain undetected discontinuities, or that
may develop cracks in service. Stress ratio was 0.05. (a) Tested at 510 C
(950 OF); (b) tested at 595C (1100 OF).
Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels, American Society ForMetals, Metals
Park OH, 1978, p 661
189
190
7-8. 2.25Cr-1.0Mo Steel: Fatigue Crack Growth Rates in Air

and Hydrogen
10-5
10-6
v
i>.
i>.V
vi>.
Vi>.
i>.
~
-.
.Jl
i>.
10-7
-.
'"
al
i0
.dt
10-8
Ol
o~ ,
o ,
Gl
:::I
oOe
.l'
~
u
0'
oil'
't:l
.....
't:l
'"
s:
;.
Gl
1i
10-9
0.,.
l-
Ol
.;::;
'"
u..
8
10- 10
Frequency
f-
10- 11
f:
2% Cr - 1 Mo Steel
R = 0.05
V
i>.
Environment
Air
2 Hz, 50 Hz
H,
0 50
5 Hz
138 kPa H2
2 Hz
0.5 Hz
~I Threshold
,f
6 78910
20
40
60
80 100
Stress-intensity factor range, L1K, MPa . m 1/2

Fatigue crack growth rates in 2.25Cr-l.OMo steel in air and in hydrogen.
Corrosion fatigue descriptions are further complicated by the fact that the environment may produce multiple effects. For example, Suresh et a/demonstrated that
dry hydrogen may produce afrequency-sensitiveenvironmental effect analogous to
SCC at intermediate t:J.K values and a frequency-insensitive environmental effect
near the threshold. This is illustrated in the above graph for 2.2SCr-1Mo steel tested
in air and in 138-kPa hydrogen gas. Because the sustained-load threshold for this
steel is on the order of 90 M Pa . m 112 (82 ksi . in. 112), the K""'j of about 22 MPa .
. 1/2) gives
K th :
v.
m 1/2 ( 20 k'
SI . Ill.
K ,h ({) ...?
~
lt can be seen for t:J.Kvalues greater than K ,h ({) that there is a large increase in
growth rate for the low test frequencies but not for the higher ones. Therefore, this
regime may be considered to be one where superposition might apply. In addition,
however, there is a true threshold, t:J.K,h , which appears to be frequency-insensitive
but which nevertheless decreased by about 30% to SA MPa . m 112 (4.9 ksi in. 112 )
because ofthe hydrogen environment. Such mulitple effects are poorly understood
and are clearly possible in a large number of material/environment systems.
Source: W. W. Gerberich and A. W. Gunderson, "Design, Materials Selection and Failure Analysis," in Application of Fracture
Mechanics for Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds..
191
7-9. 2.25Cr-1.0Mo Steel: Effect of Holding Time
120days
Compressive or
tensile hold
Type of strain hold
Strain
range.%
Compressive
Tensile
Both
2.0
loJ
104
Cycles to failure
1.0
C.}
0.5
:::2 } .
X indicates zero hold lime
Cycles to failure
(b)
Time-to-failure/cycles-to-failure diagrams for annealed 2.25Cr-l.0Mo steel

tested in strain-controlled cyclic loading at (a) 480C (900 OF) and (b) 540
C (1000 OF). Hold time indicated on
graph is length oftime that specimens
were held (during each cycle) in the
state of maximum tensile strain (open
symbols) or compressive strain (filled
symbols). Strain amplitude indicated
by shape ofsymbols and figures along
zero-hold-time line.
In these "time-to-failure/cycles-to-failure" diagrams, the lowest curve (zero hold time) indicates the corresponding time period and number of cycles to failure for continuous strain-controlled fatigue tests over the
strain range from 0.4 to 2.0% with no holding period at maximum strain. The other curves, which are
approximately parallel, are for increasing periods of holding time at maximum strain levels in either tension
or compression. The vertical curves are drawn through the number of cycles to failure for each particular
cyclic strain. For all tests at 2% strain, failure occurred in less than I000 cycles regardless of holding time or
whether the stress was tensile or compressive. The effect of reducing the strain increment and increasing the
holding time on number of cycles to failure can be determined from the appropriate curves in the figures.
Source: Metals Handbook, 9th Edition, Volume I. Properties and Selection: Irons and Steels, American Society for Metals. Metals
Park OH. 1978. pp 662-663
192
7-10. Cast 2.25Cr-1.0Mo Steel, Centrifugally Cast: Fatigue

Properties at 540C (1000 OF)
60,------r----r---r---r--r---.,.------,
5.0
I 4.0
lLJ
o
Z
<l
cr
3.0
<r
cr
I-
(f)
2.0
...J
g
1.0
OL...._ _-L-_ _---1._---1._.l.-...L-_ _--'-_ _- . J
100
200
400 600
1000
2000
CYCLES TO FAILURE
4000
Fatigue properties of 2.25Cr-1.0Mo centrifugally cast pipe,

A2l7, Grade WC9, at 540C (1000 OF).
Source: Steel Castings Handbook, 5th Edition. Peter F. Weiser, Ed. Steel Founders' Society of America. Rocky River OH. 1980,
P 15-55
7-11. H11 Steel: Crack Growth Rate in Water and in Water Vapor
0.5 ,........----,-----,--"'"""'T--"'"""'T--~--....,
"0
0.3
c
E 0.1
.....
c
Q)
.05
H -II Steel
230 ksi Y. S.
K - 30 ksi IT'"
o
Activation Energy
9,000 cal./gm-atom
o
a:::
-...
s: .03
.Jl:
U
...
o
u .01
o Water
o Relative Humidity 100% at Test Temperature
0.005
II
II
II
II
80 F,
Tested at Higher Temperature
Crack growth rate versus temperature for an HII steel in water and
water vapor.
It is of considerable interest that the strain rate and temperature

dependence of hydrogen embrittlement, as determined by ductility
measurements after rising load tests on hydrogen-charged materials,
show a characteristic behavior that resembles closely that seen with
crack growth rate measurements and external hydrogen environments.
Source: Herbert H. Johnson. "Keynote Lecture: Overview on Hydrogen Degradation Phenomena," in Hydrogen Embrittlement
and Stress Corrosion Cracking, R. Gibala and R. F. Hehemann, Eds., American Society for Metals. Metals Park OH. 1984, P 18
193
194
7-12. 9.0Cr-1.0Mo Steel: Creep-Fatigue Characteristics
9% Cr 1% Mo AT 525C WITH OR WITHOUT

TENSILE HOLD TIMES TO
hOWELL
(TOTAL STRAIN RANGE; FROM (22))
LOW C 9% Cr 2% Mo AT 550C
CONTINUOUS CYCLING
(TOTAL STRAIN RANGE; FROM (2311
~
UJ
l-'
Z
10
cr
cr
~
Vl
9% Cr 2% Mo NbV
AT 550C
(PLASTIC STRAIN
RANGE; FROM (24))
~--5
MIN. TENSILE
DWELL
CYCLING
10'
CYCLES
TO FAILURE
Illustrating the elevated temperature low-cycle fatigue and creep-fatigue properties of normalized and tempered 9% Cr Mo variants.
In this chart are presented the elevated-temperature-fatigue and creepfatigue data for the 9%Cr-1%Mo steel as a single curve in terms of total strain
range against cycles to failure; also shown for direct comparison are the continuous cycling fatigue data for the low-C, 9%Cr-2%Mo variant which, although inferior at relatively high strain ranges, suggests superior endurance
may be attained in the high-cycle region. From the limited evidence, it seems
probable that normalized and tempered 9%Cr-I%Mo steel may be used in
reactor-quality sodium at service temperatures with little effect on tensile
properties and stress rupture strengths or ductility and that the short term
low-cycle fatigue endurance will be increased and fatigue crack growth rate
reduced. This behavior is a consequence of the structural stability of the
material with respect to interstitial element transfer in liquid sodium and also
the low oxygen potential of the overall system which may be expected to
preclude oxide penetration and enable partial recohesion of the crack faces
during fatigue.
Source: S. J. Sanderson, "Mechanical Properties and Metallurgy of 9%Cr I%Mo Steel," in Ferritic Steels for High-Temperature
Applications, Ashok K. Khare, Ed., American Society for Metals, Metals Park OH, 1983, P 95
7-13. 9.0Cr-1.0Mo Modified Steel: Stress Amplitudes Developed

in Cycling
-------------------------200 t
- - - - - -__
en
w
a::
------ --- ..
----Continuous Cycle
100
e
a..
::E
Fe 9Cr IMo BV, Nb
(/)
l-
(/)
-100
- - with 30 sec T H
---
-200
r,.,..,,"'="::=-:-:.....---......:_:'":_;:"::-=-=_,,:_::- - - _...
10
100
1000
10000
CYCLES
This chart shows stress amplitudes (tensile and compressive)
that developed in the course of cycling the modified Fe-9.0Crl.OMo steel through a total strain range of 0.5% at 649C
(1200 OF). Fatiguing was carried out in vacuum. Dotted curve
indicates continuous cycling; solid curve indicates cycling with a
30-s hold at maximum tensile strain.
Source: S. Kim, J. R. Weertman, S. Spooner, C. J. Glinka, v. Sikka and W. B. Jones, "Microstructural Evaluation ofa Ferritic
Stainless Steel by Small Angle Neutron Scattering," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck
and Stanley M. Wolf, Eds., American Society for Metals, Metals Park OH, 1984, p 175
195
196
7-14. 9.0Cr-1.0Mo Modified Steel: Effect of Deformation
cb:
dQ
Fe 9Cr IMo 8 V,Nb
In 10 8
8
6
4
'i
Not deformed (N aT)
Fotl(~ued
Foti~ed ( 649C, At
(649C, A Et = 0.5 -/0,

10000 cycles, continuously
79
= 0.5 %,
cycles with 30 sec tension hold
Crepl
R~9~~,M~fb28
holn)
~
~
-2
-4
0.02
0.04
0.06
0
0
0.08
<I")
Curves of dI.ldO vs q for specimens of modified Fe-9.0Cr-1.0Mo steel which have

undergone various types of deformation. A magnetic field of ~28 kg was applied to the
specimens during the SANS measurements. A = 0.48 nm,
Source: S. Kim, J. R. Weertman, S. Spooner, C. J. Glinka, V. Sikka and W. B. Jones, "Microstructural Evaluation of a Ferritic
Stainless Steel by Small Angle Neutron Scattering," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck
and Stanley M. Wolf, Eds., American Society for Metals, Metals Park OR, 1984, P 175
197
8-1. Type 301 Stainless Steel: Scatter Band for

Fatigue Crack Growth Rates
K.If , ksi in.1/2
20
60
40
80 100
10- 3
GI
i3
~
......
10- 2
E
E
"t:l
......
I /
'"
"t:l
......
GI'
'"
/ /
..c
GI
i3
10- 4
.~
// /
i0
... 10- 3
'"
u
...u
'"
GI
:::I
'"
u..
'"
.;;
10- 4
L-.J'-
/
/
/
/
/
/
>u
'<,
"t:l
......
'"
"t:l
/
10-5
1/2 hard tvpe 301
24C (75F)
0.063 < R < 0.807
--'-_ _.1...---J'---'---'---'-...J...........L - - J
10- 6
80
Effective stress-intensity factor,
K.1f = Kmax [1 - Rl o.66 7 , MPa m 1/2
Scatter band offatigue crack growth rates of Y2hardtype 301 stainless
steel, tested at 24 0 C (75 OF), 10 Hz, andR ratios of 0.063 to 0.807 based
on effective stress-intensity factor, Kerr'
Fatigue crack growth rate data reported by Walker for Y2-hard type 301 stainless steel sheet are
summarized in the above graph. The data were obtained in air at room temperature over a series ofload
ratios (R) from 0.063 to 0.807 at a frequency of 10 Hz. These data are based on the "effective stress
intensity factor," Kerr, rather than on fj,K, to account for the effect of the range of stress ratios. Kerr is
defined as follows:
Kerr = K max (I - R)m
where m is determined empirically and R is the load ratio (minimum load/maximum load) on cyclic
loading. The crack growth rate law then becomes:
da/dN= C[Kmax(l- R)my
Results of fatigue crack growth rate tests on austenitic stainless steels have shown that the crack growth
rate tends to increase as the R ratio is increased, when compared at given values of fj,K. If tests are made
at several load ratios to determine m, then the effects of other load ratios may be estimated.
Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
Metals. Metals Park OH, 1982, P 114
198
8-2. Type 301 Stainles Steel: Effects of Temperature and

Environment on Fatigue Crack Growth Rate
tlK, ksi . in. 1/ 2
20
10- 5
Gl
>
c.>
I"'"
--.E
--'"
2'
"'"
"tl
"tl
10- 6
20C (68F)
L-T
Type 301
30
40
50
2 X 10- 7
60
Stress-intensity factor range, tl K, MPa . m 1/2

-----Annealed, tested in argon
- - - - - Annealed, tested in air
- - - - - Warm worked, tested in argon
Fatigue crack growth rates for type 301 stainless steel have been
reviewed by Pineau and Pelloux in the temperature range from - 30 to
+95 C (-22 to +203 OF).The results, summarized in this graph, were
obtained on compact specimens 7 mm (0.28 in.) thick at a cyclic frequency of 20 Hz with a sinusoidal waveform at a load ratio (R) of 0.0 1.
All specimens were tested in dry argon except one series that was tested
in laboratory air. For the annealed specimens tested in argon, fatigue
crack growth rates at a given t::.K value increased as the temperature
increased over the testing temperature range. Fatigue crack growth
rates in laboratory air at 20C (68 OF) were higher than for corresponding conditions in argon, indicating that the humidity and/ or oxygen in the air influenced the growth rates.
The warm worked specimens were reduced 65% at 450 to 500C
(840 to 930 OF), resulting in a substantial increase in strength. Fatigue
crack growth rates for the warm worked specimens (above) indicate
that the fatigue crack propagation properties of the warm worked alloy are different from those of the annealed alloy. This effect of warm
working has been observed for other austenitic stainless steels. These
differences are attributed to the extent of the strain-induced transformation at the crack tip. This transformation effect would be most noticeable in type 301, because it is less stable than the other alloys in the
UNS S3xxxx series.
Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OR, 1982, P 113
8-3. Type 304 Stainless Steel: Effect of Temperature on

Frequency-Modified Strains
AISI 304 STAINLESS STEEL

o
o
LIJ
.....J
<.,;)
..
430C
650C
8160C
(f)
a:
.....
(f)
Data of Berling and Slot for AISI 304stainless steel, showing frequency-modified elastic and plastic
strains at three temperatures in air,
In contrast to most other segments of our technology, interest in the fatigue problem in the power-generation industry generally involves elevated
temperature. Laboratory testing on both smooth specimens and specimens
designed for crack growth is performed with temperature and frequency or
strain rate as parameters. The importance offrequency or strain-rate effects
is shown in this chart. These data are for solution-treated AISI 304 stainless
steel subject to triangular wave shapes at equal-loading and reverse-loading
strain rates. Representation of the behavior here utilizes fatigue equations
known as frequency-modified fatigue equations. They describe the elastic
and plastic strains versus fatigue life and include the frequency or strain rate
of the test. For the present purposes they are useful in showing how increasing temperature acts to change the cyclic stress-strain response and the strainlife fatigue response of this alloy.
Source: L. F. Coffin, "Fatigue in Machines and Structures-Power Generation," in Fatigue and Microstructure, American Society
for Metals, Metals Park OH, 1979. P 13
199
200
8-4. Type 304 Stainless Steel: Fatigue Crack Growth RateAnnealed and Cold Worked
~K, ksi . in. 1/ 2
20
10
30
40
60
Cold worked 25%,

tested at 427C (800F)~
.,
U
I
I ~//
I
10- 3
>
u
-<;
E
E
/
/
.,
I'
'"
"t:l
....,'~
...s:
s:
IJI,'
Cl
-"u
>
u
'<,
.~
z'
Cold worked
25%, tested
at 25C
(77F)
10-5
"t:l
-..
'"
"t:l
/it'
.,u
::>
Cl
'':;
'"
10-4
y/~
z'
"t:l
'<,
U.
,1:"
I
I
Annealed, tested at
427C (800F)
-
80 100
2 X 10-4
/
10- 4
"
Annealed,
tested at
25C (77F)
Type 304
........
10
-'--_--'-_....L.._---'_....L.----lU
20
30
40
60
10- 6
80 100
Stress-intensity factor range, ~K, MPa m 1/2

Fatigue crack growth rates for annealed and cold worked type
304 stainless steel at 25 and 427C (77 and 800F), 0.17 Hz, and
an R ratio of O.
In some applications, type 304 stainless steel components are fabricated in

the cold worked condition to improve strength properties. A comparison of
fatigue crack growth rate data by Shahinian, Watson, and Smith, illustrated
in this graph, shows that the high-~K crack growth rates were lower for the
cold worked specimens than for the annealed specimens. Crack growth rates
were higher for the specimens tested at 427C (800 OF) than for corresponding specimens tested at room temperature.
Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels." in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell, William w. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park OH, 1982, p 120
8-5. Type 304 Stainless Steel: Effect of Humidity on Fatigue Crack

Growth Rate
CiK, ksi . in.1/2
10
20
40
60
80 100
10- 4
10- 5
~Roomair
10- 4
20
Type 304
25C mOF)
30
40
Stress-intensity factor range, CiK, MPa m 1/2
Effect of humidity on fatigue crack growth rates for type 304

stainless steel tested at room temperature, 0,17 Hz, and an R
ratio of O.
The effects of humid air environments on the room temperature fatigue crack growth rates of
specimens of annealed type 304stainless steel are shown in the above chart for specimens cycled
at 0.17 Hz with an R ratio of zero (Shahinian, Watson, and Smith). At the lower end of the t:.K
range, fatigue crack growth rates in humid air are substantially greater than crack growth rates
in dry air. However, fatigue crack growth rates of specimens oftype 304stainless steel tested in a
pressurized water reactor environment at 260 to 315 C (500 to 600 F) with R ratios of O. 2 and
0.7 were no greater than the fatigue crack growth rates in air at the same temperature with an R
ratio less than 0.1 (Bamford). However, variations in R ratios influenced the fatigue crack
growth rates in the pressurized water reactor environment.
Metals, Metals Park OH. 1982. p 122
201
202
8-6. Type 304 Stainless Steel: Effect of Aging on Fatigue Crack

Growth Rate
l1K, ksi . in.1/2
10-4
10- 3
.,
>
lJ
-.~
10- 5 Z
~
'"
"t:J
10- 4
Unaged
Aged
Hold time
Zero
0.1 min
1.0 min
10- 6
Type 304
593C (1100F)
10-5
.l--_ _....L..._--'-_....I....----I_L-.L-J........J
L-..JL-.I--'-
10
20
40
60
80
100
Stress-intensity factor range, l1K, MPa . m 1/2

Effect ofaging at 593 C (1100 OF)for 5000 h, and hold times of 0.1 and
1.0 min for each cycle, on fatigue crack growth rates of L-T oriented
specimens of type 304 stainless steel tested in air at 0.17 Hz and an R
ratio of O.
Because austenitic stainless steels are expected to give long service life, an evaluation of
the effect oflong-time aging at service temperatures is important. Results offatigue crack
growth rate tests on specimens that were tested in the unaged and aged conditions (5000
hours at 593 C, or 1100 OF) are shown in this graph, as reported by Michel and Smith.
After aging for 5000 hours at this temperature, precipitation of M 23C 6 carbides is essentially complete. These results indicate that at 593 C (1l00 OF) there are no deleterious
effects of aging on the crack growth rates of specimens that are continuously cycled. When
a holding time of 0.1 or 1.0 minute is included in each loading cycle, there tends to be a
slight increase in the fatigue crack growth rate at a given 11Klevel.
Metals, Metals Park OH, 1982, P 121
8-7. Type 304 Stainless Steel: Effect of Temperature on Fatigue

Crack Growth Rate
.1K, ksi . in,1!z
10
20
649O C l 1 2
.......
E
E
10-3
10-4
t<,
1/ .:
"tJi'"
..
.'"
"".
'"
100
538"C (1000"-><-/
i0
60
II>
c:;
>u
40
10- 4
f!
u
u
II>
:::l
t'I~
II>
10- 5
'"
u,
10- 5
"'"
"
.......
~~~
10-6
316C 1600F)
10
>-
u
.......
.E
z
~'
...'"
c:;
20
Type 304
40
60
100
Stress-intensity factor range, .1K, MPa . m l/Z

Effect of testing temperature on fatigue crack growth rates for annealed
type 304stainless steel tested in air at 0.066 Hz and anR ratio of 0 to 0.05.
Results offatigue crack growth rate tests on types 304 and 304L stainless
steel at room temperature and at elevated temperatures have been reported by
James and Schwenk, and by others. As shown in this graph, increasing the
exposure temperature from room temperature to 650 C (1200 F) increases
the fatigue crack growth rates at any ~Klevel within the range ofthe tests in
an air environment. These data, reported by James and Schwenk, are for
specimens of both the L-T and T-L orientations, for several different maximum alternating loads, for load ratios of 0 to 0.05, and for cyclic frequencies
from 0.033 to 6.66 Hz for the room-temperature tests and 0.067 Hz for the
elevated-temperature tests.
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American Society for
Metals, Metals Park DB, 1982, P 115
203
204
8-8. Type 304 Stainless Steel: Damage Relation at 650C (1200 OF)
STRAIN RATE,
IN.lIN.lSEC
4 x 10'3
o
().
~
'i7
HOLDPERIOD, MINUTES
TENSION
COMPRESSION
o
0
I
0
10
0
30
0
60
0
180
Ll
o
~
(3
0
30
3
30
30
0.1
0
0
3
30
3
0
0
0
-tI
4 x 10'3
4 x 10'4
4 X 10'5
O"rtlpN,
I3 v l3 ( k. I ) =
C = 1.158
x 10 5
13 = 0.895
k
= 0.756
101' - : - - -.........- -.........- - ' -.........L.;:-----''------''-----'---'--'-=----'----'----'---'---'-:-----'----'--.........- I -.....

101
105
Ostergren's damage relation for AISI 304 at 650C (1200 OF).
The damage function was proposed by Ostergren and is based on the frequencymodified fatigue approach. A damage function is approximated by the quantity U,!:iE p ,
where u, is the maximum stress in the cycle and !:iE p is the inelastic strain range. The tensile
hysteresis energy is employed to account for the facts that low-cycle fatigue is essentially a
crack-growth process and that crack growth and damage occur only during the tensile
part of the cycle. The use of the tensile-stress quantity, in conjunction with the plasticstrain range, provides a means of accounting for loop unbalance, since, for the same
inelastic strain, a positive mean stress provides a greater hysteresis energy than does a
compressive mean stress. The method is effective in accounting for hold-time effects, as
indicated in the chart above.
Source: L. F. Coffin. "Fatigue in Machines and Structures-Power Generation, "in Fatigue and Microstructure, American Society
8-9. Type 304 Stainless Steel: Fatigue Crack Growth Rate at Room
and Subzero Temperatures
10-4
10-3 1-- +-- - +-- - - - - - +-'''''10,......,

24C (75 of)
o -196C (-320 OF)
0-269 c (-452 OF)
10-4 f--+---+-------4I::L-+------=l
10
50
Stress intensity factor
range, 11K, MPa
100
vm
Fatigue crack growth rate data for type 304 austenitic stainless steel
(annealed) at room temperature and at subzero temperatures. For this
alloy, crack growth rates are nearly the same at room and cryogenic
temperatures.
Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society for Metals, Metals Park OH, 1980, p 756
205
206
8-10. Types 304 and 304L Stainless Steel: Effect of Cryogenic

Temperatures on Fatigue Crack Growth Rate
~K. ksi in. 1/ 2
20
40
60
80
Gl
~
......
E
E
10- 3
"C
......
Cll
Gl
"C
~i
Type 304L
22C (72FI
...i!=
10- 5
--
Cll
"C
tJ)
.>t!.
...u'"
~
tJ)
-.
.E
"C
...0
Gl
>
u
10- 4
Type 304L
-196, -269C
(-320, -452FI
.;;
'"
IL
10- 6

Fatigue crack growth rates for annealed types 304 and 304L stainless
steel at room and cryogenic temperatures, 20 to 28 Hz, and an R ratio
of 0.1.
Fatigue crack growth rate data obtained by Tobler and Reed on

specimens of types 304 and 304L stainless steel (annealed) at temperatures in the range from room temperature to liquid helium temperature (-269C, or -452 OF)are shown in this graph. The data for type
304 were scattered over the range shown, while for type 304L, the data
at room temperature described one curve and the data at the cryogenic
temperatures described the other curve. These results indicate that
cryogenic fatigue crack growth rates for type 304 do not deviate significantly from room temperature fatigue crack growth rates over the
/:!"K range studied. Furthermore, if design calculations for type 304L
are based on room temperature fatigue crack growth rates, the calculations will be conservative for cryogenic exposure.
8-11. Type 304 Stainless Steel: Fatigue Crack Growth Rate in Air
With Variation in Waveforms
~K,
10
ksi . in.
'/
30
20
40
50
10- 4
CIl
U
~
.......
E
E
10- 3
/ /
/
/ /
I /
/ /
II
/ /
Waveforms
Z
'1::J
.......
co
'1::J
2l'
s:
i0
..
..
( (
/ /
/ /
."
u
co
u
10- 4
/ /
/ /
/
/
CIl
:J
."
.;;
co
u,
rYYY\
10- 5
20
30
40
.E
Z
} 0.067 Hz
Type 304
538C (1 0000 F)
R = 0.05
>-
~
co
'1::J
/VVV\
10
.!!
u
.......
10- 6
50
Stress-intensity factor range, ~J<, MPa . m 1/2

Scatter band of fatigue crack growth rates for annealed type 304 stainless steel at 538 C (IOOOF)in air at anR ratio of 0.05 with two different
waveforms at 0.067 Hz.
The data presented in this graph were obtained in tests with a sawtooth waveform. Changing from a sawtooth waveform to a waveform
with a short holding period at maximum load did not influence the
overall fatigue crack growth rates according to additional data reported by James and shown above.
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds., American Society For
Metals, Metals Park OH, 1982,P 117
207
208
8-12. Type 304 Stainless Steel: Effect of Hold Time on Cycles to

Failure
30
LABORATORY TESTS
AISI304 STAINLESS STEEL
650C
z
......
180
Z
ILl
C>
z 0.01
a:::
60
a:::
~
en
-J
60
DO 01
10
0
30 30
o
I
o NO HOLD TIME
o TENSILE HOLD TIME IN MINUTESAS INDICATED
0.001 L...-_--L._ _.L..---L----L---l.-_ _.L..-_----L_..L-...l..-.L..-_-..l..._ _--'--~

100
1000
10,000
LIFE - CYCLES TO FAILURE
Effect of hold time on life for AISI 304 stainless steel.
Wave-shape effects are also important in fatigue crack

growth, as has been studied by Barsom. He observed that the
crack growth rates were greater as the loading rate increased and
the unloading rate decreased, given a fixed period of cycling.
Overload effects are also important in retarding crack growth.
Substantial damage can result from these wave shapes, particularly when the hysteresis loop is severely unbalanced, as can occur in long tensile-strain hold-time tests.
8-13. Type 304 Stainless Steel: Effect of Hold Time and Continuous
Cycling on Fatigue Crack Growth Rates
6K, ksi . in. 1/ 2
10
40
20
60
80 100
10- 1
.:E.
0.1 min hOld-...f

I
"
"
,r~
10- 2
1.0minhold
Type 304
593C (1100F)
10- 1 '--_ _---'_ _....L-_...L-_---'_---'_...L...I

20
30
40
60
80 100
Stress-intensity factor range, 6K, MPa . m 1/2
Fatigue crack growth rates per unit of time ida]dt) for annealed
type 304 stainless steel for continuous cycling (0.17 Hz), for 0.1
and 1.0-min hold times at maximum load for each cycle at 593 C
(1100 OF), and for anR ratio of O.
As shown above, the fatigue crack growth rate is greater for

specimens tested with no holding time (continuous cycling) than
for specimens held at maximum load for 0.1 or 1.0 minute per
cycle. The lowest fatigue crack growth rates occurred for specimens with the longest holding time, based on dal dt, The same
trend was observed for tests at 593C (liOO OF), as shown here.
Therefore, cyclic loading has a more damaging effect than static
loading on crack growth per unit of time.
209
210
8-14. Type 304 Stainless Steel: Effect of Cyclic Frequency on

Fatigue Crack Growth Rate
dK, ksi . in.1/ 2
10
20
30
0.0014 Hz
40
50
0.0067 Hz
10- 4
Q)
>o
--EE
z'
--'"
10-3
"tl
"tl
l!l'
Q)
...::
10- 5
s:
een
"tl
..10:
"tl
10- 4
Q)
>u
--.5
z'
--'"
;:,
en
'':;
'"
u..
Type 304
538C (lOOOF)
R = 0.05
10
20
30
40
10- 6
50
Stress-intensity factor range, d K, MPa . m 1/2

Effect of variation in cyclic frequency on fatigue crack growth
rates for annealed type 304stainless steel at 538 C (1000 0 F) for
an R ratio of 0.05 in air with a sawtooth waveform.
For fatigue crack growth rate tests on specimens of annealed

type 304 stainless steel at elevated temperatures, increasing the
cyclic frequency will decrease the crack growth rate over part of
the ~K range, as shown here for tests at 538C (1000 OF).
Metallic Structural Materials, James E. Campbell, William w. Gerberich and John H. Underwood. Eds., American Society for
8-15. Type 304 Stainless Steel: Effect of Frequency on Fatigue Crack

Growth Behavior
~K,
SiRESS INTENSITY FACTOR RA:\iGE,
kg/lmm?/2
Il
5x 101
..,
~
.l::!
s:
u
c:
z
::!2
"'
't:l
....-c.
~
co:
10-5
....
:I:
;:
0
co:
<.:l
:><:
e:(
co:
ANNEALED TYPE 304 S S.

TESTED IN AIR AT 53S"C 1l000f)
R 0.05, Ref. [45]
:::I
<.:l
....
-e
......
0.003 cpm
'il 0.4 cpm
00 4cpm
10-6
40cpm
li. 400 cpm
4000cpm
4
10
STRESS INTENSITY FACTOR RANGE,
~K, Ib/lin)3/2
5x101
STRESS INTENSITY FACTOR RANGE, oM, foIW/(m?/2

Effect of frequency on the fatigue crack growth behavior of type 304 tested in an air
environment at 538C (1000 OF).
211
212
8-16. Type 304 Stainless Steel Welded With Type 308: Fatigue Crack
Growth Rates
t.K. ksi . in. 1I2
t.K. ksi .
in. 1/ 2
40
20
20
40
60 80
60 80
10- 2
Ql
10- 4
Ql
c:;
10- 4
--
E
E
Shielded
metal arc
E
Z
c:;
>
u
--. 10...
"t:l
10- 3
"t:l
"t:l
co
-e
Ql
c:;
>
u
...eoi
j!l'
--.S ...::..
z '"
--.. e
i0
"t:l
10- 5
"t:l
s:
u
u
Submerged
arc
..'"
.;:;
~
.;:;
u..
10- 4
--.S
z
--..
"t:l
Ql
..~
Ql
c:;
>
u
s:
u..
Shielded
metal arc
10- 5
"t:l
10- 4
o SMAW 1
SMAW 2
24C (75F)
20
40
10-6
60 80 100
Stress-intensity factor range, t.K, MPa . m 1/2
593C (l100F)
20
40
10-6
60 80 100
Stress-intensity factor range, t.K, MPa . m 1/2
Fatigue crack growth rates for annealed type 304 base metal and type 308 weld metal at 24 and
593C (75 and 1100 OF), 0.17 Hz, and an R ratio of O.
Type 308 stainless steel is the alloy that is usually used for welding rod for weldments in
type 304 stainless steel when those weldments are to be exposed to room temperature or to
elevated temperatures in service. Because service experience has shown that failures are
more likely to originate in weld metal or in heat affected zones than in the base metal, it is
important to have fracture information on weldments. In general, fatigue studies at elevated temperatures on specimens from type 304 weldments have shown that the fatigue
crack growth rates in the type 308 weld metal and heat affected zones are no greater than
in comparable specimens of the base metal. Fatigue crack growth rate data obtained by
Shahinian for specimens of type 304 welded with type 308 rod by the submerged arc and
shielded metal arc processes are shown above for tests at room temperature and at 593C
(llOOF).
Source: J. E. Campbell. "Fracture Properties of Wrought Stainless Steels." in Application of Fracture Mechanics for Selection of
Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood. Eds.. American Society for
Metals. Metals Park OH. 1982. P 125
8-17. Types 304 and 310 Stainless Steel: Effect of Direction on S-N
400
Type 310, transverse
Type 310, longitudinal
<II
Q.
~" 300
~
~E 200
:::l
'iii
..><:
40 '"
.......'"
30 '"
","
'"...
,~
50
~~-~
Type 304, longitudinal
E
:::l
20 ,~
x
~ 100
10
105
106
No. of stress cycles
"'
0
108
.SoN curves for two grades of stainless steel,
Metals, American Society for Metals, Metals Park OH, 1980, P 32
213
214
8-18. Types 304, 316, 321, and 348 Stainless Steel: Effects of
Temperature on Fatigue Crack Growth Rates
AK, ksi . in. 1/ 2
20
10
60
100
40
10-2 ...-r----r----.------r---.---. 4 X 10-4
10- 4
Q)
>
Ll
--EE
z
--'a>"
"0
10- 3
"0
.!!!
...
s:
...3:
Ll
>
Ll
-z
-.~
"0
Cl
l'Il
..
"0
Ll
l'Il
10-5
Ll
Q)
::::I
Cl
.~
l'Il
IL.
10-4
L -_ _....L.
10
20
10-6
80 100
....I...._--l._...l.-..J.J
40
60
Stress-intensity factor range, AK, MPa m 1/2
Fatigue crack growth rates for annealed types 304, 316, 321, and 348
stainless steel in air at room temperature and 593C (1100 OF), L-T
orientation, 0.17 Hz, and an R ratio of O.
As reported by Shahinian, Smith, and Watson, fatigue crack growth rate tests were made on singleedge-notch cantilever specimens oftypes 321 and 348 stainless steel from L-T orientation at 0.17 Hz
with an R ratio of zero at room temperature and at elevated temperatures to 593C (1100 OF). As for
types 304 and 3I6, fatigue crack growth rates in air increased with increasing testing temperature. The
curves above show that, at room temperature, the fatigue crack growth rates for types 304, 316,321,
and 348 all fall within a narrow band. For tests at 593C (1100 OF),however, specimens of type 3I6 had
the least fatigue crack propagation resistance, whereas specimens of type 348 had the highest fatigue
crack propagation resistance, over the 11Krange studied. Results of tests on specimens of types 304 and
321 were nearly the same at 593C (1100 OF) in air.
8-19. Type 309S Stainless Steel: Effect of Grain Size on Fatigue

Crack Growth Rate
~K, ksi
10
20
40
in.1/2
100
r - - , - - - r - - - , - - - r - - - - - - - - - - , 10- 3
10- 2
10-4
Type 309S
Testing
frequency,
Hz
10
15
20
25
30
Grain size
45 fJm 480 fJm
-0-
-0-
10- 6
10- 7
10
20
40
100

Fatigue crack growth rates for annealed type 309S stainless steel
for two grain sizes, at frequencies from 10 to 30 Hz and an R
ratio of 0.05 at room temperature in air.
Types 309S and 3 lOS stainless steel are the low-carbon versions of types 309 and 310. They have higher chromium and
nickel contents than type 304 and consequently have better corrosion resistance and more stable austenite than type 304. Fatigue crack growth rate data have been reported by Thompson
for tests made at room temperature on compact specimens from
plate of type 309S and the L-T orientation after heat treating to
a grain size of 45 Jlm in one set and 480 Jlm in the second set.
Specimens with the smaller grain size had substantially higher
yield and ultimate tensile strengths than the specimens with the
larger grain size. Fatigue crack growth rates were obtained on
tension-tension loading at frequencies from 10 to 30 Hz and at
an R ratio of 0.05. The results are plotted above. These data
provide further evidence that a wide variation in grain size, and
the associated variation in strength level, does not affect the results of fatigue crack growth rate tests.
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
Metals, Metals Park OH, 1982. P 126
215
216
8-20. Type 31 OS Stainless Steel: Effect of Temperature on Fatigue

Crack Growth Rate
~K, ksi in. 1/ 2
20
40
.-------r---r-..---..---.......,."""T""'l~---.______.
.,
U
>
("l
10-4
10- 3
--
E
E
Base metal
-196, -269C
(-320, -452F)
--.,'
"C
co
"C
......co
.,
>
("l
10- 5
...
.c
--.5
z
-"C
een
co
"C
-"o
co
...
.,
("l
10- 4
:::l
en
'':;
co
u..
Type 3105
20
40
60
100
Stress-intensity factor range, ~K. MPa m 1/2

Fatigue crack growth rates for annealed type 31OSstainless steel
at 22, -196, and - 269 0 C (72, - 320, and -452 0 F), 10 to 28 Hz,
and anR ratio of 0.1, with corresponding data for SMAW welds
with type 316 filler metal.
Because of its high nickel content, type 3lOS stainless steel is

completely stable at all cryogenic temperatures and with any
amount of cold working. Therefore, it is often considered for
cryogenic applications that require a high degree of austenite
stability on thermal cycling and strain cycling. Fatigue crack
growth rate at various temperatures is illustrated above.
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
8-21. Type 316 Stainless Steel: Growth Rate of Fatigue Cracks

in Weldments
liK, ksi in. 1/ 2
10-2
20
40
60
100
4 X 10-4
Ql
10-4
13
>
u
'<,
E
E
z'
~
<0
10- 3
Ql
13
>
u
"C
'<,
fl'
E
.~
z'
"C
s:
..
'<,
<0
"C
Cl
10- 5
E
u
Ql
:::J
Cl
'.J
<0
u..
10- 4
Type 316
593C (1100F)
10- 6
20
40
60
100
Stress-intensity factor range,

liK, MPa . m 1/2
Fatigue crack growth rates in type 316 base metal and

weld metal in the unirradiated and irradiated conditions
at 593 C (BOO OF)in air [fluence 1.2 X 1022 n] cm-, >0.1
MeV at 410C (770 OF)].
Results of fatigue crack growth rate tests on weldments of type 316 stainless steel have
shown that the crack growth rates in the weld metal are generally no higher than in the
base metal and may be somewhat lower at elevated temperatures (Shahinian, Smith, and
Hawthorne). The curve shown above for unirradiated weld metal tested at 593C (1100 OF)
represents fatigue crack growth rates substantially lower than those for the unirradiated
base metal at any given I:>.K level (Shahinian). The weld was produced by the submerged
arc method using type 316 welding rod. Weldments were stress-relief annealed at 482C
(900 OF). Specimens were single-edge-notch specimens for cantilever loading and were
tested at 0.17 Hz and at an R ratio of zero. Irradiation slightly reduced the fatigue crack
growth resistance of the weld metal, but its fatigue crack growth resistance was better than
that of the unirradiated base metal.
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds .. American Society for
217
218
8-22. Type 316 Stainless Steel: Fatigue Crack Growth RatesAged vs Unaged
AK. ksi . in.1/2
10
60 80
20
40
10 2 ."......---r------,;-------r--,--..,
Type 316, cold worked,

0
0
tested at 593 C (1100 F )
III
EE
10- 2
z
~
"lJ
III
10- 1
10- 2
--.S
--..
>
u
J!l'
E
..c:
10- 3
e
;
"lJ
"lJ
E
u
III
:J
10-4
..
en
.;;
u..
10- 5
10- 4
0
Ii.
10- 5
v
L-_....L-
10
--l.
20
Aged Hold time

Zero
0.1 min
1.0 min
10.0 min
10- 6
80 100
...L-_--l._......L~
40
60
Stress- intensity factor range, AK. MPa . m 1/2

Effect of exposure at 593 C (1100 OF) for 5000 h, and hold times during
cycling, on fatigue crack growth rate of20% cold worked type 316 stainless steel at 593 C in air.
Results also have been reported by James for fatigue crack growth rate
tests in 20% cold worked specimens of type 316 stainless steel which were
cycled at frequencies of 0.0055 to 6.66 Hz, at 538C (lIDO OF) and at an R
ratio of 0.05. Over the 11Krange studied, the fatigue crack growth rates were
highest for the specimens subjected to the lowest cyclic frequency.
Metallic Structural Material, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American Society for
8-23. Type 316 Stainless Steel: Fatigue Crack Growth RatesEffect of Aging
.1K, ksi in. 112
20
40
60 80
10
10-1.......- --r-----.-------,----.--,....,
10-3
10-2
Q)
-.~
593C (1100F)
E
E
10- 4
'"
Q)
"tl
...e
U
u>'<,
Q)'
..c
10-3
.~
een
"tl
'<,
'"
"tl
u
u
10-5
Q)
:::l
en
';:;
'"
Type 316
593C (1100F)
LL.
10-4
Unaged Aged
...
Hold time
Zero
0.1 min
1.0 min
10-6
10-5
10
20
40
60
80 100
Stress-intensity factor range, .1K, MPa . m 1/2

Effect of exposure in air at 593C (1100 OF) for 5000 h, and hold times,
on fatigue crack growth rates for annealed type 316 stainless steel at
593C in air.
219
220
8-24. Type 316 Stainless Steel: Effect of Temperature on Fatigue

Crack Growth Rate
dK, ksi in. 1/ 2
10
20
40
60
100
10- 4
10- 3
Ql
1)
>u
10- 5
--.5
--
2"
"'C
III
"'C
10-4
10-6
Type 316
Cold worked
10- 5
10
20
40
60
100
Stress-intensity factor range, dK, MPa . m 1/2

Fatigue crack growth rates of20% cold worked type 316stainless
steel for various temperatures. Curves are averages for L-T and
TL specimens at each temperature in air; 3 Hz at 24C, 0.67 Hz
at elevated temperatures; R = 0.05.
Results of tests on compact specimens of 20% cold worked

type 316 stainless steel at frequencies of 0.67 and 3.0 Hz and at
an R ratio of 0.05 are summarized in the above graph. Similar
results have been reported by Shahinian for tests on cold
worked type 316 at 427C (800 OF).
8-25. Type 316 Stainless Steel: Effect of Cyclic Frequency on

Fatigue Crack Growth Rate
dK. ksi in. 1I2
10
20
30
40
50
60
10- 4
0.0067 Hz----...
II>
10- 3
>
o
--
)7 ~0.67"'
E
E
z
~
'"
0067
"C
...eai
...;:
e
"'-y' //J
'~
II>
10- 5
s:
Cl
"'..u'""
1,1
10- 4
>
u
--.5
z
"C
<,
'"
"C
U
II>
::>
Cl
.;;
'"
u,
Type 316
538C (lOOOF)
10- 6
10
Stress-intensity factor range, dK, MPa m 1/2
Effect of variation in cyclic frequency on fatigue crack growth rate of
annealed type 316 stainless steel in air at 538C (1000 OF) and an R ratio
of 0.05.
As may be observed above in tests at frequencies in the range from

0.0067 to 6.67 Hz at 538C (1000 OF), the trend is for the crack growth
rate to increase as the frequency is decreased, but there is more scatter
than for type 304.
In studying heat-to-heat variations in fatigue crack growth rates for
specimens from three heats of type 316stainless steel, James has shown
that the spread from high to low values of fatigue crack growth rates is
no greater than that represented by a factor of 2.6 over the range of 11K
values studied. One heat was produced by air melting, another by vacuum arc remelting, and a third by double vacuum melting.
Source: J. E. Campbell, "Fracture Properties or Wrought Stainless Steels," in Application or Fracture Mechanics for Selection of
Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood. Eds . American Society for
221
222
8-26. Type 316 Stainless Steel: Fatigue Crack Growth Rate in the
Annealed Condition
6K, ksi . in. 1I2
10
20
30
40
60
80 100
10- 3
/1
370C (700.F)
---t!
10- 2
""1/;/
I ,
482C {900
/1
,1
p'"
//'
fi,' I
:,/
593C
.'
(1100F) /
l'
10- 4
..
>-
"I
y: /:
.5
z
~
./ ;,'I
:i
"
I'"I
/ /'1
10- 5
10- 4
316
' - -_ _'-_J.----'_L-l......J-J....I-J..J
10
20
30
40
60
10- 6
80 100
Stress-intensity factor range, 6K, MPa m 112
Effect of testing temperature on fatigue crack growth

rates for annealed type 316 stainless steel tested in air at
0.17 Hz and an R ratio of O.
Most of the fa tigue crack growth rate testing on type 316 stainless steel has been oriented toward its use in components for nuclear reactors, but the data also are applicable
to design of equipment for fossil fuel power stations, petrochemical refineries, and chemical plants. Its improved yield strength compared with that of type 304 stainless steel is an
advantage for these applications. The austenite stability in type 316 is greater than that in
type 304, so it is advantageous to use type 316 rather than type 304 for critical cryogenic
applications. Effects of elevated temperature on crack growth rate are summarized in the
graph above.
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American Society for
8-27. Type 316 Stainless Steel: Effect of Environment (Sodium,

Helium, and Air) on Cycles to Failure
10.0
8.0
DATA
POINTS
6.0
4,0
FATIGUE
TEST MEDIA
Na (10 ppm OF 02)
AIR
HELIUM
t'
2.0
ex:
l-
V>
1.0
...J
u 0.8
>u 0.6
0.4
EXPOSED SPECIMENS
FATIGUED IN SODIUM
FATIQJEDIN AIR
FATIGUED IN HELIUM
EXPOSURE MEDI~
0.2
Na
ue ppm OF 02)
EXPOSURE CONDITION
286 hrs AT 92S K
0.1
102
8 103
8 104
CYCLES TO FAILURE
Effect of environment on fatigue characteristics of type 316 stainless steel at 92SK;

based on cyclic strain and cycles to failure.
Source: R, H. Cook and R. P, Skelton, "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
in Source Book on Materials for Elevated-Temperature Applications, Elihu F, Bradley, Ed" American Society for Metals, Metals
Park OH, 1979, P 84
223
224
8-28. Types 316 and 321 Stainless Steel: Effects of Gaseous

Environments on Fatigue Crack Growth Rates
aK, ksi . in.1I2
8 10
20
10-3 r----,,-..----....
40
60
100
---...,.:.-......:.=----..:..;
Type 316
649C
(1200F)
">
10-5
Roomair
E
E
z
~
."
Type 316
2SoC 177F)
10-'
Dry air
Wet nitrogen
Dry nitrogen
fi
f!
.c
"
>
-!:!
.~
12tn
.><
<0
."
f!
":tn>
.~
10-6
Types 316 and 321
u,
2SoC (77F)
Room air
Wet air
10- 5
Types 316 and 321

649C (1200F)
Dry nitrogen
Dry argon
10
20
40
60
100
Stress.intensity factor range, 6K, MPa m 1/2
Effect of gas environments on fatigue crack growth rates for types 316 and 321
stainless steel at 25 and 649C (77 and 1200 OF).
Fatigue crack growth rate data at 25C (77 OF) show that crack growth
rates increased slightly with increased humidity when oxygen was present but
that high humidity in an inert gas had no significant effect. Fatigue crack
growth rates in room air at room temperature were the same for types 316
and 321 stainless steel. Furthermore, in tests at 649 C (1200 OF)in dry nitrogen, fatigue crack growth rates for types 316 and 321 also were the same. In
air, however, fatigue crack growth rates in type 316 specimens increased by a
factor of about 22 over rates in an inert environment at the same temperature.
Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American Society for
8-29. Type 321 Stainless Steel: Effect of Hold Time on Fatigue

Crack Growth Rates
.6K, ksi . in. 1/ 2
10
10-'
20
40
60
80
10- 3
Type 321
593C (1100F)
10- 2
..
U
>
u
10-'
E
E
.
"
10- 3
>
.~
e
'"
""eu
..
10- 5
..
"
10-'
::I
'"
.~
u,
10- 6
Unaged Aged
10- 5
...
Hold time
Zero
0.1 min
1.0min
10-
10- 6
'--_-'-
10
..J....
20
'
.......L_ _.l----JL-..J
40
60
80 100
Stress-intensitv factor range,ll.K, MPa . m'l2
Fatigue crack growth rates for annealed type 321 stainless steelunaged
and aged at 593 C (1100 OF)for 5000 h and tested in air with continuous
sawtooth waveform (0.17 Hz), with 0.1 and 1.0-min hold time at anR
ratio of 0 at 593C (1100 OF).
Results oftests by Michel and Smith on specimens of annealed type

321 stainless steel that had been aged at 593C (1100 OF) for 5000
hours and then tested at 593C have shown that long-time exposure at
the service temperature does not reduce the fatigue crack propagation
resistance in air. Aged specimens tested with zero holding time had
lower crack growth rates than corresponding specimens that were not
aged (see above graph). Fatigue cycling with holding times of 0.1 and
I,D minute on each cycle increased the crack growth rates slightly, as
shown in the figure,
Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels," in Application of Fracture Mechanics ForSelection of
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H, Underwood, Eds., American Society for
225
226
8-30. Type 403 Stainless Steel: Effect of Environment on Fatigue

Crack Growth Rate
~K,
ksi in. 1/ 2
10
20
30 40
60 80
10-3 rr----,-----,--,--...........-,------,
I
/
II
/
10-4
10-5
Air
Type 403
In H20
III
"1:J
pH 7, 25C
pH 10, 25C
-- - -
pH 7, 100C
pH 10, 100C
10- 6
In 1M NaCI solution
- - - - pH2tol0, 100C
10- 5 L..-_ _---L_ _L.--L---L----I---L...J......L.J.
10
20
30 40
60 80 100
--'
Stress-intensity factor range, ~K, MPa . m 1/2

Fatigue crack growth rates in type 403 stainless steelin air, water,
and aIM NaCI solution at 10 Hz and an R ratio of 0.5.
Exposure to water at 25C (77 OF) resulted in intermediate

crack growth rates between those in air and those in water at
100 DC, as shown on a different scale in the above graph.
Tests in the 0.01 M (molar) and 1.0 M sodium chloride solutions were made with the solutions at pH levels of 2, 7, and 10
and with an open circuit. Fatigue crack growth rates in 0.01 M
sodium chloride at pH 10 and 100C were the same as those in
water at 100 "C. At lower cyclic frequencies, the fatigue crack
growth rates were higher than at 40 Hz at 6.K values above 20
MPa m 1/2 (18 ksi in. 1/2). For tests in the 1.0 M sodium chloride solution at 100C (212 OF) (see graph), fatigue crack
growth rates were the same as for water at the same temperature.
At 100 C (212 OF), fatigue crack growth rates in 1.0 M sodium
phosphate solution at pH 10 and at 10 and 40 Hz and in 1.0 M
sodium silicate at pH 10 and at 10 Hz were practically the same
as those in air.
Metallic Structural Materials. James E. Campbell, William W. Gerberich and John H. Underwood. Eds., American Society for
Metals. Metals Park OH, 1982. P 147
8-31. Type 403 Modified Stainless Steel: Scatter of Fatigue Crack

Growth Rates
'--_-'-_-'-......L.......L.....L....w....LJ..._ _- ' - _ - ' -
20
40
60
100
----I
10-6
200 300
Stress-intensity factor range, ll. K, MPa . m 1/2

- - - - Heat 484 in room temperature air
Heat 634 in room temperature air
Heat 933 in room temperature air
Heat 933 in 271C (520F) distilled water at
8.3 MPa (1200 psi)
Upper boundaries of fatigue crack growth rate scatter bands for
three heats oftype 403 modified stainless steel in the heat treated
condition, tested at 10 Hz and an R ratio of 0.083 or 0.067.
The curves representing the upper boundaries of the scatter

bands of the fatigue crack growth rate data indicate that there is
some heat-to-heat variation in fatigue crack growth rate properties for these heats. Furthermore, exposure at 27 1C (520 OF)in
distilled water at a pressure of 8.3 MPa (1200 psi) increased the
fatigue crack growth rate.
227
228
8-32. Type 422 Stainless Steel: Fatigue Crack Growth Rates in

Precracked Specimens
Koff' ksi in. 1/2
10
20
40
60 80 100
200
71e (160F)
57e (135F)
a>
10-
.:.::::
~
"tl
Type 422
~_---l
20
---L_---l_....L.........L.
l..-
..J
10-5
200
Fatigue crack growth rates in precracked round rotating beam specimens of type 422 stainless steel in 4.5% NaCl solution at room and elevated temperatues, 10Hz, and an R ratio of-I.
Type 422 stainless steel contains nickel, molybdenum, and tungsten,

as well as 12%chromium to improve properties. The effects of sodium
chloride solutions and elevated-temperature exposure on fatigue crack
growth rates were determined by Eisenstadt and Rajan in tests of
notched round rotating beam specimens in which the numbers of test
cycles were marked by minor stress interruptions that produced marking rings. Calculations for maximum stress-intensity factors were
based on equations for solid round bars subjected to bending loads.
The material for these tests apparently had been heat treated to a yield
strength of approximately 827 MPa (120 ksi). The specimens were one
inch in diameter in the test sections. Each specimen was rotated at 600
cycles per minute (10 Hz) while at constant load with the salt water
solution flowing over the notched section. Tests with several concentrations of salt solution indicated that the maximum corrosive effect
was obtained with the 4.5% solution. Results oftests with specimens in
the 4.5%sodium chloride solution at room temperature, 57C (135 OF),
and 71C (160 OF) are shown above. Increasing the temperature of the
solution substantially increased the fatigue crack growth rates.
Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood. Eds., American Society for
Metals, Metals Park OH. 1982, P 150
8-33. Type 422 Stainless Steel: Fatigue Strength-Longitudinal

vs Transverse
CLASS II
(Crucible 422)
turbine-wheel forgings, 12 to 30 inches in diameter,
ksi
1900 F (l 040 C)oil quench + 1200/1400 F(650/760 C)
100..--------------------------.
UNNOTCHED
x
80-
60r-
1-----------... VIBRATING
NOTCHED
401-
Kt = 2.1
60 notch
0.030inch
root radius
CANTILEVER TESTS
FOR 108 CYCLES
x 0% delta ferrite
.x
o
5% delta ferrite
15/16% delta ferrite
A 20% delta ferrite
20 .....
I
20
40
60
80
TRANSVERSE FATIGUE STRENGTH
100 ksi
Transverse fatigue strength as related to longitudinal fatigue strength for type

422 stainless steel, including effects of varying amounts of delta ferrite.
Source: J. Z. Briggs and T. D. Parker, "The Super 12%Cr Steels," in Source Book on Materials for Elevated-Temperature Applications, Elihu F. Bradley, Ed.. American Society for Metals, Metals Park OH, 1979, p 121
229
230
8-34. Type 422 Stainless Steel: Effect of Temperature on

Fatigue Strength
CLASS II
(Crucible 422) (VacuumMelted)
ksi
140
15% delta ferrite

1800 F (980 C)oil quench + tempered
to a tensile strength of 131/138 ksi
130 f120 ,....(/)

(/)
110 ,....-
IJJ
a:: 100 "-
J-
(/)
80 70 90
601()~
'.,""'-ee_
.......
room temperature
-.-
700 F(370 C)
...
ROTATING CANTILEVER-BEAM TESTS

I
105
106
CYCLES
101
108
SON curves for vacuum-melted type 422 stainless steel with 15% delta ferrite,
showing effect of temperature on fatigue strength.
Source: J. Z. Briggs and T. D. Parker, "The Super 12% Cr Steels," in Source Book on Materials for Elevated-Temperature
Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH, 1979, P 121
8-35. Type 422 Stainless Steel: Effects of Delta Ferrite on

Fatigue Strength
CLASS II
(Crucible 422)
3/4 -inch-diameter bar stock or 3/4 -inch-thick plate
1900 F (1 040 C)oil quench
ksi
110'
100
en
en
LLI
0:
I-
90
en
% DeltaFerrite TensileStrength, ksi

155
160
.... 140
5% Ferrite,Tempered 1150 F
(620 C)
~
15~20
'~~v_
~...
00% Ferrite,Tempered 1150 F
--.. . . . . .-v..
-~
"-.. 0
(620 C)
o
0.:
15/20% Ferrite,Tempered 1200 F(650 C)
0-
80
V;;
lr-+
0-
'LONGITUDINAL ELECTROMAGNETIC
CANTILEVER-BEAM TESTS
70
lOS
107
CYCLES
S-N curves for type 422 stainless steel, which demonstrate the adverse effects of
delta ferrite on fatigue strength.
Source: J. Z. Briggs and T. D. Parker. "The Super 12%Cr Steels," in Source Book on Materials (or Elevated-Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH. 1979, p 121
231
232
8-36. 17-4 PH Stainless Steel: Fatigue Crack Growth Rates in Air

vs Salt Solution
.o.K, ksi in.'/2
10
100
10-2
I
'---Hll00
,
R =0.05
'I
1 min hold
Salt soln
,~
I~
10-4
I
I
10- 3
il!
'"
~u
"5,
10- 4
."u..
10-6
10- 5
174 PH
'--
-'-
-'-
10
100
...1..1
10-1
Stress-intensity lactor range, .o.K, MPa m1/2
Fatigue crack growth rates in WOL specimens of 17-4 PH stainless steel in the HI050 and
HllOO conditions in room temperature air and in a 3.5% NaCI solution,
Results of fatigue crack growth rate tests on specimens of 17-4PH stainless steel under
comparable conditions are presented here. Those specimens that were tested in the HI 050
condition at a stress ratio of 0.67 with a one-minute holding period at maximum load in
each cycle had the highest fatigue crack growth rates (as for 15-5PH) in the upper levels of
I:!.Kvalues. Specimens in the H II 00 condition tested in a salt solution with a one-minute
holding period, however, had fatigue crack growth rates only slightly higher than those of
comparable specimens tested in air with continuous cycling.
Metals. Metals Park OH. 1982, P 156
8-37. 15-5 PH Stainless Steel: Fatigue Crack Growth Rates in Air

vs Salt Solution
LlK, ksi in. 1/2
10
100
)
"
Hl~50 ~I //'I
R - 0.67
I:/,
1 min hold
I 1//
I/Il
RT.ir
,,1
l,' l'
'li
10- 3
..
Hll00--.......,'
R = 0.05
~
1 min hold
"
I
I
'I
Salt soln
."
!l'
VI
'I1 ,r--HllOO
R = 0.05
.t::
e
'"
""ut!
..
..
---Hl050
R = 0.05
10Hz
Sine wave
RT air
10-'
I'
TO
1 min hold
RT air
u,
10- 6
H 1100 :----""'/'
R = 0.05
10 Hz
I
Sine wave
RTair
10-5
155 PH
1-
--L.
...1...
10
100
.J1O-
Stressintensity factor range, LlK,MPa m 1/2
Fatigue crack growth rates in WOL specimens of 15-5 PH stainless steel in the HI050 and
HllOO conditions in room temperature air and in a 3.5% NaCI solution.
For specimens in the HI 050 condition, increasing the R ratio from 0.05 to 0.67 and incorporating a one-minute holding period at maximum load in each cXcle substantially increased the
crack growth rates at LiKvalues over 40 MPa m 1/2 (36 ksi- in. I 2). For specimens in the H 1100
condition, exposure to a salt solution environment during tests with a one-minute holding period at maximum load increased the fatigue crack growth rates over those of specimens tested in
air with one-minute holding time or with continuous cycling (see graph).
Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood. Eds . American Society for
Metals, Metals Park OH. 1982, pISS
233
234
8-38. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates

at Room Temperature
~K. ksi . in.1/2
10
20
40
60 80100
10-2
200
4 X 10-4
Gl
>
u
E
E
---
10- 4
-e
Gl
l'O
.....
"tl
>
U
---
Gl'
..
.~
l'O
.s:
10- 3
Z'
;:
"tl
l'O
"tJ
Cl
E
u
Gl
:::l
Cl
';;
l'O
u..
10-5
PH 13-8 Mo
H1100
10- 4
20
40
60 80 100
200
Stress-intensity factor range, ~K. MPa m 1/2

Fatigue crack growth rates in cantilever beam specimens of PH
13-8 Mo (HII00) stainless steel, at L-T orientation, 0.17 Hz, and
an R ratio of 0, in room temperature air. Data are based on the
stress-intensity-factor range as shown.
Source: J. E. Campbell. "Fracture Properties of Wrought Siainiess Steels," in Application of Fracture Mechanics for Selection of
8-39. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates in Air

and Sump Tank Water
toK, ksi in. 1/2
6
8 10
20
40
60
10-2
R = 0.3
STW
L-T
10- 4
10- 3
10- 4
10- 6
PH 13-8 Mo
H1000
8 10
20
Stress-intensity factor range, toK, MPa m 1/2

Fatigue crack growth rates in compact specimens of PH 13-8 Mo
stainless steel in the HI 000 condition for room temperature tests
at I Hz, R ratios of 0.08 and 0.3, L-T and T-L orientations, in
low-humidity air (LHA) or sump tank water (STW).
Effects of increasing the load ratio,R, on fatigue crack growth

rates in low humidity air (LHA) in sump tank residue water
(STW) for specimens of PH 13-8 Mo (H 1000)are shown above.
The highest fatigue crack growth rates in this series were obtained on specimens tested at an R ratio of 0.3 in STW. Increasing the load ratio from 0.08 to 0.3 had a marked effect on the
growth rates.
235
236
8-40. PH 13-8 Mo Stainless Steel: Fatigue Crack Growth Rates at

Subzero Temperatures
~K,
8 10
ksi . in. 1/ 2
20
40
60
100
10- 4
10- 5
Q)
>
u
'-:
.!:
1:1
"nl
1:1
10-6
PH 13-8 Mo
H1000
10- 7
8 10
20
40
60
100
Stress-intensity factor range, ~K. MPa . m1/2

Fatigue crack growth rate scatter band for compact specimens from rolled bar and
extrusions of PH 13-8 Mo stainless steel in the HIOOO condition for room temperature
tests in low-humidity air and in sump tank water at frequencies of 1 and 6 Hz and anR
ratio of 0.08 for L-T and T-L orientations.
Fatigue crack growth rate data for room temperature tests on specimens from
rolled bar and extrusions of PH 13-8 Mo (H 1000)stainless steel make up the scatter
band in the above graph. Specimens of L-T and T-L orientations were tested in
low-humidity air and in sump tank residue water at frequencies of I and 6 Hz and at
an R ratio of 0.08. Under these conditions, variations in frequency and environment
had little effect on fatigue crack growth rates. For tests at -54C (-65 OF),the rates
of fatigue crack growth were lower than those at room temperature over most of the
~Krange.
237
8-41. PH 13-8 Mo Stainless Steel: Constant-Life Fatigue Diagram

Minimum stress, ksi
-150
-100
100
+50
150
200
1600
1400
200
1200
'"
a..
:2 1000
150
Jl
100
.S
x
'"
:2
",'
...~'"'"
E
E
800
S)<::l l'l'b
'x
'"
:2
'li~
:l
600
S)<::l
(c)
\e"""
.,-,;
'b~
~fl,
Axial fatigue
Unnotched specimens
Longitudinal and
400
50
transverse or-
200
ientations
o '--_-'--_...l-__....J...._----L_--''------''L-_-'--_-L.._--'--_----L_--'_ _-'----_...l-_-'
-1200 -1000 -800
-600
-400
-200
+200
400
600
BOO
1000
1200
1400
1600
Minimum stress, MPa

Constant-life fatigue diagram for PH 13-8 Mo stainless steel, condition HI ODD.
:l
238
8-42. Types 600 and 329 Stainless Steel: S-N Curves for Two
Processing Methods
1000
800
600
'"
o,
::2:
I
I
.~
.L
400
IV
I I I.
AISI 329 (electroslag remelted)
Vi
'"
e
Ul
600 Steel (STAMP)-+--+--I
0.57
100
80
'\1'
60
0.51
40
'iii
-'"
",'
'"
e
Ul
200
20
100
10'
10'
10'
10'
Cycles to failure
Steel
STAMP 600
Electroslag-remelted 329
Tensile
strength,
MPa(ksl)
Yieldstrength
(0.2% olTset),
MPa (ksi)
760
(110)
630
(91)
600
(87)
500
(73)
Mechonlcal properties
Elongationin
Reduction
50 mm (2 ln.),
%
in area, %
26
54
29
65
Impact
energy,
strength,
J (Rlb)
MPa (ksl)
Fatigue
25
(18)
35
(25)
430
(62)
320
(46)
S-N curves showing test results and mechanical properties of STAMP-processed 600
stainless steel and electroslag-remelted AISI 329 stainless steel. Fatigue ratio
(0 107/Rm) for 600 steel: 0.57. Fatigue ratio for electroslag-remelted 329 steel: 0.51.
Source: Metals Handbook, 9th Edition, Volume 7. Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, p 549
8-43. Grade 21-6-9 Stainless Steel: Effect of Temperature on

AK, ksi in.1/2
10
20
40
60
21-6-9
10- 4
Q)
>
u
......
E
E
22 to -196C
/
(72 to -320F) /
-269C
(-452 F)
10- 3
2:
"t:l
......
III
"t:l
...oj..
III
s..
0
Cl
III
>
10- 5
/
/
or.
Q)
10- 4
......u
,E
2:
"t:l
......
III
"t:l
Q)
:l
Cl
'+:i
III
u.
/
10
20
10- 6
40
60
100
Stress-intensity factor range,

AK, MPa m 1/2
Fatigue crack growth rates in specimens of annealed 21-6-9 stainless steel at 22, -196 and - 269 0 C (72, - 320 and -452 0 F), 20 and
28 Hz, and an R ratio of 0,1.
Similar tests made with specimens of 22-13-5 stainless steel

showed fatigue crack growth rates that were nearly the same as
shown here for 21-6-9.
Source: J. E. Campbell, "Fracture Properties of Wrought Stainless Steels,"in Application of Fracture Mechanics for Selection of
239
240
8-44. Kromarc 58 Stainless Steel: Effect of Cryogenic

Temperatures on Weldments
dK, ksi in. 1/ 2
20
40
60
100
200
10- 3
10- 4
I
I
Kromarc 58
Base metal
Weld metal
......_ -::':::--_--:'-_....L.._----''--_ _.L------'
20
40
60
100
10- 6
200
Stress-intensity factor range, dK, MPa . m 1/2

Fatigue crack growth rates for solution treated Kromarc 58 base
metal in air at room temperature, and base metal and weld metal
at-269C (-452 OF) in liquid helium, at 10 Hz and anR ratio
of 0.1.
For the fusion zone of a gas tungsten arc weld made with
Kromarc 58 filler metals, the KIJJ) value was 156 MPa. m l / 2
(141 ksi- in. 1/2) at -269C (-452 OF).Fatigue crack growth rate
data for the base metal at room temperature and at - 269C and
for the weld metal at -269C are shown above. The data were
obtained on compact specimens at 10 Hz and at an R ratio of
0.1. Fatigue crack growth rates for tests in liquid helium were
lower than at room temperature at the same t::.Kvalues. Therefore, if room temperature crack growth rate data are used to
estimate crack growth at cryogenic temperatures, the estimated
values will be conservative.
8-45. Pyromet 538 Stainless Steel: Effects of Welding Methods

on Fatigue Crack Growth Rates
AK, ksi . in. 1/ 2
10
20
30 40
>
E
E
200
10- 4
",-SMAW
24 and -269C
(75 and -452 F)
Gl
......u
60 80 100
10- 3
Z
't:l
......
Gl
'"
't:l
>
u
......
/"GTAW
24C (75F)
!l
'"
..r:
.~
'"
.~
10- 5 z
~
'"
't:l
E
u
Gl
10- 4
:l
'"
.;:;
u..
'"
Pyromet 538 welds
20
30 40
60 80100
200
Stress-intensity factor range, AK, MPa . m 1/2

Fatigue crack growth rates in weld metal in Pyromet 538 stainless steel at room temperature and -269C (-452 OF) and at
10Hz.
The base metal was solution annealed prior to welding. One

set of welds was made by the gas tungsten arc welding (GT AW)
process with 21-6-9 filler wire, and the other was made by the
shielded metal arc welding (SMA W) process with IN 182 covered electrodes. Results of these tests are summarized in the
graph above. Specimens with SMA W welds had the same fatigue crack growth rates at room temperature and at -269 "C
(-452 F). Specimens welded by the GT AW process had higher
crack growth rates at -269 C than at room temperature. Examination of the microstructures near the fracture surfaces for the
specimens tested at - 269 C showed tha t there was 6 to 7% delta
ferrite (produced by welding) in the weld metals along with induced martensite. The SMAW weld metal was fully austenitic.
Metallic Structural Materials, James E. Campbell, William w. Gerberich and John H. Underwood. Eds., American Society for
241
242
8-46. Duplex Stainless Steel KCR 171: Corrosion Fatigue
KCR 171
Whit. wat.r
400
pH 4. IS
T ISOC
300
CI
Q.
_ 0 .......
(\/
0 ...
200
l;
~
100
6 Hz
t:>
20 Hz
100 Hz
Nf
167Hz
eyel ..
Rotating bending S-Ntests were carried out in 50C (122 OF) white water
at different frequencies (6, 20, 100,and 167Hz) for samples polished with 240
grit emery paper and the results obtained are presented in the above S-N
diagram. The results thus far 0 btained for the two highest frequencies appear
to fall on the same S-N curve, and the indication is that this curve would
present a quite horizontal fatigue limit. In the short life regime (N,: 106
cycles), the results suggest that decreasing the frequency below 100 Hz displaces this portion of the S-N curve to shorter lives without significantly
changing its slope.
Source: M. Ait Bassidi, J. Masounave and J. I. Dickson, "The Corrosion Fatigue Behaviour in White Water of KCR 171," in
Duplex Stainless Steels, R. A. Lula, Ed., American Society for Metals, Metals Park OH, 1983, p 455
9-1. Grades 200, 250, and 300 Maraging Steel: S-N Curves for
Smooth and Notched Specimens
1500r - - - - - - r - - - - -.....----~----_,
CZZl
Em
css
'"
~
18Ni(300)
18Nj(250)
18Ni(200)
_ _-+-
1000
200
....=::j
150
'iii
c..
.:.l
v.
...e'"
If)
500
L.-
104
--'-
105
---L
106
Number of stress cycles
L.-
107
0
8
10
~
Fatigue properties of maraging steels are comparable to those of

other high-strength steels. Smooth-bar and notched-bar fatigue properties for I8Ni(200), I8Ni(250), and I8Ni(300) grades are summarized
in the S-N curves shown above, Fatigue crack growth rates in maraging steels obey the da] dN= (t1K)m relationship commonly observed in
steels and are similar to those of conventional steels. Improved fatigue
properties can be obtained by shot peening and by nitriding,
Park OH. 1978, p 451
243
244
92. Grade 300 Maraging Steel: Fatigue Life in Terms of Total Strain
18% Ni morolling (300)
-1
10
,,
\
--
/(J'f/E
-2
10
--..
- LOAD CONTROL
\
'\
- : TOTAL
~\
...
A..-
-)-Trrr~ ..c
ELASTIC
PLASTIC~\
h
-3
10
L-J....l.l..JLlJJ.U-l.....LJ..J.JJJ.U.--I....LJ.J.LWL--I...1.LI..JWJJ..--L..LUJ..LJ.ll..-l-u..uJJ.lJ
10
Superposition of elastic and plastic curves gives fatigue life in terms

of total strain. An actual example for this method of determining fatigue life is presented above for a maraging steel.
Source: Marc Andre Meyers and Krishan Kumar Chawla. "Mechanical Metallurgy: Principles and Applications," Prentice-Hall,
Inc.. Englewood Cliffs NJ, 1984, p 700
10-1. Fatigue of Cast Irons as a Function of

Structure-Sensitive Parameters
50
20
iii
a:
10
...J
I&.:
"B" BAR
BAR
~ S~
.
.1
10
MEAN FATIGUE LIMIT VERSUS (BHN)

Fatigue of cast irons as a function of structure-sensitive parameters: Bhn, elastic modulus (Eo) and damping capacity (D).
Walter has shown that the fatigue properties of irons are highly dependent on
volume of graphite and its morphology and distribution, as well as the matrix structure. He was able to reduce these factors to some easily measurable parameters, Eo,
D, and Bhn, which gave good correlation with fatigue properties over a rather wide
range of irons (see graph). It is reasonable that these parameters relate to fatigue
performance, since they are measures offatigue-related properties. Eo, the modulus
at very small strains, is controlled mostly by the volume of free graphite and to some
degree by the graphite shape. Since the graphite present detracts from the matrix
load-carrying area, the more graphite, the higher the stress on the remaining
matrix-thus lower fatigue performance. D, the damping capacity, is controlled
mostly by the graphite morphology and to some degree by the graphite volume.
Sharp-edged flakes are greater stress raisers than rounded-edge flakes and spheroids; thus the higher the D, the poorer the fatigue performance. Bhn is largely a
measure of the matrix hardness and, to some degree, ofthe graphite volume; thus
the higher the Bhn, the better the fatigue performance. These easily measured properties are put to good use in industry as specification means and process-control
criteria.
Source: D. H. Breen and E. M. Wene, "Fatigue in Machines and Structures-Ground Vehicles," in Fatigue and Microstructure.
245
246
10-2. Gray Iron: Fatigue Life, and Fatigue Limit as a

Function of Temperature
Temperature, of
200
400
400
600
800
1000
I
Fatigue life
Fatigue limit
350
- 50
- 40
300
250
IV
0-
:2
::;' 200
cil
150
, ".
-,
<,
.,
'-
_____0
17
100
---- -% - "
~
<,~
50
"
20
1o
100
300
400
500
600
Temperature,oC
o Notched specimens
200
Un notched specimens
Notched specimens.
stress based on net area
Composition: 2.84 C, 1.52 Si, 1.05 Mn, 0.07 P, 0.12 S, 0.31 Cr, 0.20 Ni, 0.37 Cu. (Ref 5)
Typical fatigue life for as-cast gray iron of the above composition (left). Effect of temperature on
fatigue limit for the same gray iron (right).
Park OH, 1978, P 21
247
10-3. Gray Iron: S-N Curves for Unalloyed vs Alloyed
34
Plain Iron
Alloy Iron. 1% Ni. 0.4% Cr.
0.6% Mo. 1.0% Mn
o
32
(f)
30
a.
0
0
0
..--
en
28
26
""",
'\.
230
220
210
200
190
180
<J)
.....
170
24
Knee
(f)
- /-----------1----0--
160
22
Endurance or
~
Fatigue Limit .,.
- ._--_.1. ________________
20
18
10'
en
(f)
<J)
(f)
L..
<U
o,
10 5
-L
10 6
150
140
130
10 7
Cycles To Failure
A fatigue crack starts in an area of high stress concentration after a large number
of loading cycles. It is always a brittle type of fracture even when occurring in
ductile metals. As the crack progresses it increases the stress concentration, and the
rate of propagation under the cyclic loading increases. When the cr~ss section of the
remaining metal becomes insufficient to support the maximum load, complete
failure occurs as it would under an excessive steady stress.
The number of stress applications that will induce a fatigue failure is less at higher
maximum stress values, and conversely a larger number of stress cycles can occur at
a lower maximum stress level before a fatigue crack is initiated. A plot of this
relation for a metal is called an S- N curve and relates the maximum applied stress to
the logarithm of the number of cycles for failure. When the number of cycles without failure exceeds ten million, the endurance life is considered infinite for bodycentered-cubic ferrous metals. The maximum stress that will allow this number of
cycles is established as the endurance limit, or the fatigue strength or fatigue limit.
Two typical S-N curves for a plain and alloy high-strength gray iron are presented
above.
Source: Iron Castings Handbook, Charles F. Walton, Ed.. Iron Castings Society, Inc.. 1981, p 246
.....
L..
(f)
248
10-4. Gray Iron: Effect of Environment
21
140
20 -
'ea.n
19
0
0
0
18
en
Fatigue
Strength
18.000
130
co
Q.
PSI
1124 MPa)
120
17
en
(JJ
Q)
....
+-'
(JJ
Q)
....
16
+-'
110
(.f)
(.f)
15
14.500 DSt
(100 MPal
100
14
10
10
10 7
10
Number of Cycles
The effect of various environments and corrosion inhibitors listed in the
table below on the corrosion fatigue properties of gray iron.
Environment
Fatigue strength
psi
MPa
Air
17,920
Water
14,560
3% sodium
chloride. . . . . . . . . . . . . . . . .. 5,600
1% borax
15,680
3% "Sobenite"*
17,920
3% sodium
carbonate
17,920
3% soluble oil.
17,920
0.25% potassium
chromate
17,920
Fatigue strength
reduction factor
124
100
1.23
39
108
124
3.20
1.14
1.00
124
124
1.00
1.00
124
1.00
* "Sobenite" is a mixture oj 10paris sodium benzoate to 1 part sodium

nitrite.
The corrosion fatigue program involved testing in air, a spray of demineralized water, and a spray of three-percent sodium chloride solution; additional tests were made with a demineralized water spray and various known
corrosion inhibitors. The S-N curves and table above indicate that both the
demineralized water and three-percent sodium chloride sprays reduced the
fatigue strength of a pearlitic gray iron. Of the various alkaline inhibitors and
soluble oils investigated, only borax was not completely effective for the
pearlitic irons. Annealed ferritic gray irons were similarly affected by the
sodium chloride solution.
Source: Iron Castings Handbook. Charles F. Walton. Ed., Iron Castings Society, Inc., 1981, p 255
10-5. Class 30 Gray Iron: Modified Goodman Diagram

Mean Stress. MPa
50
100
150
200
30 ~---"""T---r--""---T""""'---"""'"2
200
Area of Finite Life

150
20
I--------+---~;<;;;;;;;;;;;;;;;;+.;;;;;;ijit----~
'(ij
a.
o
o
o
100
eo
CL
iii
Ul
....~
(f)
50
Cl
:il
S
0
-50
-101'-
.....L.
10
.l...-
20
...J
30
Mean Stress. 1000 psi
A modified Goodman diagram relates the endurance limit to an

allowable working stress when it is superimposed on a steady
stress.
In many engineering applications, alternating stress is superimposed on a steady stress and requires special consideration. A
method of relating the effect of the combined static and alternating stresses on the endurance limit has been developed into the
Goodman diagram, of which a modified form is shown here.
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981,P 251
249
250
10-6. Class 30 Gray Iron: Fatigue Crack Growth Rates

Stress Intensity Factor Range. D.K in MPavm
10
20
40
60
100
Q)
o>-
Q)
o>-
~
o
~
E
s:
.~
10-'
'"
"0
Band for Wrought

Ferrite-Pearlite
Steels
2iIII
a:
s:
Load Range
(9
i3
2000
o 2200
2500
o 2500
6 3000
... 3300
10-8
'"
U
Ib (910 kg)
Ib (1000 kg)
Ib (1130 kg)
Ib (1130 kg)
Ib (1360 kg)
Ib (1500 kg)
'"
"0
10- 4 ~
III
a:
s:
(9
.:>L
10
20
30 40
60 80100
Stress Intensity Factor Range. D.K in ksi VIiicli

Fatigue crack growth rate. The endurance limit approach to
design utilizes fatigue data taken on smooth, defect-free test
specimens. For such specimens, fatigue crack initiation may
take 80 to 90 percent of the total lifetime while crack growth is
only 10 to 20 percent of the lifetime. Such flaws allow fatigue
cracks to initiate in a relatively small number of cycles so that
the lifetime of the component depends principally on the crack
growth rate. If the initial flaw size can be determined from experience or by utilizing nondestructive inspection and the critical
flaw size calculated using the fracture toughness value K,c' then
crack growth rate data may be used to calculate the number of
cycles required to grow a crack from an initial size to a critical
size where final fracture occurs.
Only limited fatigue crack growth rate data are available on
cast irons. These results are presented in the above chart for a
class 30 gray iron, where dol dN is the crack growth per cycle
and 6.K is the stress intensity range.
Source: Iron Castings Handbook, Charles F. Walton, Ed.. Iron Castings Society, Inc., 1981, p 250
10-7. Gray Irons: Torsional Fatigue for Various Tensile

Strength Values
MPa
1000 psi
60
400
58.000 psi (400 MPa)
50
en
en
300
40
(j)
"0
Q)
+-'
<U
Q)
li} 200
a::
'0
.~
E
24.000 psi (166 MPa)
Q; 100
~
14.000 psi (97 MPa)
..J
"0
<U
1000 psi
Q;
::J
Mean Stress
-100
Torsional fatigue strength for three levels oftensile strength with various mean
stresses.
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, p 253
251
252
10-8. Gray Irons: Torsional Fatigue Data for Five

Different Compositions
40
r---------------------.,
1.50
en
1.25 Q)
E 30
E
.I:
U
....OJ
.I:
1.0
.I:
0.75
-.J
....OJ
C
Q)
-.J
20
Q)
lO
L..
lO
L..
0.5
lO
....0
10
lO
I-
0.25 I-
O'--__ '---""'--_-'----_-'-----I._-'-----'-----'400
300
----l_ _----l
800 1000
500 600
--'
2000
Number of Cycles
Total length of six cracks (the first three cracks in each of two specimens of each iron)
as a function ofthe number of thermal cycles between 1100 and 400 C (590 and 200
C). Iron compositions are as follows:
Composition. %
Iron
B
C
0
E
Si
Mn
Cr
Mo
3.43
3.45
3.45
3.44
3.43
1.65
1.74
1.68
1.69
1.66
0.57
0.59
0.63
0.58
0.58
0.49
0.30
0.21
0.50
0.30
0.38
0.39
Ni
Cu
Sn
0.60
0.97
0.59
0.87
0.30
0.077
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, pp 288, 289
10-9. Gray Irons: Thermal Fatigue-Effect of Aluminum Additions

,..----r----,---,----,-----r----r----.5
Gray Iron
3% AI. 0.6% Si (65% Ferrite)
<>
2% Si
.!: .150
o 0.5% Cu. 0.5% Mo
D
o
c
A
2%
AI.
2%
Si
(4%
Ferrite)
.125
.175
+-'
.100 t---+----+---+---.,,>4F---t----+------l
Q)
....J
~
...ro
_ _--1
.075 t---+----+---.",..e..t----+-----.,.,.-=--
.050 I--------if------""--+---=-'f----+----+---'::=--'""=i
s:
g>
--+---t"7""'==------j4
10
12
Number of Thermal Cycles (X 1000)

Thermal fatigue resistance of different alloyed gray irons.
This graph shows results of a thermal fatigue test in which notched

disc specimens were alternately heated to 800 OF (425C) and cooled
to 200 OF (95C) in two fluid beds, demonstrating that a peariitic gray
iron containing 3.4% carbon and 2% aluminum was highly resistant to
thermal crack propagation.
253
254
10-10. Gray Irons: Thermal Fatigue-Effect of Chromium and

Molybdenum Additions
4
.15
1;..:-.,;.;;;,
..,. ,~~
~.
.10
3.7% C
3
Bridge Cracked
/~
~ ->
Unalloyed ____~
CrNi _ '?"' . ~ .
..-'
.--
.-1--.05
I ...-:.
CrNiMo0'"
CrMo(
en
.::.t.
~-.:
en
o ~
<Il
L.
'0
L.
---
3.2-3.3% C
s:
Q.
Q)
en
Q)
..c
o
c
I
Bridge Cracked
.15
..c
a.
Q)
CrNiMo /
'0
.10
'
.05
o '--__-L._ _- - '
o
100
200
.L-
300
0
600
J..._ _----l_ _----J
400
500
Number of Cycles
The depth of cracks resulting from the thermal cycling of gray irons
between 860 of (460C) and room temperature.
Alloying with molybdenum and chromium provided superior thermal fatigue resistance compared with irons that contained other alloying additions or no alloying at all. In
this case, the improved thermal fatigue resistance is believed to be directly related to the
higher elevated-temperature tensile strength and better stability of the chromiummolybdenum irons. However, it must be remembered that this improvement is related to
and dependent on the temperature cycle and base iron composition, as shown above. It
has also been indicated that the development of an acicular matrix structure, by adding
relatively large quantities of molybdenum and copper, supplies a less than desirable influence on thermal fatigue cracking.
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Casings Society, Inc., 1981, p 288
10-11. Gray Irons: Thermal Fatigue-Room Temperature

and 540C (1000 OF)
Tensile Strength. MPa
100
200
300
1500
<J)
Q)
0
>-
'0
1000
Q;
.D
E
:J
500
o Lo
' - -_ _---'
10
20
--'-
-'-
30
40
'---'
50
Tensile Strength. 1000 psi
Curves showing relation between the number of thermal cycles

for cracking and tensile strength at room temperature and 1000 F
(540 o q .
For good resistance to thermal fatigue, gray irons should have

high thermal conductivity, a low modulus of elasticity, high
strength at both room and elevated temperatures, and for
temperatures above 900 OF(500 "C), resistance to oxidation and
structural change. Because some of these properties are in opposition, a compromise must be made in selecting the most appropriate metal for each type of service. As the maximum temperature to which the gray iron is cycled and number of cycles
increase, the number and size of thermal fatigue cracks become
larger. The above curves illustrate the influence of room- and
elevated-temperature strength on the thermal fatigue resistance
of irons having similar carbon equivalents, thermal conductivities, matrix structures, and elastic moduli. Those irons with
higher room and elevated-temperature tensile strengths (achieved
by alloying) generally display higher thermal fatigue strength.
Source: Iron Castings Handbook. Charles F. Walton, Ed . Iron Castings Society, Inc. 1981, p 285
255
256
10-12. Gray Irons: Thermal Fatigue Properties-Comparisons

With Ductile Cast Iron and Carbon Steel
Maximum Cycle Temperature. F
1200
1400
1600
1o',---,--------,r--------,---,
z
d>
c
:;;; 10'
Steel
(J
Ductile
Iron
B
<f)
<Il
>-
(J
'0
Gray
AI
; 1m",
Q> 10'
.0
E
:J
16~0.,...0--------=-.L------~-----~~
700
800
900
Maximum Cycle Temperature. C
The above curves show the variation of the number of

cycles to cracking with the maximum temperature of the
cycles for gray iron, ductile iron, and carbon steel. Compositions of the four gray irons are as follows:
Analysis. %
Iron
A
B
C
D
Si
Mn
Cr
Mo
Other
3.43
3.49
3.48
3.50
2.37
2.37
0.60
2.38
0.78
0.84
0.88
0.83
0.22
0.24
0.23
0.30
0.32
0.22
0.20
0.77
0.21 Sn
2.37 AI
1.51 Cu
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, pp 286, 287
10-13. Cast Irons: Thermal Fatigue Properties for Six Grades
Ferritic Compacted Graphite
Pearlitic Compacted Graphite
Ferritic Ductile
Pearlitic Ductile
Alloyed Ductile
6
10'
10'
Number of Cycles
The number of thermal cycles required to produce thermal fatigue cracking in cast irons. Compositions are tabulated below.
%Mg
Analysis
%C
%5
%Mn
%P
Class 35 Gray Iron
2.96
2.90
0.78
0.07
Ferritic Compacted
Graphite
3.52
2.61
0.25
0.05
0.015
Pearlitic Compacted
Graphite
Ferritic Ductile
3.52
2.25
0.40
0.05
0.015
3.67
3.60
2.55
2.34
4.84
0.13
0.50
0.06
Pearlitic Ductile
Alloyed Ferritic
Ductile
0.030
0.030
3.48
0.31
0.05
0.07
Alloys
0.12Cr
0.030
1.47Cu
0.54 Cu
1.02Mo
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society. Inc., 1981, pp 393, 396
257
258
10-14. Ductile Iron: Effect of Microstructure on Endurance

Ratio-Tensile Strength Relationship
Tensile Strength. MPa
300
~o
J.
500
700
900
I
1100
1300
I
0.5 -o\'ox>
~.~ "\~t x Pearlitic
Ferritic.~.
o
.."
.o~ "x~x:
x
o ,
'"
II:
Q)
0000
0.4
::J
'U
~
\ K
Tempered Martensite
<,
j( .....
0.3
60
100
140
160
Tensile Strength. 1000 psi
In general the fatigue limit for ductile iron increases with tensile strength, but as with other ferrous metals, the increase is less
than proportional. The relation between the tensile strength and
the endurance ratio for the annealed, ferritic irons is different
from that of the irons with a matrix of pearlite or tempered martensite, as illustrated above.
10-15. Ductile Iron: Effect of Microstructure on Endurance

Ratio-Tensile Strength Relationship
50
75
100
125
150
175
0.51-----e-..... "ki~---T___+-----__1I_-----+__-----__+--_t
..o,
~
1ie
O.4I------_+_---'''--------'''r__+-----.-+--'''o,~
......I _ - - - - - _ + _ - - - - - _ _ + - - _ _ 1
::>
"0
<:
0.31------t------+-----t--------1r---'''''''=----+----j
200
Tensile strength, MP.
The influence of tensile strength and structure on the endurance ratio of ductile
iron is indicated in this graph. Endurance ratio is defined as endurance limit divided
by tensile strength. Because the endurance ratio of ductile iron decreases as tensile
strength increases, regardless of structure, there may be little value in specifying a
higher-strength ductile iron for a structure that is prone to fatigue failure. For tempered martensite ductile iron, the improvement in fatigue strength due to an increase in tensile strength is greater than for pearlitic or ferritic structures. This is
indicated in the graph above by the shallower slope for martensite.
Source: Metals Handbook, 9th Edition, Volume I, Properties and Selection: Irons and Steels. American Society for Metals, Metals
Park OH, 1978, P 45
259
260
10-16. Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades,
Using V-Notched Specimens
400'.---------,-------,-------,---------,
3S01-------1-------+------+---------1
F_I6G-40-18 .......odl
~v...otchod
3ool----------+------+-------t--------I
g.
2001--------=zilJ!l
ISOI--------+----==
20
loo'f--------f-------+-------1-----------l
106
Fatiguelift. cyeles
400,-------.-----------,--------,r--------,
3S01-----------lI----------l--------l'---------l
Pllrlitlc (80-66-06 -e...t)
46 V-notched
3001-----------1r--------l--------l---------l
40
.r.
-6 2SOI -- - - - - -+----:
~e
i!'
~o
:;
il. 2001-----------1r----"
ISOI---------1I--------f--------f---------j
20
1001-----------1r-------f-------1--------l
106
Fatiguelife. cycles
Top:S-N curves, including scatter bands, for annealed ductile iron. Bottom: Similar to above except for as-cast pearlitic ductile iron. All test
specimens were V-notched (45).
Park OH, 1978, P 43
10-17. Ductile Iron: S-N Curves for Ferritic and Pearlitic Grades,
Using Unnotched Specimens
.00r-------,.-------,--------,.-----------,
F.,rillc: (60.4o-18 annIId)
unnotc'*l
3501--------t------+------+--------I
3001--------t----::;
e:lE
'0 ]
2501--------t--
=
~
=
1
200
1501--------t------+------+--------I
20
1001--------+------+------1-------;
106
Faligue life. cYcles
.OOr--------,------.----------,r--------,
3501---------f~
3001--------;------'
t:.
'0
:lE
ii 2501--------;-------+------......,1--------1
go
a
~
J200
1501-------+------+--------1------.,
20
1001-------+------+--------1------.,
5:0'"'-------'<--------';-------:--',--------'
106
10'
Faltgue lIle,cycles
Grade
Tenolle
alrenl!h
MPa
kal
Unnolched
Endurance
UmII
Endurance
MPa
kat
rallo
6040-18
80-55-06
480
680
205
275
70
99
30
40
0.43
0.40
Notched
Endurance
Umil
MPa
kat
125
165
18
24
Endul'
ance
ratio
Slreaa
coneenlrallon
lacier
0.26
0.24
1.67
1.67
Top: Similar to upper graph on the opposite page, but here the specimens were unnotched. Bottom: Similar to lower
graph on the opposite page, but here the specimens were unnotched. Data in table pertain to graphs on this and the
opposite page.
Park OH, 1978, p 44
261
262
10-18. Ductile Iron: Fatigue Diagrams for Bending Stresses and

Tension-Compression Stresses
MPa 1000 psi
MPa 1000 psi
100
600
600
80
400
60
JOO
40
::;
~ 200
c:
400
200
Q)
5 100
20
-0
c:
-0
o
-20
-100
-40
-200
~--:-f\P'A"':ld'-?-:!:---!c:---f::---,~~I:---
1001
psi
-200
200
400
600
-JO
-40 L-_-:-'-:-_-=-~_--:-'-,----_-,-:-_-=-'-MP,
100
200
JOO
400
600
o
MPa
Mean Stress
Mean Stress
Fatigue diagrams indicating endurance limits for five grades of ductile iron under bending stresses (left) and
tension-compression stresses (right). Minimum properties of the irons are given in the table below.
Iron
No.
1
2
3
4
5
Min. Tensile Strength
Min. Yield Strength
1000 psi
MPa
1000 psi
MPa
Min. Elongation
Percent
55
61
72
87
102
38
42
50
60
70
36
41
51
61
72
25
28
35
42
50
17
12
7
2
2
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981, pp 344, 345 and 346
263
10-19. Ductile Iron: Effect of Surface ConditionsAs-Cast vs Polished Surface

Tensile strength. ksi
90 100
120 _':"';'::'_...:...r:.-....,....;.,~-, 50
350 n60r--_":,,,_-.--~......,.._--,
.........._r----r-..........
g
n8
'" 3001---+----+---i--7"'q...---f---+-----i
a..
::2:
45
40
.<:='
200 1----+---+--7"'~_+_---'
25
150
_ __'___ _-'-_ _
400
600
500
700
~
__L. _ _- - - - '
800
" ' - _ ___'__ _ __ '
900
1000
1100
Tests made on 10.6-mm (O.417-in.) diameter specimens. Fully reversed stress (R = -1).
Data given in the above graph show that the endurance limit for any given
strength level of ductile iron is significantly affected by surface conditions of
unnotched specimens. The endurance limit is much higher for the polished
specimens than it is for the as-cast specimens, which have relatively rough
surfaces.
Park OH, 1978, p 45
264
10-20. Ductile Iron: Fatigue Limit in Rotary Bending

as Related to Hardness
iii
60
0-
...
...
'E
::i
,.,
0)
';:; 40
<U
u,
"0
Q)
lD
30
<U
a:
360
20
100
----I-
...
'E
::i
Q)
300
::J
0)
';:;
<U
U.
.,....
0)
<U
0...
Q)
::J
------
.
..
.....

60
400
0
0
0
260
0)
.s
"0
Q)
lD
200
s0
a:
160
200
300
400
Hardness. Brinell
'iii
00
0
0
60
<U
0...
- 400
60 -
Q)
::J
0)
.~
u,
0)
40
r-
"0
Q)
lD
z-
<U
a:
30
100
...
'E
::i
A.
I
200
.~
..
...
'E
.,-
::i
360
Q)
::J
0)
';:;
<U
u,
300
0)
.s
"0
Q)
260 lD
z-
<U
300
400
600
600
700
a:
Micro-Vickers Hardness Number
Top: Relation between Brinell hardness and fatigue limit in rotary bending for ductile iron. Bottom: Relation between rotary bending fatigue
limit and matrix hardness for ductile iron.
Source: Iron Castings Handbook, Charles F. Walton. Ed" Iron Castings Society, Inc" 1981, p 347
10-21. Ductile Iron: Effect of Rolling on Fatigue Characteristics
Rolling
Pressures
(Pounds)
75
70
500
406
580
768
65
60
400
'(ii
o,
o
0
0
en
(J)
,,
55
50
<0
n,
~
,,
"
"",
"
u:;
83
"
45
"-----------j
,----------- - Unnotched. Unrolled
300
40
35
30
10'
Unrolled
10'
10'
200
10'
Number of Cycles of Stress (Log Scale)
Fatigue strength of ductile iron can be increased substantially

by cold working, especially when this method is applied to
stressed radii or notches. More than a 60% increase in the endurance limit was obtained with a rolling pressure that was insufficient to depress the surface a measurable amount. The improvement in fatigue properties obtained by various rolling
pressures on ductile iron is indicated above.
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc . 1981, p 348
265
266
10-22. Ductile Iron: Effect of-Notches on a 65,800-psiTensile-Strength Grade
I
36
<,
32
'iij
0.
t
0.750"
<,
2.75"R
- 250
0}17" 10.4'72"
..
225
200
0"",
~ ~()OOO
Unnotched
-"'0-
28
o
o
o
eo
n,
0
'\.
'\.0"
~
20
I'"
W
(J)
~
Ul
-i
0
o
175
"1
0.417"
0.700"
V-Notched
150
.....
0.25 mm Root Had.
125
1--0-000--
16
10'
10'
10'
100
10'
Number of Cycles
The unnotched and notched fatigue properties of an annealed

ductile iron with a tensile strength of 65,800 psi (454 MPa). The
endurance ratio is 0.41 and the notch sensitivity ratio is 1.67.
Source: Iron Castings Handbook. Charles F. Walton, Ed.. Iron Castings Society, Inc. 1981, p 341
10-23. Ductile Iron: Fatigue Crack Growth Rate

Compared With That of Steel
Stress Intensity Factor Range. t.K in MPa
v'rri
20
Q)
Q)
<3
<3
>-
>-
<,
:? 10-'
E
E
o
c
c
Band for Wrought

Ferrite-Pearlite
Steels
Z
"0
<,
'"
<Ii
m
a:
10-'
z
~
eo
"0
"0
'"
a:
s:
s:
o
(910-'
3300 Ib (1500 kg)

o 1650 Ib (750 kg)
.. 2700 Ib (1225 kg)
2200 Ib (1000 kg)
.:.!
o
~
U
<.9
.:.!
10-' '--_...J......_ _...L-_L-...L-..I...-I....L...I..J....L-_ _..I...-_L-.L.....I

10
30 40
60 80100
Stress Intensity Factor Range. t.K in ksi yinch
Fatigue crack growth rate of annealed ferritic ductile iron,

compared with that of ferritic-pearlitic steels.
Source: Iron Castings Handbook. Charles F. Walton. Ed.. Iron Castings Society. Inc.. 1981. p 349
267
268
10-24. Malleable Iron: S-N Curve Comparisons of Four Grades

60
r--------...,---------,
400
350
50
.iii
0.
<tI
c...
0
0
0
:2:
-0
300 ~
-0
--l
<tI
0
--l
40
260
30
L-
10'
-'-
10-
10'
Number of Cycles
The effect of cast surfaces on four grades of malleable iron

was also studied in high-stress, low-cycle fatigue. The results
with a 95% confidence limit are presented in this S-N diagram.
Unmachined and notched surfaces do reduce the fatigue
strength. The reduction factor is as low as 1.2 for the lowerstrength irons to over 2.0 for the higher-strength irons. Inducing
compressive stresses into the surface by rolling, coining, or shot
peening can increase the fatigue life of a component significantly. Design with adequate sections that are well blended to
reduce stress concentrations is most effective in reducing the
possibility of a fatigue failure.
Source: Iron Castings Handbook, Charles F. Walton, Ed., Iron Castings Society, Inc., 1981,p 311
10-25. Pearlitic Malleable Iron: Effect of Surface Conditions

on S-N Curves
70
'00 60
400
C-
o
0
0
<0
a..
~
60
300
enIII
40
Vi
30
en
III
Vi
200
20
10'
10'
10'
10'
Number of Cycles
The influence of as-cast surfaces, smooth machined surfaces, and machined notches on the fatigue behavior of pearlitic malleable irons. Iron 1
is grade 60003 and Iron 2 is grade 80002.
Surface finish has an important influence on fatigue properties, as shown above. Samples of malleable grades 60003 and
80002 were tested in fatigue with "as-cast" and machined surfaces. Samples of the 60003 grade were also included with a machined surface containing a sixty-degree notch that was 0.050 in.
(1.25 mm) deep. The resulting data are shown in this diagram.
269
270
10-26. Pearlitic Malleable Iron: Effect of Nitriding
1 PIECE
t----50.0
9 PIECP.S
ATHOSPIlr.RE
NITRIDED
COHPRESSIOH 11,160 POUND~

TENSION
7.600 POUNDS
1,01.--------------------------------II) 5
HUMBER OF CYCLES
Effect of gaseous atmosphere nitriding on fatigue characteristics of pearlitic malleable

iron, tested by tension-compression.
Samples were austenitized, oil quenched, and tempered to 241-269 HB prior to

nitriding or testing without nitriding. This chart indicates an increase in fatigue life
of 750,000 to 2,700,000 cycles attained by nitriding.
Source: J. A. Riopelle, "Short Cycle Atmosphere Nitriding," in Source Book on Nitriding, American Society for Metals. Metals
Park OH, 1977, P 287
10-27. Ferritic Malleable Iron: Effect of Notch Radius and Depth

Depth of notch, in.
250
002
004
006
008
~ tNotch
rad.ius
o 0.13 mm or 0.005 in.
c,
<,
:;;
-5
0>
150
.~
100
~
. r--
u.
50
30
0.75 mm or 0.030 in.
...............
t;
'"
5,
0.25 mm or 0.010 in.
200
--
E:
20
'"
5,
.~
r--- r-s-
u.
10
o
0.5
1.0
1.5
2.0
Depth of notch, mm
Effect of notch radius and notch depth on fatigue strength of ferritic

malleable iron.
Fatigue strength of unnotched ferritic malleable iron is approximately 50% of the tensile strength, or from 170 to 205 MPa (25 to 30
ksi). The graph above summarizes the effects of notches on fatigue
strength. As a rule, notch radius has little effect on fatigue strength, but
fatigue strength decreases as notch depth increases.
Source: Metals Handbook. 9th Edition. Volume I, Properties and Selection: Irons and Steels, American Society for Metals, Metals
Park OH, 1978, P 65
271
272
11-1. A286: Effect of Environment
A286
w
C)
z
:
z
c
VACUUM
o AIR 593C
VACUUM 593C
AVACUUM 20C
0.01
a:
en
~ 0.001
en
-J
Q.
Plastic-strain range versus fatigue life for A286 ferrous alloy in air and in vacuum af
593C (1095 OF). Numbers adjacent to test points indicate frequency in cycles per
minute. Note absence of frequency effects in vacuum.
Coffin has suggested that for a number of materials, virtually aU of the degradation in fatigue life at elevated temperatures can be attributed to environmental
interactions. He noted that frequency effects in the low-cycle-fatigue law could be
eliminated for a large number of metals and alloys by testing in vacuum (note
above). Additionally, it was noted that tests performed in vacuum showed transgranular crack nucleation and propagation versus intergranular nucleation and
propagation in air at elevated temperatures. These results are not unambiguous,
since Koburger has shown a frequency effect in high-cycle fatigue for directionally
solidified eutectic alloys when tested in air and in vacuum, particularly at elevated
temperatures. The primary difference in these results may be related to the lack of
intergranular cracking in eutectic alloys.
11-2. A286: Effect of Frequency on Life at 593C (1095 OF)
/.--------- - ------
~---
l!>;:;::--::
/' --1>-;;:::::::0 /' /'
:0
-:
".0
,/
)
/
A 286-593C
Kr=3.0 .MI. =60ksi
2
--:;....--:;;.--
--
v STANDARD HT-AIR
oHT#1
-AIR
HT#I
-VACUUM
0 HT#2
-AIR
I> HT #3
-AIR
0 DS HT # 3
-AIR
~DS-STD HT
- AIR
Effect of frequency on life of notched fatigue bars of A286 at 593C (1095 OF)
in air and vacuum. As indicated, decreasing frequency has a degrading effect on
fatigue life ofsamples tested in air, with little or no effect on samples tested in a
vacuum.
273
274
11-3. A286: Fatigue Crack Growth Rates at Room and Elevated

Temperatures
AK, ksi In.1I2
10
20
30
40
60
SO 100
A' /
10-3
53SC
(1000.F)~'
427C (SOOF) ,
,/
(i /
I"
10-4
10- 4
01
/ "'-24C (75F)
10-6
.5
2'
~
"tl
316C (600F)
I, /
ill
10-6
,
I
10-6
A2S6
I
10
20
30
40
60
SO
100
Stress-intensity factor range, AK, MPa ' m1/2
Fatigue crack growth rates for specimens of A286 stainless steel at room temperature and elevated temperatures
for tests in air at 3 Hz (RT) and 0.67 Hz (elevated temperatures), anR ratio ofO.OS, and at L-T, T-L, R-L, and R-C
orientations.
The austenitic precipitation-hardening stainless steel A286

(heat-resistant alloy) is the main representative in this category.
It contains titanium and small amounts of vanadium and aluminum, which precipitate as intermetallic compounds such as Ni,
(AI, Ti) and Ni 4Mo(Fe, Cr) Ti on aging. Various mill forms of
the alloy are usually supplied in the annealed conditionCondition A (980C, or 1800 OF, for one hour followed by
quenching in oil or water). Precipitation hardening occurs on
aging in the range from 700 to 760C (1300 to 1400 OF)for 16
hours. Other combinations of heat treatments may be used depending on the application. One variation is to re-solution treat
at 900C (1650 OF)for two hours, quench in oil or water, and
age at 700C (1300 OF) for 16 hours. This variation results in
improved room temperature properties but less desirable stressrupture properties.
11-4. Astroloy: S-N Curves for Powder vs Conventional Forgings

100
Conventional Forgings
'"I
2
Powder Forging
80
'
\I)
\I)
60
...
<J<:
'"
><J<:
...
40
s
<II
20
0
104
10 5
10 6
CYCLES
S-N curves for conventional and powder forgings of Astroloy (notched versus
smooth),
Testing was performed using standard methods at 705C (1300 OF) and a
combination of steady and vibratory stresses for which comparative data
were available. Cycles to first indication (crack) were comparable to conventional material. Crack propagation as judged by the number of additional
cycles from first indication to failure was slower than conventional material,
as shown above.
Source: M. M. Allen, R. L. Athey and J. B. Moore, "Application of Powder Metallurgy to Superalloy Forgings," in Source Book on
Powder Metallurgy, Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979, P 97
275
276
11-5. Astroloy: Powder vs Conventional Forgings Tested at 705C

(1300 OF)
Steady Stress = Vibratory Stress = 40,000 psi
x =
Crack
o = Failure
Powder
Forging
Conventional
Forging
10 4
CYCLES
Astroloy tested in high-cycle fatigue at 705 C (1300 OF). Vibratory

stress levels were selected to facilitate a direct comparison between conventional and powder forgings.
Source: M. M. Allen, R. L. Athey and J. B. Moore, "Application of Powder Metallurgy to Superalloy Forgings," in Source Book on
Powder Metallurgy, Samuel Bradbury. Ed., American Society for Metals, Metals Park OH, 1979,p 97
277
11-6. FSX-430: Effect of Grain Size on Cycles to Cracking

lSI 4)0
large grains
0.4
,
~
,
if
'"Z
~-LG
I I I
HI
MIN
'\
II:
t---O-I- IG
fine grain a
." 0.)
0.01
10.1~",1
HZ
AVG MAl
..
II:
~
~cb
0.01
(O.~'I
....&
......
~
~,
.,;
~
0
c
II:
E
E
II:
>
0.05
10.7621
...
c
...
II:
!:
0.0"'
11.0161
"-
...
II:
0.2
40
.0
IZO
160
ZOO
240
zeo
no
CYCLES TO CRACK
SoN curves for alloy FSX-430, showing effect of grain size on cycles to cracking,
Source: Eric Bachelet and Gerard Lesouit, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed.. American Society for Metals, Metals Park OH, 1984, P 336
278
11-7. FSX-430: Effect of Grain Size on Fatigue Crack Propagation

Rate
,",,,,/CYCLE I
O.OJ
GROWTH RATE IN llNlAl! flAIlGE

ODZ
0.01
0.04
J--O-1-fG
0.4
~-LG
CRACK AVG CflACK
0.01
t-{j----I
.
'"
~
z
:ca:
...
a:
<
0.3
....
0)1
<
...%
..
!:
c
'"
C
0.03 II:
(O.TUI~
e,
II:
)~
<
Q,
,,
It'
IE
...i
O.O!
(o.~oeio
(O.Z:>41 E
E
0.04
(1.0151
large grlolinB
0.2
I
fine grains
0.5
1.0
1.5
GROWlH RAlE IN LINEAR flANGl (IO"n.lCTCLEI
2.0
Fatigue crack propagation rate-effect of grain size on fatigue characteristics of FSX-430.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of SuperalIoys," in Superalloys: Source Book, Matthew J. Donachie, Jr.. Ed.. American Society for Metals. Metals Park OH, 1984, P 337
11-8. HS-31: Effect of Testing Temperature

70
60
.u;
a.
50
..........
o
o
Q 40
30
20
1100
Aged 50 hr at 1350 F
""""-<
Fatigue strength
100 million cycles
"'- r---.....
(["---...:
--
:--
1400
1200
1300
Testing temperature, F
1500
Effect of testing temperature on fatigue strength of HS-31 casting alloy, after aging at 730C (1350 OF), for 100 million cycles.
Source: ASM Committee on Heat-Resistant Castings, "Heat-Resistant Alloy Castings,"in Source Book on Materials for ElevatedTemperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH, 1979, p 237
279
280
11-9. IN 738 lC Casting Alloy: Standard vs HIP'd Material
y-"STANDARD CONDITlON" MATERIAL
220
<,
CII
200
E
E
:z;
180
Po.
+1
Po.
......,
Ul
Ul
160
11M>
'-.<,
<,
<,
MATERIAL
<,
-"<.
"
...... ......
E-i
Ul
<,
..............
120
tI$..............
107
CYCLES TO FAILURE
SoN curves for casting alloy IN 738 LC. High-cycle fatigue properties ofnimocast alloy IN 738 LC tested at
850C (1560 oF).
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Super alloys," in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, P 340
11-10. IN 738 LC: Effect of Grain Size on Cycles to Failure
240
/
220
E
E
2:
200
a..
"
STANDARD CDNDITIDN
-,
......
<,
.......
FINE GRAINS
.......
180
a..
111
111
W
0:
t-
111
160
......
.......
......
.......
140
MATERIAL
......
'
....... ~
120
e
10
10
10
CYCLES TO FAILURE
SoN curves for alloy IN 738 LC. High-cycle fatigue properties of extra-fine-grain and conventional material
tested at 850C (1560 OF).
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book. Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, p 340
281
282
11-11. IN 738 LC: Effect of Grain Size on Cycles to Cracking

'N - 738
0.4
.......-o--i ~
I I
M'N. HZ N ,
fG
-LG
0.0'
o.ZS41
MAX."2
'i
.!i
...
a:
0.02
_ O.J
IO.!lOI1
...
:...
f
....
c
e::
,
~~)-~......--<O)-----~
<,
........
....
eo
.20
160
200
CYCLES TO CRACK
oJ
:tIO
...
...a:
2:
0.04
11.0161
fine grains
2.40
::l
c
a:
c
0.03 a:
10.JUI
large grains
0.2
.;
0
320
S-N curves for alloy IN 738 Le, showing the effect of grain size on number of cycles to cracking.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, P 335
283
11-12. IN 738 LC: Effect of Grain Size on Crack Propagation Rate

c.ROWTH IlATE IN L1I1EAR 11&",[ l ..." u n l l
O.GI
,
0.01
,
O.OJ
I
0,04
,
. 4
0.01
O.~~I
t;
.:
O.GZ
(0.!>08)
.0.) f-
.
......
...
!
&
..~
Ci
QOJ
(0.7611
.,
f-
........
..'"
%
0.0-
1.01'1
...
.....
IN ' ) I
t-O-i :E-Ci -
0.2 f-
0.'
1.0
FG
LG
CRACI( I /IIIG CU~K I
1.5
1.0
GIIOWTH RAT( IN LINEAR RANG[ llli'in.lCYCLL,
SoN curves for alloy IN 738 Le, showing the effect of grain size on crack propagation rate.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OB, 1984, P 336
284
11-13. IN 738 LC: Fatigue Crack Growth Rate at 850C (1560 OF)
10- 5
do
dN (m/cycle)
10- 6
100Hz
Alloy I IN 738 LC
Temperature I 850C
.0
10- 10
Abo coarse - grained
10- 11
fine-grained
t'<J
0.1
Waveform: Sinusoidal
10
20
30
40
-3/2
~K(MNm
)
50
60
Fatigue crack growth rate at 850C (1560 OF) in various grain sizes of alloy IN 738 LC.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book,' Matthew J.
achie, Jr., Ed., American Society for Metals, Metals Park OH, 1984,p 341
000-
11-14. Inconel 550: Axial Tensile Fatigue Properties in

Air and Vacuum at 1090 K
50
40
30
....
C>
(a)
20
'iii
0-
vi
VI
w
a:
10
50
t-
VACUUM
VI
Z
<Cl:
AIR
40
::E
(b)
30
20
10
105
10
107
LIFE. cycles
I I II lid
I I "" II
100
1000
10000
LIFE. h
Axial tensile fatigue properties oflnconel550 at 1090 K in

air and vacuum. (a) Ratio of cyclicto mean stress= 0.125.
(b) Ratio of cyclic to mean stress = 0.667. Testing frequency = 33 Hz.
In reversed bend tests on lead at 500 cycles/min, Snowden

demonstrated a difference oftwo orders of magnitude in fatigue
life between vacuum, air, and pure oxygen. At all strain levels
vacuum endurances exceeded those in air, which exceeded those
in oxygen. Intermittent stress-free exposure to air had no effect
on the lifetime in vacuum. At high temperature (l090 K) vacuum also improved endurance, relative to air, of the Co-base
alloy S-816 and the Ni-base alloy Inconel550, although the effect was much smaller than that seen in lead. Endurances for the
nickel-base alloy converged at low stresses, indicating a possible
strengthening effect of air, as shown above.
Source: R. H. Cook and R. P. Skelton. "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
Park OH, 1979, p 81
285
286
11-15. Inconel 625: Effect of Temperature on Cycles to Failure

600
500 427C
Joo
80
of)
,~ e:--
400
<0
a.
:;;
Ii 300
200
'-
<,
...
---
29C (85 of)

538C (1000 F)_ 60
649C (1200 of)
I-
NOICh~~
specimens
(Kt~3.3)
760C ('1400 of)
29 Oc (85 of)
871C (1600 of) _
20
100
Cycles 10 failure
S-N curves for hot rolled solution treated Inconel625 bar

15.9 mm (0.625 in.) in diameter at various temperatures.
Average grain size was 0.10 mm (0.004 ln.),
Source: Metals Handbook, 9th Edition. Volume 3. Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
11-16. Inconel 706: Effect of Temperature on Fatigue Crack

Growth Rate
Stress-intensity factor range, aK, ksi in.1/2
10
20
30 40506080 100
200 300
10- 4
-196C (-320F)
Ql
10-3
Ql
>
u
--EE
z
--
"C
"C
III
III
"C
10- 5
l!l
III
..
Ql'
s:
i
e
i
e
-269C (--452F)
Cl
"C
III
s:
--z,E
-....
Cl
10- 4
I!
u
I!
U
Ql
Ql
::J
Cl
::J
III
'':;
Cl
'':;
III
II..
II..
10-6
20
30 40 50 60 80 100
200
300
Stress-intensity factor range, aK, MPa . m1/2

Fatigue crack growth rates of Inconel 706 forged billet (vacuum induction
melted/vacuum are remelted) at an R ratio of 0.1 and a frequency of 10 Hz.
Heat treatment: 980C (1800 OF) I h, AC; double aged 730C (1350 OF) 8 h,
FC to 620C (1150 OF), hold 8 h, AC.
Results of FCP tests at room temperature and at temperatures as low as

-269C (-452 OF) for Inconel 706 are shown above, At equivalent 11K
values, the fatigue crack growth rates for this alloy are slightly lower at subzero temperatures than at room temperature,
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties ofSuperalloys,"in Application of Fracture Mechanics for
Society for Metals, Metals Park OH, 1982, p 297
287
288
11-17. Inconel"713C": Effect of Elevated Temperatures on

Fatigue Characteristics
"in
50
c.
g 4 a I-+-+-f-Ht--~H-Yt---+--+
o
:i 3 a 1--+-+++t--~,*h-r--+--P't"+l:>---1--+t-t-l
Q)
"-
iii
Ol
2 a I--+--+--r-f-..---r---t-+-+--i---i--t-++t---+--t-+-+i
~
o
E 10
.2?
<i
NOTE: Higher fatigue strength at

1500 F than at 1200 F is consistent
with tensile strength relations in
graph shown above left.
OL-..I...-...L....J.....LJ...---l...~
.............1...--'--...L....J.....LJ...---L.----L~
0.1
I
10
100
SoN curves for Inconel "713C." Tests were performed at

two different elevated temperatures as shown.
Source: ASM Committee on Heat-Resistant Castings, "Heat-Resistant Alloy Castings," in Source Book on Materials for ElevatedTemperature Applications. Elihu F. Bradley, Ed.. American Society for Metals. Metals Park OH. 1979, P 235
289
11-18. Inconel "713C" and As-Cast HS-31: Comparison of Two

Alloys for Number of Cycles in Thermal Fatigue to Initiate Cracks
Thermal fatigue
Cycles to first
crack
Material
HS-31
Inconel
ri
l-rnin cycles,
100 to 1700F
Material
avg1
avg
-1&
1713C"
,r
I
Thermal fatigue
Cycles to develop
.
8I-m,
erne k
HS-31
Inconel
Thousands of cycles
IIIII
1713C"
3 tests
each motertol
Thousands of cycles
Thermal fatigue properties of HS-31 compared with those of Inconel "713C." Left: Number of cycles required to initiate cracks. Right: Number of cycles required to develop VB-in. crack.
Source: ASM Committee on Heat-Resistant Castings. "Heat-Resistant Alloy Castings." in Source Book on Materials for ElevatedTemperature Applications. Elihu F. Bradley, Ed.. American Society for Metals. Metals Park OH. 1979, P 235
290
11-19. Inconel718: Effect of Frequency on Fatigue Crack

Propagation Rate
I
5 X 10-3
f-
o
2
o 5 X 10- Hz }
05 X 10-' H,
o",~/ /
5 X 10-4 -
1>
/S
/ili.
o 00
&
/;".
/&
o
5 X
10-5 I-
tL {--
a~a / I.
I
/
&
~/
20 Hz
2Hz
0.5 Hz
&&/
&
//
& ......---/
5 X 10-6
/
&1
1..----I._ _--1.._--'_...L...--1..--'---'--'-...L........L.---l
10
15
20
30
40
50 60
Stress-intensity factor range, 61<, MPa . m1/2

Variation ofFCP rate (da/dN) with stress-intensity factor range
(LlK) and frequency at 550C (1025 OF) (sinusoidal load) for
specimens oflnconel 718.
The effect of frequency at 550C (1025 OF) was studied using

a sine wave; the results are shown above. Below 0.5 Hz, the FCP
rate was more rapid, and the crack surfaces showed an increased
amount of intergranular fracture with decreasing frequency,
with the crack path following the boundaries of the largest
grains. One may be inclined to attribute the increase in FCP to
either creep or environment, but this may not be the case, because different modes of deformation may have occurred at different strain rates.
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys, "in Application of Fracture Mechanics for
Society for Metals, Metals Park OH, 1982, P 294
11-20. Inconel718: Relationship of Fatigue Crack Propagation Rate

With Stress Intensity
25C (77F)
550C (1025F)
20 Hz
5 X 10-4

..
5 X 10-5
.. ,
:-
5 X 10-6
._~
: Twins
10- 6
......--L
10
No twins
....L._--'-_.J...-............L.....L.........L.-............
15
20
30
40
50 60
Stress-intensity factor range, Li K, MPa . m 1/2
Dependence of FCP rate (dol dN) on stress-intensity factor range (LiK) and temperature at 20 Hz (sinusoidal
wave shape signal) for specimens of Inconel 718.
An important effect is the hydrostatic state of stress in the tip

region. This idea has been considered for Inconel 718. The FCP
response is shown above for 20 Hz, where the effect of temperature is to increase the FCP rate, especially at LiK levels.
291
292
11-21. Inconel 718: Relationship of Fatigue Crack Growth Rate

With Load/Time Waveforms
a K. ksi in. '/2

10
20
30
40
60
/'v (2)
5 X 10-2 Hz
2 Hz
_ rxr
(1)
/'v
(3)
(2)
rf
Sinusoidal ~
~
2
5 X 10- Hz
5 X
10- 4
I
(1)
'Jr
~
- / (3)
-;0
2X
I. _ -Q
r:t
I
I... /
rI
,6
.5
"tl
" tl
t6
..
Sinusoidal
2 Hz
5 X 10- 6
4l
10-6
2 X 10- 6
10
15
20
30
40
50 60
aK. MPa m,/2
Load/time waveforms and FCP rates for specimens of Inconel

718. Top: Various forms of cyclic stress fluctuations used at 550
C (1025OF)at a frequency of5X 10- 2 Hz. Bottom: FCP rates at
550C under sinusoidal, triangular and square loads.
To separate out the possible effects of creep or environment from deformation mode, the authors used triangular and square wave shapes, as shown in
the top graph. The data obtained using the triangular wave at 2 Hz were the
same as the data obtained in other tests using the sine wave at the same
frequency which resulted in the lowest FCP rate. The effect of loading at the
same rate but imposing a lO-second hold time at maximum load was to increase the FCP rate only slightly, as shown in the lower graph.
11-22. Inconel 718: Fatigue Crack Growth Rate in Air vs Helium

~K, ksi
iny2
10
20 3040 6080
5 X 10-1 1'"T'"--"""--T"-"""-T""T""1'""'I""l
Air
= He
10- 3
.,
--EE
10- 2
z'
--.,'..
"tl
Gl
>
u
......
.5
"tl
.....
10- 4
.I:
i0
..
--.
"tl
0>
...
.,
..
Z'
"tl
u
u
10-3
::;,
0>
.;::;
u..
10-5
10-4 ......._ _L - - J ' - -........I-l'-'-..........

10
20
40 60 80100
~K,
MPa . m1/2
Fatigue crack growth rate data for Inconel 718 in

air and in helium. Frequency, 0.1 Hz. Temperature, 650C (1200 OF). Here it is evident that crack
growth rate at constant temperature is lower in inert gas.
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys,"in Application of Fracture Mechanics for
293
294
11-23. Inconel 718: Effect of Environment on Fatigue Crack

Growth Rate
~K, ksi in. 1/ 2
10
20
30 40 60 80
5 X 10-1
10- 3
.,
~
E
E
--
10-2
.,
U
~
~
III
"'C
....,'
10- 4
--.5
z
-"'C
.t:
III
e
"'u..."
"'C
CI
III
U
10-3
.,
:::l
CI
'';::;
III
u..
10- 5
10- 4
He
+ 0.5%
0=
He
+ 5% S02
H 2S
L-_----I_~---L___L.....I.....L.J...L.J
10
20
~K,
40
60 80100
MPa m 112

helium + 0.5% hydrogen sulfide and helium + 5%
sulfur dioxide. Frequency, 0.1 Hz. Temperature,
650C (1200 OF).
From the data above it becomes obvious that fatigue crack

growth rate increases greatly in aggressive environments compared with exposure to helium alone.
11-24. Inconel 718: Fatigue Crack Growth Rate in Air Plus

5% Sulfur Dioxide
AK. ksi in. ' / 2
10
20
30 40 60 80
5 X 10- 1
Air
+ 5% S02
10- 3
II>
Q
>
u
--
10- 2
E
E
--'"
"C
II>
>
"C
i
...'"
10- 4
~
.S
'"
"C
."
~
eu
10- 3
II>
:s
."
.;;
'"
II..
10-4
'--_ _.J...._-'---'--'-...L..J.....L..L..J
10
20
40
AK, MPa
60 80100
m' /2

air + 5% sulfur dioxide.
(The effect of air plus 5% S02 was similar to the effect of air alone.) It was observed that in
the helium atmosphere, which was used to establish a baseline, cracking was generally transgranular with well-defined striations. In the air, oxygen-bearing and sulfur-bearing environments, the crack path changed from transgranular to intergranular, indicating that an important
effect of the environment was to degrade the boundary strength by mechanisms that were not
clearly defined. It was suggested that oxygen diffusion along grain boundaries and localized
oxidation may have occurred. Another very important observation was that the effect of a given
environment on FCP could not be predicted on the basis of unstressed exposure tests. The attack
on the surfaces of unstressed specimens in aggressive S02 environments was minimal, but the
S02 environments caused substantial increases in FCP.
295
296
11-25. Inconel 718: Fatigue Crack Growth Rate in Air

at Room Temperature
Spec. 1290" /
(CHT)
'f
~
E
III
Spec. 158 & 803'"

(CHT)
/
v'
10-3
(!.
.. ,I:!> 0
/ ,
.. I
Z'
~
"~.'"
Inconel718
Tested in air at
24C (75F)
500 < f < 600 cpm,
R = 0.05
Conventional
heat treatment
I:!> Spec. 1290, heat I

0 Spec. 158 } heat II
V Spec. 803
0 Heat III
20
40
60
80
Modified
heat treatment
I:!> Spec. 1283, heat I

0 Spec. 253, heat II
I
I
(b)
20
40
60
80

Fatigue crack growth rate behavior ofInconel718 tested in air at 24 C (75OF).CHT=
conventional heat treatment. All testing was done at R = 0.05 and at a frequency of
0.67 Hz.
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys,..in Application of Fracture Mechanics for
Society for Metals, Metals Park OH. 1982, p 276

at 316 C (600 OF)
Spec. 210 (CHT)
-:
y "IJ./
.. ,
Spec. 1294 (CHTl
.I IJ./
/ ,~
Inconel718
Tested in air at
316C (600F)
f = 40 cpm, R = 0.05
t:J.
..
..I1J.4'
,0
/ IJ.I 0
.. IJ. .,
/
/
./
Conventional
heat treatment
IJ.
IJ.
Spec. 1294, heat I
o Spec. 210, heat

o Heat III
(a)
Modified
heat treatment
10- 6 L..-_ _--I._--'_....L--L.-L-'-...L...J-'40

60 80
20
Spec. 1282, heat I
o Spec. 254, heat
II
(b)
II
.L..-_....L--''--J.-........--L....J-I
20
40
60 80
Stress-intensity factor range, Ll. K, MPa . m 1/2

Fatigue crack growth rate behavior of Inconel 718 tested in air at 316C (600 OF).
CHT = conventional heat treatment. All testing was done at R = 0.05, and at a frequency of 0.67 Hz.
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Superalloys," in Application of Fracture Mechanics for
297
298

at 427 C (800 OF)
10-1
...----r--,-__r-..___..___.....-r::I:~--__r--..___.....,..____.____.___r.... T.:I
Inconel718
Tested in air at
427C (800Fl
f = 40 cpm, R = 0.05
Conventional
heat treatment
10- 2
D.
Modified
heat treatment
Spec. 1291, heat I

Spec. 162, heat II
D.
Spec. 1286, heat I

spec. 255, heat II
:1
Spec. 162 (CHT1,,/1
/'1
/.~
10-3
Spec. 1291",: ~
(CHT)
lID.
0
0
.. I
II
.. I
II
10-4
.. I
.II
//
10-5 L...-
L . - _ - ' - - - ' - _ . L - . L -...................
20
40
60
80
<9
----'_ _.L---L----'----''---L...............
20
40
60 80
Stress-intensity factor range, AK, MPa m1/2

CHT = conventional heat treatment. All testing was done at R = 0.05 and at a frequency of 0.67 Hz.
Source: Stephen D. Antolovich and J. E. Campbell. "Fracture Properties ofSuperalloys, "in Application of Fracture Mechanics for
Society for Metals, Metals Park OH. 1982, p 278

at 538 C (1000 OF)
10- 1
1:""""----,--...-~-,.._r_1r_T"T"']r:__--__r--.___r-,.._r_1r_T"""1"":I
Inconel 718
Tested in air at
538C (1000F)
f = 40 cpm. R = 0.05
10- 2
10- 3
10- 4
10- 5
Conventional
heat treatment
Modified
heat treatment
Spec. 1288, heat I

Spec. 165, heat II
heat III
Spec. 1284. heat I

Spec. 251} heat II
Spec. 250
'--_ _----'C--_.1.--L_.l.-.L.....JL..-L....L.J'--_ _----'_ _-'------'_.l.-.L:--JL.....L~
20
40
60
80
Stress-intensity factor range, LiK. MPa . m 1/2

Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of'Superalloys, "in Application of Fracture Mechanics for
299
300

at 649C (1200 OF)
1
10- ~---r---r-'---'-"'T1-r:r-----r----,--,--r""'T"r-r""']
Inconel 718
Tested in air at
649C (1200F)
f = 40 cprn, R = 0.05
V'
Spec. 1289
(CHT)
10- 3
./
./
!SJ
/
10- 4
Conventional
heat treatment
b.
Spec. 1289, heat I
o Spec. 156, heat
o
20
Modified
heat treatment
II
Spec. 1281, heat I

Spec. 252, heat II
heat III
40
Stress-intensity factor range, l1K, MPa . m 1/2
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties ofSuperalloys."inApplication of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell. William W. Gerberich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH. 1982, P 280
11-30. Inconel 718: Fatigue Crack Growth Rates at

Cryogenic Temperatures
22 C (72 F)
}
-78 C (-108 F)
2.54 cm
'V -196 C (-320 F)
thickness
o -269 C (--452 F)
l>
Q)
10- 3
22 C (72 F) 0.51 cm
thickness
EE
z
<tI
"0
~f
~
.r:
e
Cl
.:.!
CJ
10- 4
Q)
:I
Cl
.;:;
<tI
u..
10
50
100
Stress-intensity factor range, AI<, MPa m 1/2
Fatigue crack growth rates of Inconel 718 forged bar at an R

ratio of 0.1 and a frequency of 20 Hz. Heat treatment: 980C
(1800 OF) '% h, AC; double aged 720C (1325 OF) 8 h, FC to
620C (1150 OF), hold 10 h, AC. At the constant frequency the
effect of higher temperature is to increase the FCP rate.
Source: Stephen D. Antolovichand J. E. Campbell, "Fracture PropertiesofSuperalloys,"in Application of Fracture Mechanics for
301
302
11-31. Inconel718 and X-750: Fatigue Crack Growth Rates

at Cryogenic Temperatures
Inconel 718 (Ref 8.44)

~ 22 to -269C (72 to -452F)
Inconel 718 (Ref 8.49)
22C (72F)
Inconel X-750 (Ref 8.48)

~ 27 to -269C
(80 to -452F)
10
50
100
Fatigue crack growth rates for Inconel718 and Inconel X-7S0in

the subzero temperature range.
A comparison of FCP values from room temperature to

-269C (-452 OF)for Incone1718 and Inconel X-750 is shown
in the above chart, along with room temperature FCP data for
Inconel 718 from Shahinian et al. The FCP data for these two
alloys overlap in the t!K range shown. Under some conditions,
the FCP rate for Inconel 706 is slightly less than those for Inconel 718 and Inconel X-750 at corresponding temperatures and
t!Klevels. However, results of FCP tests depend on both melting practice and thermomechanical processing.
Society for Metals, Metals Park OB, 1982,P 300
11-32. Inconel X-750: Effect of Temperature on Fatigue Crack

Growth Rates
22C (72F)
-196C (-320F)
CIl
!E
10- 3
o -269C (-452F)
E
Z
~
"tl
:!l'
E
ie
en
CJ
10-4
CIl
5,
.~
'"
II..
100
Stress-intensity factor range, llK, MPa . m 1/2
Fatigue crack growth rates of Inconel X-750 at an
R ratio of 0.1 and at frequencies of 20 to 28 Hz.
Heat treatment: solution treated and double aged.
Within this frequency range, the effect of higher
temperature is to increase the FCP rate.
303
304
11-33. Jethete M 152: Interrelationship of Tempering Treatment,

Alloy Class, and Testing Temperature With Fatigue Characteristics
ksl
100r------------------,
...:::>~
'"
~
0.6
)(--------x
0.4
=ksi
100
80
0-_0______
50
i!:
...~
0"
...
x tempered at 510 F(300 Cllor one hour

to alensile slren&lh 0labout205 ksl
0_
o temperedal1200 Fl650 ClIorone hour

toa tensilestreneth abouI150ksl
40
20
LONGITUOINAL SPECIMENS
ROTATING8EAM TESTS FOR \0' CYCLES
0'
200
400
600
I
50
60
0:
Ii;
1000 F
800
1
450
O!----..!:--....,-k".--""*';---=----.od;;--__=!
1200 I
I
200
300
400
600
TEST TEMPERATURE
Left: Interrelationship of prior tempering treatment and testing temperature with limiting fatigue stress, and
with fatigue ratio for Jethete M152. Right: Influence of alloy class and testing temperature on fatigue
strength for the same alloy.
Source: J. Z. Briggs and T. D. Parker, "The Super 12%CrSteels," in Source Book on Materials for Elevated-Temperature Applications, Elihu F. Bradley, Ed.. American Society for Metals, Metals Park OH, 1979, P 123
11-34. Lapelloy: Interrelationship of Hardness and Strength With

ksi
9o,------------..----,
CLASS II (Lapelloy)
90,---------------,
2000 F C1 095 Cl sail quench,

marlemper 650 F(345 C) +
lemper 1150/1600 F(620/B70 C)
CANTILEVER ROTATINGBEAM
TESTS
mechanically polished,
BO
rms 2.5/4.0 mlcrelnehes
alhersurface Irealmenls,
o Including surface lolline.
rough e:rinding andlough

machining. rms5.5/40
microlnches
50
70
95% conlidencelimits
50
4~'='00,..----,-!;"....----:-!-::-----:*"--""IBO ksi
Left: Relation between surface hardness and mean fatigue limit for Lapelloy. Right: Relation between
tensile strength and mean fatigue limit for the same alloy.
Source: J. Z. Briggsand T. D. Parker, "The Super 12%Cr Steels," in Source Book on Materials for Elevaled-Temperature Applications, Elihu F. Bradley, Ed., American Society for Metals, Metals Park OH, 1979, p 123
305
306
11-35. Mar-M200: Effect of Atmosphere on Cycles to Failure
10
8
o
6
AIR
DRY AIR
WET AIR
VACUUM
PREOXIDIZED SPECIMEN. VACUUM TESTED
10
10
8
101
o
z
'"a:
V>
V>
a:
lV>
10
10
10 5
CYCLES TO FAILURE
S-N curves showing fatigue life at 10 Hz of single-crystal

low-carbon alloy Mar-M200 at 295-1200 K.
Convergence of air and vacuum data was noted for AISI 3I6
steel at 1090 K, and a crossover of the air and vacuum curves
occurred for nickel, where it was suggested that oxide in cracks
could prolong life in air at low stresses. Crossovers have also
been seen in a ferritic stainless steel and a Nil Cr alloy in the
range 875-1025 K, where tests in purified argon gave shorter
endurances than those in air, impure argon, or sulfur dioxide.
Also, in single crystals of the alloy Mar-M200, air endurances
were less than those in vacuum at room temperature whilst the
reverse was true at high temperature (above). A thin oxide film,
formed during testing, suppressed surface crack initiation, but
oxide formed during pre-exposure did not.
Source: R. H. Cook and R. P. Skelton, "Environment-Dependence of the Mechanical Properties of Metals at High Temperature,"
Park OH, 1979, P 81
11-36. Mar-M509: Correlation of Initial Crack Propagation and

Dendrite Arm Spacing
5.0
..
3,0 u>.
U
U
>.
u
4.5
N""-
E
E
'fg
'1'0
:il~ 4. 0
:
...
:il~
N""-
u
2.5 1;
or:
or:
iii
0\
:!.
0
iii0\
3.5
:!.
0
~
ij
......
;;;
;::
...
......
u
no
2.0
3,0
2.5 L-..
20
~c
.L-
40
Dendrite Arm Spacing
...r........
.....
60
1~1
Correlation between the initial crack propagation rate and the dendrite arm
spacing for Mar-MS09.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys,' in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, p 338
307
308
11-37. Mar-M509: Correlation Between Number of Cycles Required

to Initiate a Crack and Dendrite Arm Spacing
300l
r-----~---____r----
......---__r---_,
~
~
...
... 2000
;
c
-a......
- - - PresentStudy
u
'0
~
e
:::J
IlXXl
OL.-_.L.-_--L.
20
-'-
40
---'
-L..
60
---'
100
Dendrite Arm Spacing (1111)

Correlation between the number of cycles required to initiate a crack and the
dendritic arm spacing for cast alloy Mar-MS09.
Source: Eric Bachelet and Gerard Lesoult, "Quality or Castings or Superalloys," in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1984. P 337
11-38. MERL 76, P1M: Axial Low-Cycle Fatigue Life of As-HIP'd

Alloy at 540C (1000 OF)
100
90
00
80
0
'in
70
KT
0
=
1.0
m'
fA
60
50
40
30
Life, cycles
Axiallow-cycIe fatigue life ofas-HIP'd P/M alloy MERL 76 at 540C (1000

OF) at notch severities as indicated.
Source: J. H. Moil, V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications," in Powder MetallurgyApplications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, p 275
309
310
11-39. Nickel-Base Alloys: Effect of Solidification Conditions on

Cycles to Onset of Cracking
1000
F"t Solidification
(Condition f)
SlolI Solidlficotion
(Condillon S)
...
~...
......:>
Q
......... 100
..J
0.
...
Q
...'"
...
..J
s
~
...a:
......r
10
Q.
...
Q
...a:
CD
r
:z
:>
M21 71) 1I~71a IN!U

M21 713 IN738 '''939
lC
LC
lC
lC
"~_ _..L.----1_......L_..L.._..L-_---1L.-_-L..._"""'---1_......L_-L...
(;;:s;J
Creln
10 Ihl anul 01 crackIng
Talal eycln 10 !raw crack to 25 "''''
Bar chart showing effect of solidification conditions on cycles to onset of

cracking and total number of cycles required to grow cracks to 2.5 mm
in several nickel-base casting alloys.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superal1oys: Source Book, Matthew J, Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1984, P 339
11-40. Rene 95 (As-HIP): Cyclic Crack Growth Behavior Under

Continuous and Hold-Time Conditions
Stress-intensity factor range, L1K, ksi in. 1/ 2
6
10
20
40
60
80 100
10-2
10- 1
Stress-level
dependence
Ql
Ql
10-3
>u
-<;
E
E
z'
Z'
15-minute hold
time at maximum
tensile stress
-IV
..
"C
Ql'
-"C
IV
.....
"C
Ql'
IV
f!
10- 4
..c
Cl
>u
'<,
.~
10-2
"C
...0
10- 3
..c
~
2Cl
f!
f!
Ql
::3
Ql
Cl
::3
'';:;
Cl
'';:;
IV
10- 5
IV
U.
U.
10- 4
'--.1...--'-_---1._ _........." --_ _- ' -_ _...1-_"'-----'-1
10
20
40
60
80
10- 6
100
Stress-intensity factor range, L1K, MPa m 1/2

Cyclic crack growth behavior for as-HIP Rene 95 under both continuous and hold-time conditions at 650 C (1200 OF).
The effect of environment need not always lead to more rapid crack growth. It has been
proposed that oxidation products could form in the crack tip region and prevent crack
resharpening during the unloading portion of the cycle. If the stresses are sufficiently low,
the oxidation products in the crack tip region will not be cracked and, in some systems, an
elevation of the threshold might occur. Such effects would be pronounced at high temperatures and long hold times and have actually been observed in Rene 95, as shown in the
above chart. Once the stress intensity is high enough to crack the oxides, the rate of crack
growth would be expected to increase due to the severely degraded region in the crack tip
zone.
Source: Stephen D. Antolovich and J. E. Campbell. "Fracture Properties of Superalloys. "in Application of Fracture Mechanics for
Selection of Metallic Structural Materials. James E. Campbell. William W. Gerberich and John H. Underwood. Eds.. American
Society for Metals. Melals Park OH. 1982. P 284
311
312
11-41. Rene 95: Effect of Temperature on Fatigue Crack Growth Rate

Testing temperature, F
100
10_2 20
1000
..
2000
10-4
.5
z
-.r
....
.!
...
...
i'
10- 3
:i
l!
~
e
11'"
11'"
.."
l!
10-6
:::J
..~
:::J
'"
.~
...
'"
...'=
10- 4
o
Testing tempereture, C
Effect of temperature on fatigue crack growth rate at constant

for Rene 95.
bJ(
That the effect of environment can be large may be inferred from some low-cyclefatigue
studies of Rene 95 in which surface and subsurface cracking was observed at comparable
strain ranges and defect sizes. As expected, the life of the subsurface crack was much
greater than that of the surface crack, leading to the hypothesis of a strong environmental
effect. This possibility is considered in more detail in an analysis of FCP properties of
Rene 95. The FCP rate was plotted as a function of temperature for a given !:J.K range, as
shown in the above chart. It is noteworthy that there is a minimum in the FCP rate at all
!:J.K levels except 22 MPaM 1/2 (20 ksi-in.b"), where the data are at least suggestive of a
minimum. Because any environmental interaction is thermally activated, the crack
growth rate at a given !:J.K level and frequency may be written as:
da
dN = Aexp - Q(!:J.K)/ RT
where A is a constant and Q(!:J.K) is the apparent activation energy.
Source: Stephen D. Antolovich and J. E. Campbell, "Fracture Properties of Super alloys, "in Application of Fracture Mechanics for
11-42. 8-816: Effect of Notches on Cycles to Failure at 900C

(1650 OF)
40 ,----,r----,r--....,...-...,..--r---r---r--..---r---,...---. 280
ltJ
a..
~
vi
~
30
210
"-
"-
'Iii
'n. ...........
E
:J
.~
vi
~
20
A::;oo
0-_-0... K,
= 3.4
----__
Reversed stress fatigue
0
r-v
--00"--
106
Number of cycles
SoN diagram for 8-816 heat-resisting alloy tested at 900C
(1650 OF), notched (broken curve) versus unnotched (solid
curve).
Source: High-Temperature Fatigue, p 245
140
'Iii
E
:::J
E
'x
ltJ
~
313
314
11-43. Udimet 700: Fatigue Crack Growth Rates at 850C (1560 OF)
AK, ksi in. 1/ 2
10
40
20
60 80100
10- 3
10- 2
Q)
Q)
>
t.l
--EE
>
t.l
--.E
z'
"tl
'<,
'"
10- 4
"tl
...~
Q)'
.t:
01
-"
t.l
...e
Q)'
s:
'"
"tl
e
01
10-3
-"t.l
t.l
t.l
.'"
Q)
Q)
...
01
.01
u,
'"
u,
::J
::J
10- 5
Stage I
0
m = 16
KN
10- 4
Ib
c 582 1310
v 711 1600
01067 2400
l> 1244 2800
160 3600
20
10- 6
40
60 80 100
AK, MPa m 1/2

Crack growth rates in terms of stress-intensity factor range for
Udimet 700 at 850 0 C (1560 0 F). Crack growth rates for this alloy
are greatly accelerated by increases in temperature.
Source: Stephen D. Antolovich and J. E. Campbell. "Fracture Properties of Super alloys,"in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerberich and John H. Underwood, Eds.. American
11-44. U-700 and Mar-M200: Comparison of Fatigue Properties

I
- - COLUMNAR GRAINE\i
AND SINGlE CRYSTAl
-- ..
-- --.
M~R.:-,~OO
10-2
-c
1700 0 F
-c
a:
Z
a:
1'-.
l-
V)
.....
-(
I-
-. -- ..
I-
w- 2
oI-
1-..
C>
- - - CONVENTIONAllY CAST
MAR.M200
I I I
---WROUGHT POIYCRYSTAlliNE
UDiMEl 700
.~
I':::--
--_.~~
r--
-.
'--.
I-.. r--..,.--. .1---.

- r-
!--.
'.
1400 0 F
,
103
CYCLES TO FAILURE
Comparison offatigue properties at 760 0 C (1400 0 F) and 925 0 C
(1700 OF) for a typical wrought nickel-base alIoy (U-700) with
conventionalIy cast, directionally solidified and mono crystal ,
Mar-M200.
Source: Francis L. Versnyder and M. E. Shank, "The Development of Columnar Grain and Single Crystal High Temperature
Materials Through Directional Solidification," in Source Book on Materials for Elevated-Temperature Applications, Elihu F.
Bradley, Ed., American Society for Metals, Metals Park OH, 1979, p 358
315
316
11-45. Waspaloy: Stress-Response Curves
a. 1200
::iE
Ul
STRAIN CONTROLLED
1100
:::l
t-
:J
a.
::iE
V)
V)
Ul
0::
t-
WASPALOY
V)
10
100
N, CYCLES
1000
Stress-response curves for Waspaloy having nonshearable precipitates.
During aging of precipitation-hardenable alloys, the coherent

precipitates grow, and accommodation strains build up. At
some point the energy associated with the accommodation
strains exceeds that necessary to create a precipitate-matrix interface, and the precipitates become partly incoherent. This is
accompanied by a change in precipitate-dislocation interaction
from one of shearing to that of dislocation looping or bypassing
the precipitates. Since the reasons for strain localization have
been removed, deformation becomes more homogeneous. Local softening is thus prevented, and the cyclic-response curve
shows hardening to saturation, or to failure, as illustrated
above.
11-46. X-40: Effect of Grain Size and Temperature on Fatigue

Characteristics
~ %
}
-'---}
----
Z
....
<
at 750 DC
at 650 DC
~ 0.6
1Il
L!J
0.5
0.4
z< 0.3
0::.
Small
I-
...J
-e
0.2
-....
0.1
2
100
2
1000
5
10.000
CYCLES TO FAILURE
S -N curves for X-40 showing effects of grain size and temperature on fatigue characteristics of this alloy.
Source: Eric Bachelet and Gerard Lesoult, "Quality of Castings of Superalloys," in Superalloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH. 1984, p 335
317
318
11-47. Cast Heat-Resisting Alloys: Ranking for Resistance to

Thermal Fatigue
IlJ
a::
:J
='300
it
o
F-200
~150
...J
~IOO
<l
::!:
a::
ILl
~ 50
u, 40
ffi
30
z~
20
ILl
15
ILl
10
~
a::
10 15 20 30 40 50
100
200 300
CYCLES TO FAILURE OF INDIVIDUAL TESTS
The design of components that are subject to considerable temperature cycling must also
include consideration of thermal fatigue. This is particularly true ifthe temperature changes are
frequent or rapid, and nonuniform within or between casting sections. Fatigue is a condition in
which failure results from alternating load applications in shorter times, or at lower stresses,
than expected from constant-load properties. "Thermal fatigue" denotes the condition when the
stresses are primarily due to hindered expansion or contraction. Good design helps minimize the
external restraint to expansion and contraction. Rapid heating and cooling may, however, impose temperature gradients within the part causing the cooler elements of the component to
restrain the hotter elements. Finite-element computer analysis has shown that, for some industrial applications, these thermally induced stresses may exceed those resulting from the mechanicalloads.
An example of results from thermal fatigue data is presented above. This graph offers a ranking of many cast heat-resistant high-alloy grades relative to their resistance to thermal fatigue.
Such rankings are indicative of general alloy properties only because most thermal fatigue tests
are based on an arbitrary set of experimental conditions rather than on their fundamental material behavior. Nevertheless, such test results have been useful in considering alloy selection questions, and in identifying the superior thermal fatigue resistance of nickel predominating grades
and the good performance of some HH type compositions.
P 19-7
12-1. Corrosion-Fatigue Properties of Aluminum Alloys Compared

With Those of Other Alloys
500,.------typical
corrosion
-.,
fatigue strength
(sea water , N 108 cycles

ambient temperature
70
R -1 )
60
400
';:;'"""'
-.
z
duplex staness steels,

titanium alloys
50
(e.g. n -6A1-4V)
300
40
J:.
0.
c:
-F
200
.~
'iii
.><
L-J
iii
,.....,
100
30
.!j
nickel alloys
(e.g. aJloys 600 and fn
r.. .-/ .
.S!'
~
20
ferritic
_ stainless
martensilic_ \ steels
copper - nickel alloys

,carbon steels,
I bw alloy
aluminum
m
iii
.~
8
1)
As shown above the corrosion-fatigue strength of bare aluminum alloys is superior only to that of magnesium alloys. Careful
surface protection may bring the corrosion-fatigue strength up
into the range of bare stainless steels or copper-nickel alloys.
Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material," in Aluminum Transformation Technology and Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, R. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OR, 1980, P 617
319
320
12-2. Comparisons of Aluminum Alloys With Magnesium and Steel:

Tensile Strength vs Endurance limit
6.
250
()~ /
':>.o~'b
200
"3 /
6.
6./
r0-
ll..
~
E::::: 150
:.= (J)
.'=
0)
0)-
c i:)
moo 100
~O
"O~
Aged aluminium alloys

Non-heat treatable
aluminium alloys
Magnesium alloys
Steels
C )(
UJ~
50
O~---JL...-_----I.
100
x
0
6.
_ _---L_ _-'-_ _....I..-_ _.L......_---l
200
300
400
500
600
700
Tensile strength (MPa)

Fatigue ratios (endurance limit/tensile strength)
for aluminum alloys compared with those of magnesium and steel.
It is well known that, in contrast to steels, the increases that

have been achieved in the tensile strength of most nonferrous
alloys have not been accompanied by proportionate improvements in fatigue properties. This feature is illustrated in the
graph above, which shows relationships between fatigue endurance limit (S X 108 cycles) and tensile strength for different alloys. It should also be noted that the fatigue ratios are lowest for
age-hardened aluminum alloys and, as a general rule, the more
an alloy is dependent upon precipitation hardening for its total
strength, the lower this ratio becomes.
Source: I. J. Polmear, Light Alloys, Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981, P 39
12-3. Aluminum Alloys (General): Yield Strength vs Fatigue Strength
500
ultimate tensile strength

yield strength
fatigue strength in air , N - 5 X 101
1
latiguestrength in sea water I N _10
fatigue strength
8
in river water I N - 10
o
o
R - -1 , .. 60 Hertz I
stnooth specimens I
ambient temperature
4()()
70
60
+
50
';:;"'"'
--z
E
.--.
300
'iii
""
'--'
40
corrosion fatigue strength

01 aluminum alloys
s:
s:
e;,
e;,
c
30
l!! 200
OJ
l!!
OJ
000
0CF,air ,
N-5x1Q1
-Orr-----l:I
20
100
000
I
~ ~
~~
II
~ ~~
I
~I
f2,
<0
~
I
Cl)
iij
...
!2 <0, <0 <0

~
I
I
,
~
Ie ~ 'Ie
R
;::
R
~
An analogous conclusion can be drawn from a review of corrosion fatigue

tests with smooth aluminum alloy specimens as shown in 'the above graph.
Here aluminum alloys are listed in order of increasing yield strength. As the
yield strength goes up, so does the ultimate tensile strength, but the fatigue
strength in air soon reaches a limit which is roughly the same for alloys of
greatly different yield strength. In other words, medium- and high-strength
aluminum alloys all have about the same fatigue strength. The above graph
shows that the same is true for the corrosion-fatigue strength: there is as yet
not a single commercial aluminum alloy available with a high-cycle
corrosion-fatigue strength significantly higher than all the other aluminum
alloys. Thus, corrosion fatigue is still a limiting factor for the application of
aluminum alloys.
Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material," in Aluminum Transformation Technology and Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 616
321
322
12-4. Comparison of Aluminum Alloy Grades for Crack Propagation

Rate
1O- 3 r--- - - - - - - - - - - - - - - - - - - ---,

7050 - T 736
7175 - T 736
2219 - T 851
7079 - T 651
2618- T 6
Q)
10
~
E
'----'
10-6
Z
~
<I
10- 7
10
typical scatter in experiments
crack orientation L - T
specimens CNP, DCB I CT
ambient temperature
environment air
R =0,0-0,1
0,1 -30 Hz
10- 9
---+--------------------4
1).-....
10
20
cyclic stress intensity range
50
40
I
.t:.K I
60
MN m- 3f2]
Many commercial aluminum alloys show similar fatigue

crack propagation rates in air, as indicated in the above
comparison.
Source: Markus O. Speidel, "Aluminum asa Corrosion Resistant Material, "in Aluminum Transformation Technology and Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloniand D. E.
Embury, Eds.. American Society for Metals. Metals Park OH, 1980, p 613
12-5. Alloy 1100: Relationship of Fatigue Cycles and Hardness for

HO and H14 Tempers
50r----.,---.------.,---.-------r--r-----r--r-----r--r--,---r--,---r----,
~ 30
-----------
LU
cz::
<:
:I:
20
><::
10
AI 1100 HO
AI 1100 H14
Comparison of the Knoop hardness for well-annealed (HO) and coldrolled (H14) aluminum as a function of fatigue.
Microcrack initiation is often triggered by a dislocation rearrangement.

For instance, in the case of well-annealed Al 1100 (RO), the material will
harden in the early stages of fatigue (see S-N curves above) as the dislocation
density in the bulk of the material increases, accompanied by pronounced
slip-step formation on the surface. On the other hand, in the case of the
cold-worked material Al 1100 (R 14), the material will soften in the early
stages offatigue (above curves) as the dislocation density, introduced by the
cold work, decreases. Slip-step formation in this situation is much less pronounced than it is during hardening, because the initial dislocation-loop
length is much shorter. In either case, during this initial rearrangement, the
dislocations form a cell structure with individual dislocations of long loop
length shuttling to and fro between the cell walls. This latter part of the fatigue life is called the saturation stage offatigue, during which the dislocation
shuttling leads to local instabilities, or "extrusions-intrusions," and finally to
microcracks, which can be observed after about 25% of the fatigue life has
been expended. The micro crack density is about the same for both materials.
Source: O. Buck and G. A. A1ers,"New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 128
323
324
12-6. Alloy 1100: Interrelationship of Fatigue Cycles, Acoustic

Harmonic Generation and Hardness
8xlf3
16xlf
80
......
Cla
-e
zlA I
60
Vl
......
Vl
.....
z
-c'"
40 oe:
::J:
ll.
K.H.
0
0
ZO
z~
AI 1100
0
0
80
120
0
160xllY
FATIGUE CYCLES
Normalized second harmonic displacement and
Knoop hardness as a function of fatigue.
The effects of dislocation rearrangements on harmonic generation within the bulk of the material during fatigue are shown
in the above chart. Using 3D-MHz longitudinal waves, the normalized second harmonic amplitude of an initially compressiondeformed Al 1100 single crystal was monitored and found to
increase as a function of compression-compression fatigue. At
the same time, the surface hardness (Knoop) decreased. Apparently, the dislocation-loop length prior to fatigue was quite
short, since the initial amplitude of harmonic generation was
small. During fatigue softening, the cell structure that developed
(with its individual dislocations within the cells) became quite
large, so that a change of the dislocation-induced harmonic generation, A 2d , increased. Application of this technique to highstrength aluminum alloys failed, however, apparently because
of an immediate rep inning of the long loops by interstitials in
this alloy.
Source: O. Buck and G. A. Alers, "New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 131
12-7. Alloy 2014-T6: Notched vs Unnotched Specimens; Effect on

Cycles to Failure
50
~ 40
)(
N.
c:
~ 30
VI
VI
....
cr:
:n
20
10
Effect of notch on fatigue of 2014-T6. As is true for most

alloys, notches greatly reduce the fatigue properties of
aluminum alloys.
Source: P. C. Varley, The Technology of Aluminium and Its Alloys, Butterworth & Co. Ltd., London, England, 1970, p 43
325
326
12-8. Alloy 2024-T3: Effect of Air vs Vacuum Environments on

Cycles to Failure
..
40
~ 35
Ii
Q)
J; 30
Cl
.- Ultrahigh Vacuum
0- Air
,
~
'"
.........
..........
<,
<,
"0
; 25
~,
<,
.........
En'durance Limit
15 s
10
I I I I III"'
2
8 10'
6 8 10
.
"~
...
J
7
2.5 x 10
l/)
~~
-3
)(
-3
'~
r""- ...
III
E
::J
, 20
r-
- 3.5 x 10
1.5 x 10- 3
Number of Cycles-to-Failure
The effects of air versus vacuum on the fatigue life
of a 2024-T3 aluminum alloy.
For most materials, environment appears to be most effective

early in the crack-growth process, with little or no effect at high
crack-growth rates. Additionally, the majority of S- N curves diverge at decreasing stresses, the increase in fatigue life caused by
vacuum becoming greater at lower stresses. In contrast to this
behavior, however, aluminum and aluminum alloys have been
shown to exhibit conflicting results. For example, a 2017-T4
alloy tested in air and at 2 X 10-6 torr and a 2024-T3 alloy tested
in air and at 10-10 torr in rotating bending exhibit convergence of
S- N curves at low stresses, the effect of environment apparently
becoming less important at decreasing stresses, as shown in the
above chart.
12-9. Alloy 2024-T4 Alclad Sheet: Effect of Bending on Cycles to

Failure
400
350
rf.
:E
VI'
VI
...e
(/)
50
300
40 'w
...:
250
~.
30
200
150
-----2 1
Stress ratio. 0.1
100
0.01
....::.4"-3
0.1
en
20
10
Effects of bending on fatigue characteristics of

aluminum alloy sheet.
For the data here, sheet 1.02 mm (0.040 in.) thick was annealed, solution heat treated and quenched, and then fatigue
tested. The sheet represented by curve I was not bent. All other
sheet was bent 90 in the annealed condition. Flattening (unbending) was done in either the annealed condition (curve 2) or
the solution heat treated and quenched condition (curves 3, 4
and 5). Details of bending and flattening were as follows: (I) Not
bent. (2) Bend radius, 3.18 mm (Ys in.); flattened in annealed
condition. (3) Bend radius, 3.18 mm (Ys in.); flattened in
quenched condition after 3 days of storage at-18 to-12C (0 to
IOOF). (4) Bend radius, 3.18 mm (Ys in.); flattened in quenched
condition after 14 days of storage at -18 to -12C (0 to 10 OF).
(5) Bend radius, 1.59 mm (1/ 16 in.); flattened in quenched condition after 3 days of storage at - 18 to - I2 C (0 to IO F).
Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
327
328
12-10. Alloy 2024-T4: High-Cyclevs Low-Cycle Fatigue

40
~
LOW CYCLE t R -0 2 C. RAOIAJlON I

-_. HIGH CYCLE (R. - 1 Mo RADIATION I
I PANGBORN e I ill I
III
'5
C
30
C::P
/
/
20
/
/
/
~a:
a:
8
A
/,
/,
10
/
/
0
0
0.5
1.0
FRACTION OF LIFE In/n,1

Dependence of 13 on n/nior low-cycle fatigue and
bulk properties of high-cycle fatigue of AI 2024.
After correcting for the difference in initial /30 values, it can be

seen in the above diagram that the two fatigue processes, although radically different in strain history, exhibit similar behavior throughout most of the fatigue life.
Source: Sigmund Weissmann and William E. Mayo, "Determination of Strain Distributions and Failure Prediction by Novel X-ray
Methods," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck and Stanley M. Wolf, Eds., American
Society for Metals, Metals Park OB, 1984, p 195
12-11. Alloy 2024-T4: Relationship of Stress and Fatigue Cycles
32
30
28
u
26
.....0
24
VI
-;'"
c::
22
I~
20
sc;
18
..,.z
.c::
e'"
16
0'
'"
>
14
,,
12 :
5,000
10,000
15,000
20,000
Number of fatigue cycles, N

Dependence of 13 on number of cyclesN at various
stress levels of AI 2024-T4.
Here is shown that for the maximum stress of241 MPa with R
= 0.1, the 13 value increased during the first several hundred cy-
cles. This was more pronounced for the surface grains (Cr
radiation).
KO'I
Methods," in Nondestructive Evaluation: Application to Materials Processing, Otto Buck and Stanley M. Wolf, Eds., American
Society for Metals, Metals Park OB, 1984, P 194
329
330
12-12. Alloy 2024-T4: Dependence of the Average Rocking Curve

Halfwidth 13 on Distance From the Surface
20..--------------------,
19
18
~-_:_------,O----..."....----_1
ICQ.
14
13
_ _ _ _ _ ~o (inillal ~alf.i~_~_)
12
IIL..-_ _..........
50
.L--_ _- - ' -
100
150
....l...- _ _---J
200
250
I'm
Dependence of the average rocking curve halfwidth 11 on depth distance from surface for different fractions of corrosion fatigue lives, N F' of Al
2024-T4.
X-ray rocking curve measurements were carried out as a function of depth distance from the surface, and typical results ofthe
dependence of 13 on depth distance for an alloy cycled with a =
276 MPa, corresponding to the static yield stress, are shown
above. It may be seen that the minimum 13 values at the surface
layers were larger than those in the interior. The 13 values declined up to a depth of about 50 Mm from the surface and subsequently retained a plateau value throughout the interior of the
specimen for each fraction of the life.
Methods." in Nondestructive Evaluation: Application to Materials Processing, Otto Buck and Stanley M. Wolf. Eds., American
12-13. Alloys 2024 and X2024: Effect of Alloy Purity on Cycles to

Failure
+
260
tV
a...
:2:
CIl
CIl
220
CIl
Ol
C
.~
~ 180
<i:
140 ~
105
---'106
o ---'----'
Cycles to failure
Effect of reducing the concentration of submicron
particles in an AI-Cu-Mg alloy. X2024 is a highpurity version of the commercial alloy 2024.
The disappointing fatigue properties of age-hardened aluminum alloys are also attributed to an additional factor, which is
the metastable nature of the metallurgical structure under conditions of cyclic stressing. Localization of strain is particularly
harmful because the precipitate may be removed from certain
slip bands, which causes softening there and leads to a further
concentration of stress, so that the whole process of cracking is
accelerated.
The fatigue behavior of age-hardened aluminum alloys
would therefore be improved if fatigue deformation could be
dispersed more uniformly. Factors which prevent the formation
of coarse slip bands should assist in this regard. Thus it is to be
expected that commercial-purity alloys should perform better
than equivalent high-purity compositions because the presence
of inclusions and intermetallic compounds would tend to disperse slip. This effect has been confirmed for an Al-Cu-Mg alloy, and fatigue curves for commercial-purity and high-purity
compositions are shown in the above S-N diagram. Here the
superior fatigue behavior of the former alloy arises because slip
is more uniformly dispersed by submicron dispersoids such as
MnAI 6
Source: I. J. Po 1mear, Light Alloys, Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981, P 40
331
332
12-14. Alloys 2024 and 2124: Relationship of Particle Size and

Ii K, ksi in. 1/ 2
10
15
20
10- 3
10- 2
Q)
Q)
>
u
--
E
E
10-4
10- 3
10-5
'C
10-6
10-5
..
.'"
~
0
en
10- 7
en
'';:::;
Q)
'"
'0.
c,
:l
en
'~
'C
.~
Q)
e,
UI
.. -~
'u
U.
10-6
1......
.:.:
o
Q)
Q)
:l
'"
....
...
en
U.
'C
'"
.s:
~
0
--'"
Q)"
10- 4
s:
.:.:
--.5
'C
'C
Q)"
>
e
--'"
.....
...'"
.
'".
Q)
"---'~_ _"""""_ _"""""---''--_''''''''''_'''''''''---' 10- 8
20
Stress-intensity factor range, LiK, MPa m 1/2
Comparison of typical particle sizes in aluminum alloys
with crack advance per cycle on fatigue loading.
The above graph represents Staley's work in summarizing the

role of particle size on fatigue crack growth in aluminum alloys,
Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys." in Application of Fracture Mechanics for
Society for Metals, Metals Park OK 1982. P 191
12-15. Alloys 2024-T4 and 2124-T4: Comparison of Resistance to

Mechanically Polished R=-I
2.6
2.4
...
in
co
..r:::
CIl
200
fJl
"0
0>
0
300
fJl
fJl
Ql
AA
400
...
Ql
fJl
Nij =Cycles to ~ 15 fLm Crock
fJl
fJl
Ql
in
600
500
~
<,
700
2024 T4 oNjj
2124 T4 ANjj
2.8
co
2.2
A Ao
...J
2.0
102
103
105
..r:::
AA
"0
z
106
107
Cycles to fatigue crack initiation for specimens of aluminum alloys 2024-T4 and 2124-T4 versus stress at notch
base (computed using Neuber stress-concentration factor).
The 2124 alloy studied had 1/ 10 the inclusions of the 2024

alloy studied (0.2 vol% compared to 2 vol%) but a larger grain
size (45 /-lm compared to 20 /-lm in the transverse direction normal to the loading direction). With 2124-T4, slip-band cracks
not associated with inclusions formed at the lowest stress studied. They also formed more easily in 2124-T4 than in 2024-T4
at high stresses, in keeping with the larger grain size. Thus, as
shown in the above chart, at high stresses 2124-T4 is less resistant to fatigue crack initiation than is 2024-T4, but it is more
resistant at low stresses.
Source: M. E. Fine and R. O. Ritchie, "Fatigue-Crack Initiation and Near-Threshold Crack Growth," in Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, P 251
333
334
12-16. Alloys 2024-T3 and 7075-T6: Summary of Fatigue Crack

Growth Rates
t.K, ksi . in.1/2
10
100
10- 1 r - - - - - - - , - - - - - -....- - - - - - - . ,
7075-T6
9 investigations
10- 4
2024-T3
8 investigations
>u
'<,
10- 5
10-
Note:
Bounds defined
by mean curves
of separate
investigations
Cl
~
E
u
~
G>
10-
.~
z
"C
......
01
"C
10- 6
Cl
.~
U.
10-7
10- 6
10
100
Stress-Intensity factor range, AK, MPa m 1/2

Summary of fatigue crack growth rate data for
aluminum alloys 7075-T6 and 2024-T3.
Considerable use has been made of the fracture mechanics

approach in measurement of fatigue crack growth rates in aluminum alloys. These data have been generated by methods comparable to those of ASTM Method E647 for measuring fatigue
crack growth rates. In general, fatigue crack growth rates are
found to fall within a relatively narrow scatter band, with only
small systematic effects of composition, fabricating practice or
strength level, as illustrated by the data in the above chart.
Source: J. G. Kaufman and J. S. Santner, "Fracture Properties. of Aluminum Alloys," in Application of Fracture Mechanics for
Selection of Metallic Structural Materials, James E. Campbell, William W. Gerherich and John H. Underwood, Eds., American
Society for Metals, Metals Park OH, 1982,P 189
12-17. Alloys 2024-T4 and 7075-T6: Effect of Product Form and

Notches
'"
rn
UI
414
_ - -----+---j--+_
345
~---l~-~~
e,
- --
::E
276
207 I
a:
l-
rJ)
138
2024-T4 ---+----1
."
~.,-::'~---+---+---j
I,..... 1'I,--'~-r-''"'
". - ~.....;-"'~_ _+-_--j
~ ..
"
'"'-I' ,t'i:::::~"
J
-, <,
:.:~,.;-.~
c ~- -
<,
'q".
~.
"..
69
0
I---L.--+~-t---+-
414
I", -,,
7075-16
~----I
345 I----i---"~~.-:-l.c---+----jf----+----j
'"
c,
::E
rn
rJ)
a:
~'
.\ . f~,~~.l,.
~
276
207 1---+---''':-....'
lUI
''':".":--
--+
,,--
- - -
.~":''-1''''.1
"
t': ':', .... ~'~r, ::""II'jl
~.
,', " .,'\ ~,.
"
"e'::;"',"-j--""'-f'--..
138 ~c-'-cc-:--''''.j,-,''-...,.;t-,:--,:,r-'- ,,"00.
69
o
1~
:
,
~.,"
: ~gmg
~~~TE
FORGINGS
1'"
....::.:,
EXTRUSIONS
010 NOT FAil
----L
1~
1~
1~
---~..!...:'.I,
'---->":"
-.#_.s
. _ _ . __ _ _
1~
10'
!
I.
-. --}.
"-,.J.'
108
109
CYCLES
Fatigue performance of smooth and notched (K, > 17)

rotating-beam specimens from various product forms of
2024-T4 and 7075-T6 aluminum alloys.
Numerous methods have been developed to evaluate response of materials to cyclic deformation.
The earliest method was by use of S-Nplots. Typical examples are depicted above. Basic specimens
include rotating beam, axially stressed and sheet flexure. Notches have been employed to provide stress
concentration, and special specimens have been used to simulate a variety of other conditions. The S- N
response is strongly influenced by a number of conditions, including surface condition, stress ratio, and
environment. The various alloys differ widely in their response to fatigue testing-specifically, in the
number of cycles where a "level out" condition is attained. As shown in the above S- N diagrams, the
SoN response for aluminum alloys tends to level out as the number of applied cycles approaches 500
million.
Based on SoN data of smooth and sharply notched specimens and of similar tests of specimens
designed to simulate joints in structures, the following conclusions have been drawn. From fatigue
results for aluminum alloys obtained with smooth specimens . . . rather wide variations can exist
without causing appreciable differences in fatigue strengths. . . . When severely notched specimens
are used, the effects of composition and temper are even less pronounced and generally are of no
practical significance. . . . As in the case of simple notch fatigue tests, there is a lack of significant
differences in the fatigue strength of the joints of the various alloys.
Despite these laboratory data, users discovered that certain aluminum alloys performed decidedly
better than others in service when fluctuating loads were encountered. For example, airframe manufacturers determined that fatigue performance of alloy 7075-T6 was unquestionably inferior to that of
alloy 2024-T3.
Source: T. H. Sanders, Jr. and J. T. Staley. "Review of Fatigue and Fracture Research on High-Strength Aluminum Alloys." in
Fatigue and Microstructure. American Society for Metals, Metals Park OH, 1979. P 470
335
336
12-18. Alloys 2024-T351 and 7075-T73XXX: Comparison of P1M

Extrusions and Rod
500 r------,r------,-----,.---~---_,_---__r---__, 70
1.35 mm (0.053 in.)
6.43 mm (0.253 in.)
f
---l_+_~=1B
--F
+-__-+-_ _
400 1-_ _
<0
0..
::E
Stress ratio: R = 0
Ambient air
vi
III
-+-_ _--+
t -_ _-1 60
Notch tip radius:

e = 0.33 mm 10.013 in.)
K, = 3
50
300 f----I------1I-------1I-----j----j-----t-----l
~
III
E
:J
E
'x
<0
::E
E
:J
E
'x
<0
::E
200 1 - - - - / - - - + - I f - 1 . * - - - I - - - - - t - - - - - t - - - - - + - - - - - - l
20
-:::~~~~~"'F;;;;;:;:;;:::::j::~-__jl--I
100 f -_ _ 2024-T351
rod and bar band -
~"t----I-~ 7075T73XXX
10
products band
oL -_ _-.l.
10'
103
....L
...L.
10'
105
L -_ _---'
10'
10'
....L_ _~
10'
10'
Cycles
Comparison of room-temperature axial stress notch fatigue strength ofP 1Malloy

extrusions and ingot metallurgy alloy rod, bar and products, 0 ,X7090T7E71 in
the longitudinal direction; , X7091-T7E69 in the longitudinal direction; 6.,
X7091-T7E69 in the long transverse direction; - denotes test specimen did not fail,
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,p 468
12-19. Alloy 2048-T851: Longitudinal vs Transverse for Axial Fatigue

500
'"
70
60 ""'"
Q..
:i:
400
::I"
::I"
I;.
'" 300
E
:::J
'x 200
'"
:i:
100
103
R = 0.1
o 0 I;. Longitudinal
I
... Longtransverse
10 4
I;.
105
106
Numberof stress cycles
50 ~
'"
40 E
:::J
E
30
'x
'"
:i:
20
107
SoN axial fatigue curves for unnotched specimens of

aluminum alloy 2048T851 plate, showing effects of R
value and direction upon fatigue properties.
Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals. American Society
337
338
12-20. Alloy 2048-T851: Notched vs Unnotched Specimens at

Room and Elevated Temperatures
500
s:
:::E
400 ~
Ii:
E
300
.~
200
10
Unnotched
R
r-,
0.1
60 ]
50
24C 176of)
176C 13600F)~_
g
E
/120oC 1260 of)
40 ~
-=-
30
:::E
20
101
...:::E..
400
g ~
300
E
~ 200
.~
:::E 100
Notched (K t '" 3.01

R
~ s,
0.1
24C (76 of)
~J
120 c 1260 of)
176C 1360 OF)
50 ]
- 40 g
30
_ 20
--;;;;1
"i
:::E
S-N curves for unnotched (upper graph) versus

notched (lower graph) specimens of aluminum alloy 2048-T851 plate.
Source: Metals Handbook. 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals. Metals Park OH, 1979, P 82
339
12-21. Alloy 2048-T851 : Fatigue Crack Propagation Rates in LT and

TL Orientations
1
AK. kSi'in. 2
AK. kSI'in.2
10
10
LT crack oriant!tion
TL crack orianLtion
l~3.4
"
,
10-8
10- 5
{.:
.5
~.
..:!
rot"
-1i
~
10-5
.5
..:!
..:!
10-6
M-a.:
M~7.1
61
10
"iI
10- 6
61
10
Fatigue crack propagation in aluminum alloy 2048-T851 plate, showing

propagation data for both LT and TL (longitudinal and transverse)
crack orientations.
for Metals, Metals Park OH, 1979, p 81
340
12-22. Alloy 2048-T851: Modified Goodman Diagram for Axial

Fatigue
Minimum stress,ksi
s:
60
::;:
~
1;;
"E
'x
..
E
E
"
200
.~
::;:
::;:
Minimum stress, MPa
Modified Goodman diagram for axial fatigue of un notched specimens

of aluminum alloy 2048-T851 plate.
12-23. Alloy 2219-T851: Dependence of Relaxation Behavior on the

Cyclic Hardening Parameter
O~-----r-------,------r-------,r-----...,
0% RH
50% RH
-300
L--
....J..
10
...!-
20
.l..-
30
---I.
40
...J
50
CYCLES (x 10-3)
The dependence ofrelaxation behavior on the cyclic hardening parameter, (J. (J was varied by changing the relative humidity (RH), which affects the near surface ductility in this alloy. Values used were: (J= 6 X 10.5
for 50% RH and (J= 2 X 10-5 for 0% RH. The cyclic stress amplitude was
0.88 a yleld for both samples.
Source: M. R. James and W. L. Morris. "The Relaxation of Machining Stresses in Aluminum Alloys During Fatigue." in Residual
Stress for Designers and Metallurgists. Larry J. Vander Walle. Ed.. American Society for Metals. Metals Park OH. 1981. P 184
341
342
12-24. Alloy 2219-T851: Effect of Strain Amplitude on the

Relaxation of Residual Surface Stress With Fatigue
Or-------.-------,r------.------,----...,
0.7" YIELD
o
0.64 a YIELD
-300
.....
o~-----'--------l----...J-.-------l----
10
20
30
40
50
CYCLES (x 10- 3)
The effect of strain amplitude on the relaxation of surface residual stress

with fatigue. The symbols are the residual stress value measured by the
x-ray diffraction peak shift technique. The solid curves are the predicted
mean residual stress values during fatigue.
Surface milling produced the shallowest stress gradient and resulted in the
slowest rate of relaxation of the surface stresses. A comparison of measured
to predicted values of residual stress during fatigue is made for four "as machined" specimens in the above chart. The residual stress values were measured parallel to the external stress axis. A value of f3 = 0.0004 was used to fit
the data for all specimens. Residual stress measurements were also made in a
direction transverse to the applied stress axis. Within experimental error, the
cyclic relaxation rate was the same as in the longitudinal direction.
Source: M. R. James and W. L. Morris, "The Relaxation of Machining Stresses in Aluminum Alloys During Fatigue," in Residual
Stress for Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, P 182
12-25. Alloy 2219-T851 : Relationship of Fatigue Cycles to Different

Depth Distributions of .Surface Stress
O....------r-----r-----,-----r-----,
-300
l--
--L
-L-
10
20
..L....-
30
----I
40
.....
50
CYCLES (x 10-3)
The relaxation behavior oftwo samples having different depth distributions of residual stress. Note the difference in the peak cyclic stress, a.
0= rolling (10%reduction); f3 = 0.012; a= 0.91 ayield'. = sand blasting;
f3 = 0.003; a = 0.71 ayield'
Relaxation of a compressive surface stress requires an expansion of the

material normal to the surface. Of necessity, this involves slip at an acute
angle to the surface. If the slip does not penetrate the surface, the residual
stress cannot relax. Supporting this picture are our observations that the
relaxation rate in Al 2219-T851 is more rapid in dry air. It is known that
humidity increases the rate of cyclic hardening of a thin (less than I Jlm) layer
at the surface. The effect of humidity on relaxation is therefore simply to
make it more difficult for dislocations to penetrate to the surface.
Source: M. R. James and W. L. Morris. "The Relaxation of Machining Stresses in Aluminum Alloys During Fatigue," in Residual
Stress for Designers and Metallurgists. Larry J. Vander Walle. Ed. American Society for Metals, Metals Park OH. 1981, P 183
343
344
12-26. Alloy 2219-1851: Probability of Fatigue Failure
LOG (NUMBER OF FATIGUE CYCLES)

Schematic curves of constant probability for failure (actual failure = 100%).
The solid line in the graph represents failure; the dashed lines
indicate the percentage offatigue life expended. The exact location of these lines is highly sensitive to the material and its
microstructure as well as the influences of environment.
Source: O. Buck and G. A. Alers, "New Techniques [or Detection and Monitoring of Fatigue Damage.t' in Fatigue and Microstructure. American Society for Metals. Metals Park OH, 1979. p 104
12-27. Alloys 3003-0, 5154-H34 and 6061-T6: Effect of Alloy on

Fatigue Characteristics of Weldments
300
40
260
30
If. 200
:t!
::;;
g
E
E
160
20
"
..
'x
::;;
E
E
.."
'x
::;;
100
10
50
Number of cvcles
The fatigue life of welded joints at high loads varies with the
alloy. As the load is decreased, differences disappear until, at
about one to ten million cycles of axial loading (R = 0), the
fatigue strength of an arc-welded joint is approximately the
same regardless of alloy and is 50 to 70% that of the unwelded
alloy. Typical data are given in the above graph for three aluminum alloys. Specimens were from 9.5-mm (Ys-in.) plate; weld
reinforcement removed; axial loading; R = O.
345
346
12-28. Alloy 5083-0 Plate: Effect of Orientation on Fatigue Crack

Growth Rates
.1K, ksi in. 112
10- 4
..
u
10- 3
>
~
-.
-.t.l
E
E
10-5
'C
-.
Cll
'C
.E
'C
-.
Cll
'C
10- 4
10- 6
T-S
10-5
Compact specimen thickness

= 46 mm (1.8 in.)
R = 1/3, f = 13 Hz
Room temperature, dry air
.1K, MPa . m 1/2
Effect of orientation on fatigue crack growth rates

in 180- and 196-mm (7.0- and 7.7-in.) 50830
plate.
From the data shown above there is obviously no great effect

of specimen orientation on fatigue crack growth rates.
Source: J. G. Kaufman and J. S. Santner, "Fracture Properties of Aluminum Alloys," in Application of Fracture Mechanics for
12-29. Alloy 5083-0 Plate: Effect of Temperature and Humidity on

~K, ksi in. 1/2
10- 2
,/
RT, moist
air~i
,,
,,,I
,
10- 3
Ql
"C
Ql
EE
z
~
10- 4
10- 5
.E
~
III
"C
10-4
10-6
Compact specimen thickness
= 46 mm (1.8 in.)
R = 1/3.
f=13&18Hz.
T-L orientation
10-5
'--
...L-_--L.
5
~K.
10
'--~
10- 7
50
MPa m1/2
Effect of temperature and humidity on fatigue

crack growth in 180-mm (7.0-in.) 5083-0 plate.
As shown in the above graph, growth rates for alloy 5083-0

are appreciably higher in moist air than in dry air. Growth rates
in water solutions of sodium chloride are similar to those in
moist air.
Selection of Metallic Structural Materials, James E. Campbell, William w. Gerberich and John H. Underwood, Eds., American
347
348
12-30. Alloys 5086-H34, 5086-H36, 6061-T6, 7075-T73 and

2024-T3: Comparative Resistance to Axial-Stress Fatigue
0.7
R~O
Frequencv = 1.1 kHz

0.6
0.5
0.4
....
.~
0.3
a:
0.2
0.1
5086-
5086-
6061-
7075-
2024-
H34
H36
T6
T73
T3
Ratio of axial-stress fatigue strength of aluminum alloy

sheet in 3% NaCI solution to that in air.
Fatigue strengths of aluminum alloys are lower in corrosive

environments such as seawater and other salt waters than they
are in air, especially when evaluated by low-stress, long-period
tests. As shown in the above bar chart, such corrosive environments produce smaller reductions in fatigue strength in alloys of
the more corrosion-resistant types, such as 5xxx and 6xxx alloys, than in less resistant alloys, such as those of the 2xxx and
7xxx series. Like stress-corrosion cracking of aluminum alloys,
corrosion fatigue requires the presence of water. In contrast to
stress-corrosion cracking, corrosion fatigue is not appreciably
affected by test direction, because fracture resulting from this
type of attack is predominantly trans granular.
12-31. Alloys 5083-0/5183: Fatigue Life Predictions and

Experimental Data Results for Double V-Butt Welds
IOOr---r-,--...--I'""T'"T'TTT"--,----.-r....T"T"TT,----r-,--...--I'""T'"TT"n
~
~
60
400
40
300
200
.:
1II
<I
10
8
CT,=+18kSi
KI~
5083 -0/5183 Double - V Butt Welds
I 4
10
s~s
' mol =2.60, 0ls0.0Iln., R =0, 1= 3/8 .n.

CT, + 18 ksi
30
20
10
10~
NT' Cycles
Total fatigue life predictions and experimental results for

5083-0/5183 3/8-in. (10-mm) butt welds.
Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 114
349
350
12-32. Alloys 5083-0/5183: Predicted Effect of Stress Relief and

Stress Ratio on Fatigue Life of Butt Welds
1OO~"""""""""""'--''''''''''''''''''''''~-~''''''''''''''''''''''''''''''"T''T"r--''''''---'---'''''''''''''''''''''''''600
80
eo
40
~3-o1!l183
Bull Weld
Kt..." 3.HI.
400
300
1" 0.0111\ (0.2~41Ml)

."90",' "SO", '" 112ir\ 112 7mm)
- - .,.5,
200
--- .,0
R"O
~-.
S~S
1(,...
'
30
20
10
Predicted effect of stress relief and stress ratio on

5083-0/5183 butt weld fatigue life.
Because of the high notch-root plasticity during the first few

cycles, before the material cyclically hardens, the aluminum
weld considered here (5083/5183) exhibits little dependence
upon either residual stress or stress ratio, even though the relaxation of the stabilized mean stress (uos) is very slow-as indicated in the above chart.
Source: F. V. Lawrence, "The Predicted Influence or Weld Residual Stresses on Fatigue Crack Initiation," in Residual Stress for
Designers and Metallurgists, Larry J. Vander Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 113
351
12-33. 7XXX Alloys: Cyclic Strain vs Crack Initiation Life

100
.fIN
AllOY
7075T6
7050T6
7075T7
7050-T7
W
0
:::l
f-
::J
Q..
::iE
UTS
kSi (MPa)
810(558)
887(612)
73.2(505)
74.1(511)
<>:
TYS
ks; (MPa)
769(530)
82 7(570)
658(454)
63.8(440)
./,
fl. 2"
95
12.0
11.0
14.0
z
<i:
a:
f-
en
...J
10- 2
/7050.T6
<>:
:-~-~ ----.
--"""""''''''''""""",""
f0
f-
-------
"""'''''''''''''1;;./11
1O-3 ':'":-_ _---'100
-'-
-':-_ _----J'-_ _--'-
-'-_ _............
10 7
REVERSALS TO INITIATION, 2N
Cyclic strain versus initiation life for laboratory-fabricated high-strength 7XXX aluminum alloys,
Plots of elastic-strain amplitude versus life have seen relatively little use for commercial alloys, but plots oflog total strain
amplitude versus life have been used more frequently to compare materials. This approach offers the advantage that both
high- and low-strain fatigue may be characterized with one plot.
As illustrated above, fatigue resistance at low total strain amplitude is governed by the elastic-strain amplitude. Fatigue lives
for total strain amplitudes less than about 5 X 10-3 generally
increase with increasing strength. On the other hand, fatigue
lives for total strain amplitudes greater than about 10-2 generally
increase with increasing ductility.
Source: T. H. Sanders, Jr., and J. T. Staley, "Review of Fatigue and Fracture Research on High-Strength Aluminum Alloys," in
Fatigue and Microstructure, American Society for Metals, Metals Park OH, 1979, p 472
352
12-34. Alloy 7050: Influence of Alloy Composition and Dispersoid

Effect on Mean Calculated Fatigue Life
Zn
Mg
Cu
5.5
2.2
2.3
to
to
to
6%
2.4%
2.4%
Aged 3 h at 121C (250F)

+ 9 h at 163C (325F)
1.0
Alloy 7050 sheet
6%Zn
2.2% Mg
0.4% Mn
1.2
Low humidity
Aged 3 h at 121C
(250F) + 9 h at
163C (325F)
Alloy 7050
1.0
0.8
High humidity
0.6
0.4
0.2
% Zr
%Mn
0.1
0.4
0.1
0.4
%Cu
1.0
2.3
1.0
2.3
Effect of dispersoid type (based on composition) on fatigue crack propagation life of 7050 alloy sheet.
The influence of alloy composition on dispersoid effect is shown in the

above bar chart. The general trend in this chart is that for more finely
dispersed particles the fatigue crack propagation life is increased. Whereas
dispersoid type appears to have a relatively small effect on mean calculated
life, the smaller precipitates provided by aging produce a much larger effect.
There is some evidence that new processing practices may provide the fine
microstructures needed to enhance fatigue resistance. The potential of intermediate working (commonly referred to as ITMT treatments) remains attractive but has not been proven for notched specimens.
Society for Metals, Metals Park OH, 1982, p 192
12-35. Alloy 7050: Effect of Grain Shape on Cycles to Failure
Q.
.400
UJ
::J
I...J
Q.
~ 3
(J)
(J)
UJ
7050 AR
7050 HR
II:
:n
200
...
105
CYCLES
106
TO FAILURE
Stress-life curves for two 7050 alloys having fine,

equiaxed grains (AR) and pancake-shaped grains
(HR).
As indicated in the above graph, grain shape showed no perceptible difference in life over a range of stress amplitudes.
353
354
12-36. Alloy 7075 (TMP. T6 andT651): Effect of Thermomechanical

Processing on Cycles to Failure
300
~ od~
\
m
a..
~
tl
c
200
0\0
&
6-
<i:
0
"-
\6,
'
0 0'0Q,
""4_
6
100
L..-
10
TS
632
627
573
567
'6
\
0
6'h
CJ)
.~
\~
\
If)
If)
PS
587
600
516
488
7075 TMP
... 7075 TMP
o 7075-T651
67075-T6
"""'---
'lIQ .Q..
--_~~
"""'---
.........
........
Cycles to failure
Effect ofthermomechanical processing (TMP).on the unnotched fatigue properties of the commerical AI-Zn-MgCu alloy 7075. PS = proof stress (MPa); TS = tensile
strength.
Detailed studies of the processes of fatigue in metals and alloys have shown that the initiation of cracks normally occurs at
the surface. It is here that strain becomes localized due to the
presence of pre-existing stress concentrations such as mechanical notches or corrosion pits, coarse (persistent) slip bands in
which minute extrusions and intrusions may form, or at relatively soft zones such as the precipitate-free regions adjacent to
grain boundaries. Density has also been found to improve the
fatigue performance of certain alloys, although this effect arises
in part from an increase in tensile properties caused by such a
treatment (see above diagram). It should be noted, however,
that the promising results mentioned above were obtained for
smooth specimens. The improved fatigue behavior has not been
sustained for severely notched conditions, and it seems that the
resultant stress concentrations override the more subtle microstructural effects that have been described.
Source: I. J. Polmear, Light Alloys. Edward Arnold Ltd. London, England, and American Society for Metals, Metals Park OH,
1981.P 41
12-37. Alloys 7075 and 7475: Effect of Inclusion Density on

Cycles to Failure
1:1
400
Q.
~
ILl
;:)
I-
300
:J
Q.
~
c:(
I/)
I/)
200
ILl
II::
l-
I/)
100
104
105
106
CYCLES TO
107
108
FAILURE
Effect of inclusion density on the stress-life behavior oftwo 7XXX alloys: high-inclusion density, alloy 7075; low-inclusion density, alloy 7475,
355
356
12-38. Alloy 7075: Effect of TMT on Cycles to Failure
400
'E
z
7075
oY;I~Nm-' ]
7075TMT 600
:gw
300
a:
~
til
Cl
Z
200
a:
w
100-'----,..---.-----,-----.----.--10'
10'
10'
107
10'
CYCLES TO FAILURE
Influence of TMT on SN curves (R = -1).
There is evidence in the literature that a uniform dislocation

density introduced by cold working improves the fatigue life
also in connection with FTMT. The above graph shows an example taken from the work of Ostermann. Most of these improvements are due to an increased yield stress.
Source: G. Lutjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, p 195
12-39. Alloys 7075 and 7050: Relative Ranking for Constant

Amplitude and Periodic Overload
fi
. . . . .
?<
::'::)
2.0 X 106
Total life
c~ns.tant amplitude life

Life Increase due
to retardation
Overload ratio
= 1.8
Applied every
4000 .cycles
Overload ratio
= 1.8
Applied every
8000 cycles
'0
1.0 X 106
E
::::J
0.5 X 106
Constant
amplitude
(Overload ratio
= 1.0)
Overload ratio
=1.4
Applied every
4000 cycles
I
Relative ranking of fatigue life of 7075 and 7050 aluminum alloys under constant
amplitude and periodic single overload conditions.
Crack-growth retardation is caused by tension overloading during fatigue testing.

The variable-amplitude test is believed to be more sensitive to alloy difference, and it
clearly provides more useful information for alloy-development investigations. For
example, as illustrated by the data for alloys 7075 and 7050 in the above graph, quite
different results are obtained in constant-amplitude tests than in tests with single overloads every 4000 or 8000 cycles. Thus, information on the variation in load level during
fatigue cycling is required for correct characterization of the fatigue behavior of aluminum alloys.
357
358
12-40. Alloy 7075: Effect of Environment and Mode of Loading

o
~
160
AIR
NaCl
'0
;f
-0
---
.0... - - 0 -
.; 120
'"
~
a:
I;;
u
::;
80
MODE I LDADING
'"
,I
7
10
AIR
Nael
;f
.;
'"
'"
....a:
'"
u
120
uJ
::;
--_~
80
u
>u
40
0 ....._
10~
MODE J LOADING
........-1.-1.........................._
o
0
--0

........- - ' - - '.....................' - - _........- - ' - - ' - ' "....., ...............
Fatigue behavior of 7075 aluminum alloy in air and aerated sodium

chloride solution: (above) under mode 1 loading; (below) under mode 3
loading.
Tests performed on a commercial 7075 alloy in a mode 3 loading condition (torsion) indicated that
the reduction in fatigue resistance associated with cathodic charging was considerably less than it was
under mode I loading (note above charts). Although total immunity to corrosion fatigue was not
observed, the slight reduction in fatigue resistance can be associated with conditions that did produce a
true mode 3 loading condition both on a micro-scale and on a macro-scale.
To summarize the aluminum alloy results, it appears that corrosion reactions liberate hydrogen,
which effectively embrittles the region in the vicinity of a crack tip. The specific details of the embrittlement are not known, but it appears that dislocation transport of the hydrogen is involved. It has been
speculated that hydrogen may collect at the semicoherent precipitate-matrix interface, thus explaining
the reported fracture plane; however, a great deal more research will have to be performed before a
more definitive answer will be available.
12-41. Alloy 7075-T6: Effects of Corrosion and Pre-Corrosion

-------------------
200
7075 AI T6
0.5 M NoCI
276 MN/m (40 ksil mean stress
RT
- 28
24
N-
E 150
20
<,
'iii
V>
V>
w
cr 100
l-
V>
u
u
u>-
:i
50 -
- 16
""
12
A AIR
B CORROSION
C PRE CORRODED/AIR FATIGUE
D PRE CORRODED/HEAT TREATED /AIR FATIGUE
L.LLL.J...U _ _L
I
I I l l d _ _ .I_-.l...0--4
5
10
106
10
0
7
10
The effects of corrosion and pre-corrosion on the fatigue lives of a 7075T6 alloy. Note that re-solutionizing and re-aging the alloy after precorrosion results in a significant increase in fatigue resistance.
Fatigue resistance of high-strength aluminum alloys is severely affected by corrosive

solutions, especially chloride solutions, and this behavior has been attributed either to
preferential dissolution at the tips of the growing cracks or to preferential adsorption
of damaging ionic species. Experiments on a 7075-T6 commercial alloy and on a highpurity analog of the alloy (AI-5.0Zn-2.5Mg-1.5Cu) indicate that localized hydrogen
embrittlement may be responsible for the poor corrosion fatigue resistance of these
alloys. For example, the above diagram shows the results of fatigue tests performed on
the 7075 alloy under simultaneous exposure to cyclic stresses and a corrosive environment (curve B) compared to tests performed in laboratory air (curve A). If specimens
are pre-corroded and tested in laboratory air, there is also a significant reduction in
fatigue resistance (curve C). The reduction in life at low Nfis associated with pits which
form at nonmetallic inclusions. If the alloy is re-solutionized and aged, equivalent to a
low-temperature bake, a significant amount offatigue resistance is regained, indicating at least partial reversibility of the damaging phenomenon and strongly suggesting a
solid-solution effect arising from environmental interaction.
Source: D. J. Duquette, "Fundamentals of Corrosion Fatigue Behavior of Metals and Alloys," in Hydrogen Embrittlement and
Stress Corrosion Cracking, R. Gibala and R. F. Hehernann, Eds., American Society for Metals, Metals Park OH. 1984, P 265
359
12-43. Alloy 7075: Effect of Cathodic Polarization on

Fatigue Behavior
2U
.-... ,.
-,
~"
' ~~'
I~
L .. .' ,.
.....
10
. / " " 13v Na2S04
13v NaCI
.---- .
N,
Effect of cathodic polarization on the fatigue behavior of

7075 AI alloy in NaCI and Na Z S0 4'
It had been previously observed that halide ions are particularly damaging to the fatigue behavior of Al alloys; however, if
the alloy is cathodically charged during stressing, sulfate ions
prove to be equally damaging, particularly at long NJ , At lower
NJ the slight decrease observed in cr solutions appears to be
associated with damage to the passive film, as shown in the
above S-Ndata, In SO~ solutions, a crack must initiate to break
the protective film to allow access to the bulk alloy, Cathodic
charging of the high-purity analog of the 7075 alloy also shows a
reduction in fatigue resistance, In many cases, fatigue crack
initiation in the equiaxed-grain high-purity alloy is intergranular, and at more active cathodic potentials there is a tendency toward a higher percentage of transgranular cracking.
Source: D. J. Duquelle, "Fundamentals of Corrosion Fatigue Behavior of Metals and Alloys," in Hydrogen Embrittlernent and
Stress Corrosion Cracking, R. Gibala and R, F. Hehemann, Eds., American Society for Metals, Metals Park OH, 1984, P 266
361
362
12-44. Alloy 7075-T6: Effect of Surface Treatments and Notch

Designs on Number of Cycles to Failure
o PAlMlIIC ACIO
" ANODIZED AND WATER SEALED
50
41
40
"j
ij
!:1'"
}I
I;;;
......'------------~
}O
I~
4
10
4
10
NUMBER OF CYCLES
lOB
Tension fatigue test of 7075-T6 aluminum alloy sheet, notch factor K T = 1.

}D
o PAlMIIIC ACID
" ANODIZED AND WATER SEALED
21
~20
"
ol-
----------
ij
~15
t;;
10
10
I~
104
NUMBER OF CYClES
Tension fatigue test of 7075-T6 aluminum alloy sheet, notch factor K T = 2.37.
363
45
BAREPQISHffi
o ANODIZm AND NOI WAHR SALD
o PALMIIIC ACtO
40
000
ANOD Izm AND HOI

WATER SAUD
20
101
6
10
NUMBER OFCYCLES
Flexure fatigue test of 7075-T6 aluminum alloy sheet.
The three charts show the effects of notch designs and surface treatments on fatigue properties
of aluminum alloy 7075-T6 sheet; the table shows the effects of 17 surface treatments.
Surface treatment
No. of cycles
to failure
Polished
.
125,000
Anodized and water-sealcd
125,000
Propionic acid 2,800,000
Vale ric acid 15,000,000
Caproic acid 9,200,000
Octanoic acid 12,300,000
Decanolc acid
.. 7,500,000
Lauric acid 8,600,000
Myristic acid 11,600,000
Surface treatment
No. of cycles
to failure
Palmitic acid 30,000,000

Stearic acid 8,700,000
Docosanoic acid 6,000,000
Sebacic acid 13,700,000
Octyl alcohol 6,000,000
Dodecyl alcohol 7,000,000
Dodecylamine 18,500,000
Hexandeiamine 3,000,000
Sheet was anodized: 15% sulfuric acid, 23 C, 15 amp/sq ft, 40 minutes.

Stress amplitude: 26,000 psi.
Source: Irvin R. Kramer. "Improvement of Metal Fatigue Lifebya ChemicalSurface Treatment, "in Fatigue-An Interdisciplinary
Approach, John J. Burke, Norman L. Reed and Volker Weiss, Eds., Syracuse University Press, Syracuse NY, 1964, pp 250,251
364
12-45. Alloy 7075-T6: Effect of R-Ratio on Fatigue

Crack Propagation
1 -
10
0.00
.20x
I;
.33e
.50l>
.700
10- 1
f--
f--
a...
~
.80-
10
R
0.00
.20
.33e
.50l>
.700
.80 -
f--
Q)
>.
Q)
U
>.
o
.....
E
E
.....u
10- 2
E
E
f--
~10
<I
I
z
.....
"
C
"
lie:
~
It
10- 3 f--
...
1-
~
~
~
.....
.g
~
o
o
101
10
20
30
10
30
tJ.K (MPa Ifffi )
AK(MPavm)
Fatigue crack propagation in aluminum alloy 7075-T6 showing the effect of R ratio
and the applicability of the Forman, Keraney, and Engle relation. The scatter in the
data is much less in the latter.
The above diagrams show that data scatter is much less when the rate daldN is
computed according to the equation due to Foreman et al. R< O. The proposed Foreman equation is:
da
C(t::.K)1I1
dN
(I -R) K c -
Source: Marc Andre Meyers and Krishan Kumar Chawla, "Mechanical Metallurgy: Principles and Applications," Prentice-Hall,
Inc., Englewood Cliffs NJ, 1984, p 716
12-46. Alloy 7075: Effect of Predeformation on Fatigue

Crack Propagation Rates
VACUUM
-2
10
oJ
.9!
10
u
>.
E
E
z
Ci
"tl
"tl
,~
10
12
15
20
11K [ MNm-JI2 )
30
40
50
Influence of predeformation by cold rolling on fatigue

crack propagation rates for 7075. Ih 100C, SEN specimens, vacuum, R = O.I,j= 30 Hz.
Cold deformation also increases the fatigue crack propagation rate as shown in the above graph, which compares an undeformed structure with 10% and 20% cold rolled structures.
Source: G. Liitjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
D. E. Embury, Eds.. American Society for Metals. Metals Park OH, 1980, P 207
365
366
12-47. Alloys 7075 and 2024-T3: Comparative Fatigue Crack

Growth Rates for Two Alloys in Varying Humidity
10"
c:--,---,--,r-r-,.,..-----,---r--____:::>
78910
20
1
(MPa m / 2 )
Comparative fatigue crack growth rates for aluminum alloys 2024-T3 and 7075-T6 in air ofvarying humidity.
Relationships between rate of growth of fatigue cracks and

stress intensity for the alloys 2024-T3 and 7075-T6 are shown
above. Other 2xxx series alloys show rates of crack propagation
similar to tha t of 2024-T3 over most of the range of test conditions. In general, these alloys have rates of crack growth that are
close to one-third those observed in the 7xxx series alloys. It is
now common to use precracked specimens to assess comparative resistance of alloys to stress-corrosion cracking, since this
type oftest avoids uncertainties associated with crack initiation.
Source: I. J. Polmear, Light Alloys. Edward Arnold Ltd, London, England, and American Society for Metals, Metals Park OH,
1981,p79
12-48. Alloy 7075-T651: Fatigue Life as Related to

Harmonic Generation
FATIGUE LIFE EXPENDED
50
o
30*>
60*
1000
2000
I..LI
0
l:::Ji=? 4
-' - '
<t:a..
::?:::?:
""<t:
~~
2
00
LL.
I..LI::?:
=>""
-'<t:
<t:::c
>0
>2
L58
a.. I..LI
Vl
00
FATIGUE CYCLES
Peak value of normalized second harmonic generation as a function of fatigue life.
Recent experiments on flexural-fatigue specimens (aluminum

alloy 7075-T651) clearly show the potential of harmonic generation for fatigue monitoring. The above chart shows the peak
value of the harmonic generated as a function of fatigue life. At
60% ofthe fatigue life expended, the harmonic had increased by
about a factor of four.
Source: O. Buck and G. A. Alers, "New Techniques for Detection and Monitoring of Fatigue Damage," in Fatigue and Microstructure, American Society for Metals, Metal Park OR, 1979, P 137
367
368
12-49. Alloys 7075-T6 and 7475-T73: Effect of Laser-Shock

Treatment on Fatigue Properties
.25" OPEN
HOLE
0_
.75" 0.0.
LASER SHOCK PROCESSED AREA
2
(26 J/cm / SIDE)
Fatigue Test Specimen Geometry

FATIGUE TEST RESULTS FOR 7075-T6 ALUMINUM
"MAX = 20 KSI NET R = .t
LASER-SHOCKED
CONTROL SPECIMENS
SPECIMEN NO.
NO. OF CYCLES
TO FAILURE
51,500
n,loo
2
3
SPECIMEN NO.
NO. OF CYCLES
TO FAILURE
1
2
473,800
520,700
AVG.
SCATTER
497,250
1.1
385,000
AVG.
SCATTER
169,600
7.5
FATIGUE TEST RESULTS FOR 7475-T73 ALUMINUM

"MAX = 20 KSI NET R= I
LASER-SHOC KED
CONTROL SPECIMENS
SPECIMEN NO.
1
2
3
AVG.
SCATTER
NO. OF CYCLES
TO FAILURE
41,500
74,300
109,300
75,033
2.63
SPECIMEN NO.
NO. OF CYCLES
TO FAILURE
1
2
171,800
266,200
AVG.
SCATTER
218,950
1.5
The fatigue test specimens were 0.25 inch thick by 1.5inches wide and approximately 9.5 inches long, as shown
in the above sketch. The specimen blanks were laser-shock processed, and then the 0.25-inch-diameter hole was
bored through the center of the laser-shock-processed area. The diameter of the laser-shock-processed area is
three times the fastener hole diameter. All of the specimens had machined surfaces of less than 125RMS. All of
these open-hole specimens were fatigue tested to failure at a maximum net section stress of20,000 psi, and an R=
0.1 under constant-amplitude load control. Three control specimens for each material were tested to establish
the typical fatigue life for the material. Two LSP specimens were tested for each material to establish the degree
of improvement due to the laser-shock processing.
The fatigue test results for the 7075-T6 material are summarized in the upper tabulation. The LSP specimens
showed three times better fatigue lives on the average and much less scatter than the unprocessed material. The
results for the 7475-T73 material are summarized in the lower tabulation; these show the same typically large
increases in fatigue life and reduced scatter. It should be noted that the 7075-T6 material shows better fatigue
resistance than the 7475-T73 material, whether or not it is laser-shock processed. This is largely due to the
differences in dislocation/precipitate interactions that result from the T6 and T73 heat treatments. The dislocations appear to shear through the precipitate particles in the T6 condition. The precipitate particles are apparently so strong in the T73 condition that the dislocations just loop around the particles.
Source: William F. Bates, Jr., "Laser Shock Processing of Aluminum Alloys," in Source Book on Applications of the Laser in
Metalworking, Dr. Edward A. Metzbower, Ed., American Society for Metals, Metals Park OH, 1981,pp 256-258
12-50. Alloy 7075-T6: Effect of Laser-Shock Treatment

on Hi-Lok Joints
3/16~ HI-LOKS & TAPER-LOKS
(4 PLACES)
0-
-0
c:::::=:=========$$:Jr====
.072"
Joint Fatigue Test Specimen Geometry

106
.....-- - - - - - - - - - - - - - - --== 0=----.

IZZI
SYD HOLE
~ LASER SHOCKED HOLES
LOAD
CYCLES TO
FAILURE
Iff
"MAX = 25 KSI NEJ
"MAX = 17 KSI NET
"MAX 14 KSI NET
Fatigue Test Results for Laser-Shack-Processed

7075-T6 (Clod) Hi-lok Joints
The full load transfer joint shown in the above sketch was made from
7075-T6 clad aluminum alloy and fatigue tested. The purpose of this test was
to evaluate the fatigue life improvement of laser-shack-processed fastener
holes when the holes are loaded by the Hi-Lokfastener in bearing. A secondary purpose was to find out if the cheaper Hi-Lok fastener system in a
laser-shack-processed hole would show as good a fatigue life as the much
more expensive Taper-Lok fastener system. The above bar chart shows the
test results for three different stress levels. At each stress level, three specimens with standard holes and three specimens with laser-shack-processed
holes were tested. The specimens tested at the l4-ksi stress level showed severe fretting at the intersection of the hole wall with a badly galled area of the
fretted faying surface. All of the fatigue origins occurred at or near the hole
wall corners on the faying surface.
Source: William F. Bates, Jr., "Laser Shock Processing of Aluminum Alloys," in Source Book on Applications of the Laser in
Metalworking, Dr. Edward A. Metzbower, Ed., American Society for Metals, Metals Park OH, 1981, pp 262-263
369
370
12-51. Alloy 7075 (High Purity): Effect of Iron and Silicon on

Cycles to Failure
400
n
I
:::E
u
If)
If)
300
UJ
a:
Iii
o
Z
200
~
Z
a:
UJ
<
100-'--,----r-----.,-----.-----,-10'
10
10'
10'
10'
CYCLES TO FAILURE
Influence of Fe and Si content on SoN curves (R = -1).
The large Fe- and Si-containing inclusions are detrimental to

the fatigue life of smooth specimens, because these inclusions
serve as easy nucleation sites for cracks. Comparing two alloys,
one containing these inclusions (Commercial Purity 7075) and
the other one not (High Purity 7075), shows the improvement in
fatigue life due to the removal of these inclusions (see the above
SoN curves). The alloy termed High Purity 7075 in this figure
still contains Cr and therefore the small Cr-containing inclusions. This is important because the removal of these small inclusions would have the opposite effect on fatigue life.
Source: G. LUtjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
D. E. Embury, Eds. American Society for Metals, Metals Park OH, 1980, p 193
12-52. Alloy X-7075: Effect of Grain Size on Cycles to Failure

o
0..
::r
300
UJ
...::>
::::;
0..
::r
Vl
Vl
200
UJ
a::
Vi
100L~,----------~;----------'-=--------~;-
__
CYCLES TO FAILURE
Influence of grain size on S-N curves (R = -1,

f= 100 Hz) for X-7075 with PFZ (20 hat 160C,
or 320 OF).
The resulting improvement in fatigue life due to the grain size

reduction for this crack nucleation mechanism is shown in the
above S-N curve. Again, the tensile yield stress was equal for
both grain sizes. Also for low-cycle fatigue it was found that
red ueing the grain size of 7XXX series alloys results in increased
fatigue life of smooth specimens in the averaged condition.
Source: G. Lutjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys." in Aluminum Transformation Technology and
D. E. Embury, Eds., American Society for Metals. Metals Park OH. 1980, p 192
371
372
12-53. Alloy X-7075: Effect of Grain Size on Stress-Life Behavior
400
"
c,
::E
300
::::i
a,
::E
<I:
lI)
lI)
0:
200
l-
ll)
CYCLES TO FAILURE
Aluminum alloy X-7075; 24 h at 100 C (212 OF).
The above chart shows the grain-size effect in a stresscontrolled test for a high-purity 7075 alloy (X-7075) aged to contain shearable precipitates. Since the flow stress is determined
by the interaction of dislocations with the coherent precipitates,
the yield stress is approximately the same for both alloys. Opti- .
cal examinations of the specimen surfaces show that cracks nucleate much earlier in specimens with the large grain size. Cracks
nucleated at intense slip bands for both grain sizes.
Source: Edgar A. Starke. Jr., and Gerd Lutjering, "Cyclic Plastic Deformation and Microstructure." in Fatigue and Microstructure.
American Society for Metals. Metals Park OH, 1979. P 225
12-54. Alloy X-7075: Effect of Environment; Air vs Vacuum
-I
10
..
A: 24h 100'C
VACWM
C:48h 180'C
AIR
E
u
-)
10
_,
10
:E
a
'0
-i
10
I06..L..----,--r--..,.....--..--..,.....---.------r----r--..,......
Influence of environment (laboratory air) on fatigue
crack propagation rates for underaged (A) and overaged
(C) condition. X-7075, CT specimens,R = O.I,f= 30 Hz.
A basic correlation between microstructural parameters and

fatigue crack propagation rate can only be determined so clearly
if the tests are performed with the exclusion of any aggressive
environment. To illustrate this point, the above graph shows the
comparison between underaged and overaged microstructure
also for tests performed in laboratory air. The aggressive environment has a much more pronounced effect on the underaged
condition, leading even to an opposite ranking of the alloy conditions. In laboratory air the cracks propagate still along slip
bands at low dajdN rates.
Source: G. LOtjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
D. E. Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 204
373
374
12-55. Alloy X-7075: Effect of Environment on Two Different

Grain Sizes
-2
10
GS 200.lJrn
GS
-J
46.lJrn
10
>-
.!2
-4
E 10
.g
-5
,,/~
3.5NoCI
10
I06.L----.-----.---'-r---..-----r-----.---...----.7
10
15
20
30
40
50
t:. K [MNrn-3I2 )
Influence of environment (3.5% NaCI) on fatigue crack
propagation rates for two different grain sizes. X-7075,
24 h 100 DC, CT specimens,R = 0.1,1= 30 Hz.
The same tendency is observed for the grain size dependence

of crack propagation ifthe tests are carried out in a 3.5% NaCI
solution (note above curves). The influence of environment is
larger for the large grain size. For this highly aggressive environment the cracks propagate at low dal dN rates along grain
boundaries in a complete brittle fashion.
12-56. Alloy X-7075: Effect of Grain-Boundary Ledges on

Cycles to Failure
..-
Q.
LLJ
300 r--~"-T'"""...-r'1""""'-r---'----r-"'T"""T"T"l"T~---'----"--"-"""""""'"
co..
'PO
\
~ 250 I---~ID-----':~_----_--t-----____l
:J
Q.
~
~
~o.
....
"
200t------O-~~--""'~-+__---____l
X-7075
150 t----::------:-~-:-----t-----_l
0---0 If = 0,
8h 160C
--If = 0.5,4h 160C
105
CYCLES
Effect of grain-boundary ledges on the stress-life
behavior of an alloy containing nonshearable
precipitates and PFZ.
One method that may be employed to reduce the slip length in

the PFZ is thermomechanical processing. If enough cold deformation is employed to introduce steps (or "ledges') into the
grain boundaries, the effective slip length within the PFZ is
drastically reduced (similar to a small grain size) with corresponding improvement in resistance to fatigue-crack nucleation. The above chart shows the results of a stress-controlled test
for two high-purity 7075 alloys, one cold-worked 50% to produce grain-boundary steps. The cold work drastically reduced
the incidence of grain-boundary cracking and improved the fatigue life at high stress amplitudes. At low stress amplitudes and
long fatigue lives, crack nucleation occurred at inclusions for
both alloys. This effect is most likely due to stress concentration
at inclusions.
375
376
12-57. Alloys X-7075 and 7075: Effects of Chromium Inclusions on

24 h 100C
-1
VACUUM
10
n
~
u
'"
-)
10
E
E
u
-.
10
"0
-5
10
-6
10
40
50
Influence of Cr-containing inclusions on fatigue crack

propagation rates by comparing aluminum alloys X-7075
and 7075.24 h 100 C, CTspecimens, vacuum,R = 0.1,!
= 30Hz.
As shown above, the small inclusions have a much stronger

influence on fatigue crack propagation because they lower the
reversibility of slip and they crack within the plastic zone ahead
of the crack tip. Furthermore, they normally increase fatigue
crack propagation rates also indirectly by their effect on grain
size and shape.
12-58. Alloy 7475-T6: S-N Diagram for a Superplastic

Fine-Grain Alloy
80
r------------------------------.
o
SPF Conditions:
= As pr......d
A = SPF - highsup.rpl.stic st"inl
o = SPF - low luporpl.stic st"inl
T= 516C
i=21104S-1
70
-+ = No,.ilu"
R =+0.1
60
AS PROCESSED SN CURVE
MAX
STRESS 50
(Ksi)
40
AOC>AO
-...
6.
...
(]I
_-------0-+0-+
3D
20 I-.....I._..I....I..I-I._..................I-.....I._...."""'"-"u.._"'-..........I."""".........._
1~
,~
10'
10'
10'
....."""'"-"U
10'
CYCLES TO FAILURE
SN curves for a superplastic aluminum alloy: fine-grain

7475. All testing was done with smooth specimens.
Tests on fine-grain 7475 alloy have shown improved fatigue

life as superplastic strain is increased, as shown in the above SoN
diagram. An even more dramatic improvement is obtained in
damage tolerance.
Source: C. Bampton, F. McQuilkin and G. Stacher, "Superplastic Forming Applications to Bomber Aircraft ." in Superplastic
Forming. Suphal P. Agrawal. Ed.. American Society for Metals. Metals Park OH. 1985. p 77
377
378
12-59. Alloy 7475: Effect of Alignment of Grain Boundaries on

Cycles to Failure
300
0
,.....
0
Q..
-..J
250
- II
UJ
0
:J
-<9=0.5
...J
Q.
o <9=0
200
16h 160C
6h 160C
~
<{
I/l
I/l
UJ
0:
150
II)
7475
o
I
CYCLES
1.
TO FAILURE
Effect of alignment of grain boundaries-and alignment

plus steps in grain boundaries-on the stress-life behavior
of a 7475 aluminum alloy containing nonshearable precipitates and PFZ.
If the stress axis is parallel or perpendicular to the long grain

dimension, there will be no shear stress parallel to the grain
boundary, and preferential deformation within the PFZ will be
restricted. Grain-boundary alignment is then as effective in restricting deformation in the PFZ as are steps produced by
thermomechanical treatment: this is shown by the stress-life
curves in the above graph.
12-60. Alloy 7475-T6: Superplastic vs Nonsuperplastic, as Related

to Fatigue Crack Growth
1 X10-] r - - - - - - - - - , . - - r - - - - - - - - ,
7475 T-6
FINE GRAIN
~ 1 X 10-'
(.)
.....
7075 T-6 -
(J)
::I:
(.)
1 X 10- 5
<,
<
o
1 X 10- 6
1 X 10- 1
1.----I._.L.I.....L..J..J.._...l----JL...-..JL.I...l.----I._.....L....J...J..J
10
100
1000
K (I-R)M-l (KSI "'lINCH)
This comparison of conventional, coarse-grain, nonsuperplastic aluminum alloy 7075 with superplastic alloy 7475 shows
almost an order-of-magnitude reduction in crack growth for the
superplastic material.
Source: C. Bampton, F. McQuilkin and G. Stacher, "Superplastic. Forming Applications to Bomber Aircraft," in Superplastic
Forming, Suphal P. Agrawal, Ed., American Society for Metals, Metal Park OH, 1985, p 77
379
380
12-61. Alloys X-7075 and 7075: Effect of Chromium-Containing

Inclusions on Cycles to Failure
400
:g
300
I-
:::;
Q.
<l:
200
~ J~----,--;,------------'--r:---""""'----""'-3
10
10'
10
la'
CYCLES TO FAILURE
Influence of Cr-containing inclusions on SoN curves

(R = -1,1= 100 Hz) comparing aluminum alloys x-7075
and 7075, 24 h 100 0 C. (Arrows indicate crack nucleation
visible by LM at u a= 200 MNm -2. j
The above S- N curves compare results obtained from testing

commercial 7075 alloy with the alloy X-7075 which does not
contain Cr. These small inclusions, as in the tensile test, inhibit
the formation of intense slip bands, thus retarding crack nucleation as indicated by arrows on the graph. Due to these small
Cr-containing inclusions, the grain size of the 7075 alloy was
somewhat smaller as compared to that of X-7075, which also
may have contributed to the observed improved fatigue behavior.
Source: G. LUtjering and A. Gysler, "Fatigue and Fracture of Aluminum Alloys," in Aluminum Transformation Technology and
12-62. Aluminum Forging Alloys: Stress Amplitude vs

Reversals to Failure
"t
IJ.
o
400
2.:
7075-T73 squeeze formed

SF1-T61T73 squeeze formed
6082-T6 squeeze formed
6082-T6 extruded bar
SF1-T6/T73
300~-T6
S
~
~200
.~~o
x
In
10 3
10 4
10s
10 6
REVERSALS TO FAILURE (2N,)
SON fatigue data for several squeeze-formed

forging-type aluminum alloys compared with extruded AA 6082-T6.
The above chart presents results from push-pull, about mean

zero, fatigue tests which have been carried out on a servohydraulically controJled machine. The tests have been carried out
on samples cut from actual components, not from separately
made testpieces. The results from conventionaJly extruded AA
6082 (H30) are included for reference: in this case, the data are in
the longitudinal direction, it not being possible to obtain samples of sufficient size from the transverse direction.
The results show good fatigue properties for squeeze-formed
material, which in one case compare favorably with conventionally extruded material. This further substantiates the claim
that squeeze formings in general are comparable with forgings
with respect to mechanical performance.
Source: G. Williams and K. M. Fisher, "Squeeze Forming of Aluminium-Alloy Components," in Production to Near Net Shape:
Source Book, C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals, Metals Park OH, 1983, p 367
381
382
12-63. AI-5Mg-O.5Ag: Effect of Condition on Fatigue Characteristics

180 ~------'--------r------.,----,
0.2% proof Tensile
AI-5%Mg-0.5%Ag
stress IMPa) strength IMPa)
150
ST and quenched
85
Aged 1 day 175C
200
310
175
270
X Aged 70 days 175C
lU
a..
(/] 120
(/]
Q)
(/]
Cl
'16
90
~
.~
oho
~ ~....
0_
~
<i:
&0
L60
10&
87
X 72
7
10
48
~
lOll
Number of cycles
Fatigue (SoN) curves for the alloy AI-SMg-O.SAg
in different conditions.
The fact that microstructure can have a greater influence

upon the fatigue properties of aluminum alloys than the level of
tensile properties has been demonstrated for an AI-Mg alloy
containing a small addition of silver. It is well known that binary
AI-Mg alloys such as AI-5Mg, in which the magnesium is present in solid solution, display a relatively high level of fatigue
strength. The same applies for an AI-5Mg-0.5Ag alloy in the
as-quenched condition, and the above diagram shows that the
endurance limit after lOS cycles is 87 MPa, which approximately equals the 0.2% proof stress. This result is attributed to
the interaction of magnesium atoms with dislocations, which
minimizes formation of coarse slip bands during fatigue. The
silver-containing alloy responds to age hardening at elevated
temperatures due to the formation of a finely dispersed precipitate, and the 0.2% proof stress may be raised to 200 MPa after
aging for one day at 175C (350 OF).
Source: I. J. Polrnear, Light Alloys. Edward Arnold Ltd. London, England, and American Society for Metals, Metals Park OH.
1981, P 42
12-64. AI-Zn-Mg and AI-Zn-Mg-Zr: Effect of Grain Size on

Strain-Life Behavior
100
r----------------...,
a,
" ",
GI
.':::.
.. _- "
........
24h at 150GC
...........
::---'.::: ....
'
...
(\J
-...
e--
...... ,
:.:~~
Q.
......
<.&>
<l
o Failure AI-Zn-Mg
Failure AI-Zn-Mg -Zr

, Crack Initiation AI-Zn-Mg-Zr
10
2Nf
Effect of grain size on the strain-life behavior of an

alloy having nonshearable precipitates plus PFZ.
The AI-Zn-Mg alloy had large grain size; the AIZn-Mg-Zr, small grain size.
The above chart shows Coffin- Manson life plots of two averaged AI-Zn-Mg alloys. The small-grained AI-Zn-Mg-Zr alloy
has a much longer life than does the large-grained AI-Zn-Mg
alloy. The improvement in life is attributed to increasing the
cycles to crack initiation, as indicated in the chart. A convergence is noted for long lives (low plastic-strain amplitudes) for
this strain-controlled test. Since the fine-grained material
hardens more than the other at low strains, the stress to enforce
the applied strain is greater at long lives, and this affects the life
improvement due to the fine grains.
American Society for Metals, Metals Park OH, 1979. P 228
383
384
12-65. AI-Zn-Mg: Strain-Life Curves of a Large-Grained Alloy
4h at 120C
.96h at 150C
10.
c..
wiN
..........
<l
I.
AI-Zn-Mg
0.1
10 1
Strain-life curves of large-grained AI-Zn-Mg alloy having

shearable precipitates when underaged (4 h at 120 e, or
250 OF) and nonshearable precipitates plus PFZ when
overaged (96 h at 150 "C, or 300 OF).
Since the strain localization occurs in a region free of solute, overaging the matrix precipitates or adding dispersoids does not homogenize
the deformation. This is clearly illustrated by comparing the Coffin
Manson life curves of underaged and overaged specimens of largegrained AI-Zn-Mg alloy (see above chart). The tensile yield strength
and strain to fracture are approximately the same for both specimens.
The underaged alloy has shearable precipitates, which results in strain
localization, the formation of intense slip bands, and early crack nucleation under cyclic loading. Overaging was one method described for
homogenizing deformation; however, this method is not effective for
large-grained material. Preferential deformation in the PFZ also leads
to strain localization and results, for this particular case, in the same
fatigue life. Dispersoids distributed throughout the matrix would not
inhibit strain localization in the PFZ for the same reason.
Source: Edgar A. Starke, Jr., and Gerd Liltjering, "Cyclic Plastic Deformation and Microstructure,"in Fatigue and Microstructure,
12-66. Aluminum With a Copper Overlay: Stress Amplitude vs

Cycles to Failure
70
g
Q.
60
~
-b
.......
0
<,
50
+1
c-
<,
~o
<,
o Al
D
<,
_-- -----
Al (Cu)
<,
0.....
40
10
10
10
10
Stress amplitude U a versus number of cycles to failureNj

for AI and AI with a Cu layer. Note the pronounced improvement in the latter at large N j
The fraction of fatigue life spent in crack nucleation, N; / NJ ,

increases with decreasing load amplitude (i.e., at high N's). It
would be expected that the treatment suggested above will produce a great effect in large fatigue life regimes (i.e., under conditions where initiation of fatigue crack is more important than its
propagation). The above graph shows this phenomenon in the
case of pure aluminum and aluminum with a copper surface
layer.
Source: Marc Andre Meyers and Krishan Kumar Chawla. "Mechanical Metallurgy: Principles and Applications," Prentice-Hall,
Inc .. Englewood Cliffs NJ, 1984, P 707
385
386
12-67. P/M Alloys 7090 and 7091 vs Extruded 2024

60
400
l>
50
'in
"'"
40
30
..
20
E
OJ
E
'x
::;;;
7091T7E69 (2 lots)
7090T7E71
Open symbols-longitudinal
Solid symbols-long transverse
60
O~;"
300
- - I ndicates did not fail

/to:D
10
o,
::;;;
Notch tip radius = 0.013 in.

KT = 3.0 4
~-
200 E
OJ
E
..
'x
::;;;
Band for 2024T351

rod and bar
100
Stress ratio R " 0

Ambient air
Ol-_ _-'-_ _--'--_ _--'-_ _-'-_ _--'-_ _--'-_ _--'_ _--.J

10
10'0
10
10
10'
10'
107
10'
Cycles
S-N diagram that provides a comparison of notched axial

fatigue strength for P 1M alloy 7090 and 7091 extrusions
vs 11M alloy 2024-T351 rod and bar.
The notched axial fatigue strengths of alloys 7090 and 7091

are 35 to 40% higher than those of alloys 7050, 7075 and 2024 (an
11M alloy often selected for its resistance to fatigue) at one million or more cycles.
Source: Robert H. Graham, "Wrought Aluminum PIM Alloys," in Powder Metallurgy-Applications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 240
12-68. P1M Alloys 7090 and 7091 vs 11M 7050 and 7075 Products
60
r-----.-----.----.-------.----.-------.---~-------,400
A 7091-T7E69 (2 lots)
7090-T7E71
Open symbols-longitudinal
Solid symbols-long transverse
o
50
40
30
60
O~"
300 ..
Indicates did not fail

Notch tip radius
KT = 3.0
a.
0.013 in.
2:
~~
200
:>
E
:>
..
'x
20
10
2:
Band for 7050 and

7075 products
Stress ratio R = 0
Ambient air'
100
O'----:----'-,-------'-c-----'-:c-----L,---'--:---'----:----'-,------'
10 7
103
10'
10"
10
10 9
10 2
10 10
Cycles
S-N diagram that compares notched axial fatigue strength

for P 1M alloy 7090 and 7091 extrusions vs 11M 7050 and
7075 products.
The notched axial fatigue strengths of alloys 7090 and 7091

are 35 to 40% higher than those of alloys 7050,7075 and 2024 (an
11M alloy often selected for its resistance to fatigue) at one million or more cycles, as shown above.
Source: Robert H. Graham. "Wrought Aluminum PIM Alloys." in Powder Metallurgy-Applications. Advantages and Limitations. Erhard Klar, Ed. American Society for Metals, Metals Park OH, 1983. p 240
387
388
12-69. P1M Aluminum Alloys: Typical Fatigue Behavior
971B-in. R
o
o
Forged
'u;
40
Forged 601AB-T6
?:
-=fL-i-----3-"1!I
201 AB-T6
0.300-in. diam
c.
201AB-T6
202AB_T2~
.~ 20
x
2''"
60 1AB-T2'>.
601A~
_ _..::::===_=====a=_=_=_=_=_-1
OL------''------''-----'----'-----'-----'----'
10'
10'
10
Cycles to failure, N
10'
Typical fatigue behavior of alloys 601AB, 201AB and 202AB.
Fatigue is an important design consideration for P / M parts

subject to dynamic stresses. The above S-N diagram shows typical fatigue behavior of specimens of alloys 6OIAB, 20lAB and
202AB in the T2 (as-cold-formed after sintering) and/ or T6
tempers.
Source: John D. Generous and Wayne C. Montgomery, "Aluminum PI M-Properties and Applications." in Powder MetallurgyApplications. Advantages and Limitations, Erhard Klar, Ed.. American Society for Metals, Metals Park OH, 1983, p 214
12-70. P1M Aluminum Alloys: Comparison With Specimens Made

by Ingot Metallurgy
500
~ 1.9;mm (0.075
E--t--]- E= f-- t
tf--+-1=
7.62 mm . _ _ 8.39 mm
(0.300 in.)
(0.3.30 ln.I
diarn.
diarn.
400
'"
::E
..
300
1;;
::>
-1.0
in.)
60
12.19 mm (0.480 in.)
Notch tip radius:

e ~ 1.254 mm (0.01 in.)
K, ~ 3
-,
0..
Ii
70
60 sharp' V
251 mm (9'10 in.)
50
40
.'"
~
1;;
::>
'x
'x
'"
::E
'"
::E
Smooth
200
30
a---
100
..
Solid lines represent

bands for 707517352
!three lots), longitudinal direction
Dashed lines represent
bands for 2014-T61 (nine lots smooth.
five lots notch I. longitudinal direction
0
10'
-- -
10'
10'
---
10'
20
10
-~
10'
0
10'
Cycles
Rotating-beam fatigue strength for die forgings of P 1M alloy X709I-T7E76 and ingot
metallurgy alloys 7075-T7352 and 2014-T6I. For P/M X709I-T7E76: 0 ,smooth,
transverse direction; ,notched, transverse direction; - denotes test specimen did
not fail in number of cycles indicated.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society ForMetals, Metals Park OH, 1984,p 469
389
390
12-71. P 1M Aluminum Alloys: Comparison With Forged 7175 for

Cycles to Failure
.500
t."
L10."" J,~.L
mm
10.,)'
~t~
400
70
in.]
60
Notch tip radius:
e = 0.33 mm (0.013 in.)
=3
K,
ul
50
Ul
e
1il
300
E
E
:>
'x
co
::E
-"
ul
Ul
e
1il
'iii
40
'x
co
m:a
0
200
-.we
7175-T73~ ~
It
30
o ~
10'
r r
103
10'
::E
20
Band for
100 _forgings (six lots)
Stress ratio: R = 0.0
:>
10
o
10'
10'
10'
10'
10'
Cycles
Comparison of axial-stress notch fatigue strength of P 1Malloy X7091-T7E69 die forgings and ingot metallurgy alloy 7175-T736 die forgings. 0 , longitudinal direction,
one lot; , short transverse direction, two lots; - denotes test specimen did not fail in
number of cycles indicated. Stress ratio: R = 0.1.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,P 469
12-72. Various Aluminum Alloys: Comparison of Grades for

Corrosion-Fatigue Crack Growth Rates; Air vs Salt Water
'(f3_-------------------"'"
specimens: SEN , GNP, DCB , CT
R 0.0 - 0.1 , 0.1 - 30 Hz
ambient
temperature
11 AI - alloys in salt water:

2048-T851
2219 - T 87
2618 -T6
5456-H117
5456-H321
6061 -T651
~
2cro - T651
2024 - T 3
2024-T~
2048 - T851
2219 - T851
2219 - T87
20
cyclic stress
18 AI- alloys in air:
2618 - T 6
5456 -H117
5456-H321
6061 - T651
7'005 - T63
7039 - T6X31
30
7005 - T63
7050 -T736
7fJ15 - T651
7175 - T736
7475 -T651
40
7050 - T 736
7075 - T 6
7079-T651
7106 - T 63
7175 - T 736
7475 - T 651
50
60
iltensity range, 6K, [MN 'm-2]
Comparison of scatterbands of corrosion-fatigue

crack growth rates and fatigue crack growth rates
of many commercial aluminum alloys.
Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material," in Aluminum Transformation Technology and Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OH, 1980, P 615
391
392
12-73. Various Aluminum Alloys: Comparison of Grades for

Corrosion-Fatigue Crack Growth Rates in Salt Water
163~------------------,
7475 - T651
7005 -T63
2618 - T6
7175 - T 736
7075 - T 651
2219 - T 87
5456 - H 321
,----,
~
~
~
10- 5
'------'
~~
6061 -T651
2048 - T851
-6
10
7050 - T736
5456 - H 117
10- 7
Q)
~ typical experimental scatter
::J
.21
....
....ro
specimens: CNP I DCB I CT
environment: salt water
ambient temperature
R 0,0-0,1 I 0,1 - 30 Hz
-10
10 0,......--+---+----+----+----+---.....
10
20
30
40
50
60
cyclic stress intensity range,
~K , [MN. m-~]
Corrosion-fatigue crack growth rates in salt water

for aluminum alloys exceed the scatterband.
As shown in the above graph, curves for growth rate are

somewhat higher than the air-test scatterband, but at very low
and very high stress-intensity ranges, no significant difference
between fatigue and corrosion fatigue crack growth rates is
observed.
Source: Markus O. Speidel, "Aluminum as a Corrosion Resistant Material, "in Aluminum Transformation Technology and Applications (Proceedings of the International Symposium at Puerto Madryn, Chubut, Argentina), C. A. Pampillo, H. Biloni and D. E.
Embury, Eds., American Society for Metals, Metals Park OH, 1980, p 614
12-74. Various Aluminum Alloys: Wrought vs Cast, and Influence of

Casting Method on Fatigue Life
500
400
300
250
'"
~
200
<;;
150
rf.
t:
en
~
~
201---"':
18t----
161----+~
141----+----,.:-'
~
iU
":;'l
1.
II)
100
121----'--+---'--~
75
1 0 1 - - - - + - - ' -......- +-..........:
8!---'---'-f--'-----if--"""'""-,..-t----'
71-----+--.-:.+------1------+--,,-:~
50
61-----+---+--...--:'":-...;,......1----'---+..,.....-..:.,......+---~
106
lOS
109
Life N (cycles (log))
Representative S-N curves for various aluminum alloys are

shown in the above graph. Note the absence ofa sharply defined
"knee" and true endurance limit. This is typical of nonferrous
metals. In the absence of an endurance limit, the fatigue strength
at 108 or 5 X 108 cycles is often used. (To give a "feel" for the time
required to accumulate this many cycles, an automobile would
typically travel nearly 400,000 miles before anyone of its cylinders fired 5 X 108 times.) As is true for most metals and alloys,
the wrought versions of aluminum alloys have greater fatigue
strength than the cast (see graph). It will also be noted in the
graph that there is an overlapping of fatigue strength for the
sand and permanent mold casting methods (same alloy).
Source: Robert C. Juvinall, Fundamentals of Machine Component Design, John Wiley & Sons, New York NY, 1983, p 207
393
394
12-75. Aluminum Casting Alloy AL-195: Interrelationship of Fatigue

Properties With Degree of Porosity
26
...
24
... ...
, , 1\
,~
"
4
16, [,>..
22
,,1,- 3,
... ...
5)
~ 1-1',
S-N CURVES
CASTING ALLOY AL - 195
DEGREE
FATIGUE* DEGREE
FATIGUE*
POROSITY STRENGTH POROSITY erRENGTH
- - 5 - - 9,000
;t
2
10,500
6
8,150
rr;ooo
1'>,
,""1\
i"
,~
I,
" ...
x 18
"
f' ~"
41- I - -
- .
' ... ,
8./
...
...
~I'
....
"of-.
f'
'"
I"'-\l:i:
I'
"o
.......
8
} DEGREE
:-:-- POROSITY
10 Q
,,5
10
. I-
",I'. ,
1'1
10
"I",
I""' l ......
"o
12
A~~~
-- f - -
.~
....
...
-- -
,~
r-,
_~
"'I'
1,000
.... REPRESENTATION OF
NO FAILURE
'I\.
<,
,'\..
~ .......
9,000
9,000
* AT 10 000 000 CYCLES
\,
,1"\
20
3
4
1)
IV
,,"'.
;.. "~ r-;

...
<,
~
..
I-
...
r-,
r>-
J'.... ..... ~ ..... I-
.. .. .. r-...... .....-r--
10,,6
101
Interrelationship offatigue properties with degree of porosity for AL-195

casting alloy.
Source: N. E. Promisel, "Evaluation of Non-ferrous Materials," in Materials Evaluation in Relation to Component Behavior
(Proceedings of the Third Sagamore Ordnance Materials Research Conference), Syracuse University Research Institute, Syracuse
NY, 1956, P 65
12-76. Aluminum Casting Alloy LM25-T6: Squeeze Formed vs Chill

Cast; Effect on Reversals to Failure
400
~300
w
o
=>
l-
ii 200
2
(/)
~ 100
LM 25-T6 chill cast
a::
In
O'-:--_--L
2
10
1~
--'4
'-::-_ _--'-
1~
--'---'
1~
REVERSALS TO FAILURE(2N,)
S-N curves for aluminum casting alloy LM25-T6; chill
cast versus squeeze formed.
Fatigue tests have been carried out with LM25 samples,

which were cut from a bracket component. A servohydraulically
controlled fatigue machine was used to execute push-pull tests
about mean zero. The results are presented in the above chart,
which includes for reference the results of similar tests carried
out on conventionally cast LM25. It can be seen that a significant improvement in the fatigue performance has been achieved
by squeeze forming this type of alloy.
Source: G. Williams and K. M. Fisher, "Squeeze Forming of Aluminium-Alloy Components," in Production to Near Net Shape:
Source Book, C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals, Metals Park OH, 1983, p 367
395
396
13-1. Copper: Effect of Air and Water Vapor on Cycles to Failure

5.6
<,
s:
u
c
...5.4
Vi
Q;
~5.2
'"
c
--
Ol
C
<0
a:
"E
4.6
OJ
(J)
~ -,
~
OJ
---
l'
'~
I--
::f 5.0
e
Vi
'0 4 .8
~"
<,
......
Q(~I'("
1110
~fiedAir ....
_______ T-D
r-...
I'
~
----
If) 0
.... PlJr i f iltr)fJ
I---
II
10
- --
~-- 1---
'ea"
--I" "~~.
InAir"_
4.4 6
10
.... iA
\.\~I
\
---
~
5.10
Endurance, Cycles to Fracture (Log. Scale]

- - + Indicates Specimen Unbroken
The effect of air and water vapor on the fatigue life of annealed copper.
The effect of atmospheric oxygen on fatigue life of copper has

been investigated; oxygen and water vapor reduce fatigue life in
copper. Alternate static exposure to air and dynamic exposure
to vacuum do not affect fatigue life, and SoN curves diverge as
applied stresses are reduced (see graph). Based on these experiments, the investigators concluded that:
Fatigue cracks form early, because the majority of life is
concerned with crack propagation (environment has little
or no effect on nucleation and initial growth).
2 Oxygen and water vapor are the primary damaging constituents in air (water vapor alone is effective).
3 Oxygen must be a gas (preoxidation or intermittent exposure is not effective).
Source: D. J. Duquette. "Environmental Effects I: General Fatigue Resistance and Crack Nucleation in Metals and Alloys," in
13-2. Copper: Applied Plastic-Strain Amplitude vs Fatigue Life
Cu
prestrain 20"1.
annealed
"prestrain 40/.
_-
A-
Applied plastic-strain amplitude versus fatiguelife curves for copper at long life.
Helgeland was the first to observe and claim a fatigue limit for
copper (actually the plateau stress, although it was not recognized as such at the time). Unfortunately, his results were apparently contradicted by those of Kettunen, who observed failures
at stresses down to 17.7 MPa. This difficulty was resolved by
Laird, who showed that Lukas and Klesnil's long-life CoffinManson plots showed failures to occur only down to the plasticstrain fatigue limit; at lower strains, no failures were observed in
the testing time available (see above chart). However, Lukas et
al. also carried out stress-cycling tests, in which they monitored
the plastic strain. Specimens that had been stress-cycled yielded
a plot of saturation plastic-strain amplitude versus life, where
failures occurred at strains as low as 10-5 . The difference between these tests is that in strain cycling, the stress is low in the
initial cycles and increases to saturation, whereas in stress
cycling, full application of the load in the first cycle causes a
large strain in a soft material. This initial large strain creates the
PSB cell structures, which would not otherwise form in a
constant-strain test. Since Kettunen applied the full load to his
specimens, failures were observed at stresses below that of the
plateau. Helgeland, on the other hand, although he was stress
cycling, imposed a low stress at the start of his tests and increased it gradually to the chosen value.
Source: Campbell Laird. "Mechanisms and Theories of Fatigue," in Fatigue and Microstructure, American Society for Metals.
Metals Park OH, 1979, P 195
397
398
13-3. Copper Alloy C11 000 (ETP Wire): Effect of Temperature on

Fatigue Strength
200 r - - - - - - - - - - - - - - - , - - - - - - - - - - - - - - - ,
26
'"
Cl.
::E
22 ]1
150
s:."
...en
s:."
'C,
18
1;;
Q)
::J
::J
en
en
'+>
'"
u.
~
'+>
100
14
'"
u.
10
Stress cycles
Rotating-beam fatigue strength of electrolytic tough

pitch copper, CllOOO wire, 2 mm (0.08 in.) diam, H80
temper when tested at various temperatures.
Source: Metals Handbook, 9th Edition. Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
.
13-4. Copper Alloy C26000 (Cartridge Brass): Influence of Grain

Size and Cold Work on Cycles to Failure
60
50
'"
<,
a.
0
0
0
40
I'--~
....0
::l
><
II!,,,)"
+~
20
..... 1'0- ...
.........
r-,
Q)
l J)
"0
<,
........
<fl<fl
.... 30
<,
I'
~~~
...e
~ a
t;.1--
~?"
I"'--.:!""...
1'; I-
Q)
--
f_
:: ..
r-rx_
IJ..
-i
-+
10
10
107
Cycles for Failure ,N
Legend
o Group G (60% Cold
I:>. Group D (40% Cold
o Group A (20% Cold
Anneal I Grain Size
... Anneal 3 Grain Size
x Anneal 4 Grain Size
+ Anneal 5 Grain Size
Drown)
Drawn)
Drawn)
0.0/2 mm.
0.026 mm.
0.051 mm.
0.131 mm.
Influence of grain size and cold work on fatigue

strength of copper alloy C26000 (cartridge brass).
Changes in grain size and in degree of cold work which result

in increased tensile strength or hardness usually result in improved fatigue strength. The above S-N curves illustrate this and
indicate the influence of grain size and cold work on the fatigue
strength of alpha brass. All specimens were prepared from the
same heat and, therefore, had the same nominal composition.
Source: George M. Sinclair, "Some Metallurgical Aspects of Fatigue.vin Fatigue-An Interdisciplinary Approach, JohnJ. Burke,
Norman L. Reed and Volker Weiss, Eds., Syracuse University Press, Syracuse NY, 1964, P 69
399
400
13-5. Copper Alloy C83600 (Leaded Red Brass): S-N Curves;

Scatter Band
180 ~------..---------r--------''-------~_-----~
25
160 f---~~~-+-------+------+-------+--------l
rf
140 f------~
20 ]
:::E
~
~
.J::.'
1;;
1201--------+------'
1;;
Ql
:::l
Ql
'"
:::l
.~
'"
.~
u.
100 f - - - - - - - + - - - - - - - + -
------+-------1
80 f-------+--------+-------I-----=
0-
15
10
Stress cycles
S-N curves (scatter band) for copper alloy C83600 (leaded red brass).
u.
13-6. Copper Alloy C86500 (Manganese Bronze): S-N Curves;

Scatter Band
400
If
300
:E
~c
e
t:
200
Cll
;:)
.~
'"
u.
100
- 50
<ttIOo. 0
- 40
~~
..................
--
.r."
",
30
~~
n
t:
Cll
- 20 .;::;c:
u.
- 10 '"
;:)

S-N curves (scatter band) for copper alloy C86500 (manganese bronze). All testing was performed at room
temperature.
Source: Metals Handbook, 9th Edition, Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, ArnericanSociety
401
402
13-7. Copper Alloys C87500 and C87800 (Silicon Brasses): S-N

Curves; Scatter Band
300
40
D-
.,;
Ii
1;;
Cl
..
c:
';;
...
Ql
<{
';;;
~
250
30
200
.....~
Cl
';;
E
Ql
150
100
104
20
105
106
107
108
109
Stress cycles
S-N curves (scatter band) for copper alloys C87500 and
C87800 (silicon brasses) tested at room temperature.
13-8. Copper Alloy C92200 (Navy "M" Bronze): S-N Curves;

Scatter Band
'"
200
0-
25
::i:
Ii 150
20
'"
Ci:
...~..
Cl
Cl
...'"E
'iii
.>t
15
100
50
104
105
106
107
108
sc
.....
'"
10 Ci:'"
109
S-N curves (scatter band) for copper alloy C92200 (Navy

"M" Bronze, or steam bronze) tested at room temperature.
403
404
13-9. Copper Alloy C93700 (High-Leaded Tin Bronze): S-N Curves;

Scatter Band
180
25
0~
't!'
160
140
20
Cl
.g
E
l!l
<i:
120
~_----=l15
100
80
104
105
S-N curves (scatter band) for copper alloy C93700 (highleaded tin bronze) tested at room temperature.
Source: Metals Handbook. 9th Edition. Volume 2, Properties and Selection: Nonferrous Alloys and Pure Metals, American Society
for Metals, Metals Park OH. 1979, P 426
405
13-10. Copper Alloy No. 192: Effect of Salt Spray on Tubes

liAs Received" Cycles to Failure
Tube Oia. 3/16-io. 1/4-io.
5/16-io.
Cycles to Failure After 180-0ays Exposure to Salt Sorcv

3/8-io.
Tube Die, 3/16-io. 1/4-io.
5/16-io.
3/8-io.
I----+I---J----o
0
-r----
-~
I
:0
0
0
0
0
10
10
'Q,
a
0
0
..!
uc-,
.....
M
0
Results of fatigue tests on copper alloy tubes before and after salt spray exposure.
The tubes made from the copper alloy failed in the range of
105 to 106 cycles. After exposure for 180 days to salt spray, the
fatigue performance level was not lowered (see plot above).
Brazed steel tubes, prior to salt exposure, failed in the same
test in the range of 105 to 107 cycles. After 30-days exposure to
salt spray, the resistance to fatigue was 105 to 106 cycles. After
90-days exposure, the steel tubes showed no fatigue strength in
this particular test.
Source: Donald K. Miner, "An Effective Solution to the Problem of Hydraulic Brake Line Corrosion," in Source Book on Copper
and Copper Alloys, American Society for Metals. Metals Park OH, 1979, p 356
406
13-11. Copper Alloy 955: Goodman-Type Diagram
1.5 UTS: 152025 ksi
1----+----+----+-----t----2F------i
150
120
UTS: 101.5 ksi
~
ili
ili
90
iii
.....
""
60
30 I-:JIC---I
..,.<........,h~---fI+_-I+_---1t+_-H__-I::l=-_H-+++-_I_If--I
.3
G _----'
30 '--_
30
---'-
-'-
60
90
-L.
120
--'-_ _- - - '
150
180
Goodman-type diagram (after Creech) for annealed copper alloy 955.
Two variable-speed, flat-plate testing machines of the fixeddeflection type were used for the test work. These machines have
a speed range of 750-2000 cpm with a maximum deflection of I
in. The yield, ultimate tensile strength and 1.5times the ultimate
are plotted in the graph above. The fatigue limit at zero mean
stress was determined and was found to be 22.0 ksi.
Source: J. M. Cieslewicz. "A Modified Goodman Diagram to Predict the Fatigue Limits of Copper Alloy 955." in Source Book on
Copper and Copper Alloys. American Society for Metals. Metals Park OH, 1979, p 40
14-1. Magnesium Casting Alloy QE22A-T6: Effects of Notches and

Testing Temperature
150
If.
:2
125
+1
100
t:
75
E
::J
E
50
..
'x
:2
-.
\.
20
<,
/Unnotched
~
+1
"",-...............
Unolched
/(K t ~ 2)
~.
E
E
::J
Vnotched
/IKt = 31
'x
~
25
Cycles of stress
150
If.
125
:2
+1
~.
100
75
E
::J
'x
~
50
- 20
<,
<;
<,......
/20 oC (68FI- 16 ~
.,
i!
200C 1392F)_ 12 :;;
E
~25~ c (482 of)
::J
E
'x
~
25
Cycles of stress
S-N curves for magnesium alloy sand castings, QE22AT6, showing effects of type of notches (upper graph) and
testing temperature (lower graph). Rotating-beam (Wohler) tests. Machine speed was 2960 Hz.
407
408
14-2. Magnesium Casting Alloy QH21A-T6: S-N Curves; Effects of

Notches and Testing Temperature
25
175
150
rr.:::;;
125
...
100
1;;
E
::l
E
'K
:::;;
20 c (68 IF)
-, r-,
20
............... '-- Unnolched

"-
<,
75
I'-...
50
]
15
--
Unolched K t=2110
...
E
::l
E
'K
:::;;
25
106
Cycles of stress
175
25
Unnolch.d
150
rr.:::;;
+~ 100
~
E
20
125
75
'I\.
"'---
-,
20C (68 OF)

15
~50 c 1480 OF)
I"--.
10
01
g
E
::l
E
'K
:::;;
50
25
Cycles of stress
S-N curves for magnesium alloy castings, QH21A-T6,

showing effects of notches (upper graph) and testing
temperature (lower graph). Rotating-beam (Wohler)
tests; machine speed 2960 Hz.
409
14-3. Mg-AI-Zn Casting Alloys: Effects of Surface Conditions on

Fatigue Properties
Meanstress. ksi
Meanstress,ksi
20
20
20
.16
.16
~
Ii
:t12
-l---I----l
e
~
."
!i
4-----11-----l
t---+--+-----j---t---+--....=j4
100
!II
ell
f---t---+---+----'=-+--+--, '4
125
125
Mean stress,MPa
M&an stress,MPa
Meanstress,ksl
Mean stress, ksi
10
15
20
.150,------,,--,-----,,---y-,-----...---,---r-.
20
10
'150 r ---,----i---.---'T...---.:n--r--T--.
O
20
20
Cast plates
'1251---+---+-- 107 cycles. loaded in bending
.16
"6
Xi
~-
.12
tl2
---f----f----;
--+--1---;
I---t---+----f---""'---+-...:....-=J '4
!II
1-----J-=...."""-+-----1F"'-...;...f---+--..., 4
125
Mean stress,MPa
Mean stress,MPa
Effect of type of surface on fatigue properties of cast Mg-AI-Zn alloys.
Machining improves fatigue properties of castings, as shown in the above

curves. Small radii, notches or fretting corrosion are more likely to reduce
fatigue life than are minor variations in composition or heat treatment.
'"
.~
.
~
410
15-1. Molybdenum: Fatigue Limit Ratio vs Temperature

0.9
0.8
0.7
~r\
';:;
a::'"
0.6
0.5
0.4
100
200
300
400
500
600
Temperature,OC
Ratio of the fatigue limit of molybdenum at various
temperatures to its tensile strength at the same temperature.
16-1. Tin-Lead Soldering Alloy: S-N Data for Soldered Joints

5000
4000
I'~
......
~~
A
'iii
3000
--....r- ....
a.
ui
'"
~
in
2000
.......
:--.... ........
~ :--....
~
~
"'r-..... ~
-- r- --...........,
I~
<,
4 6 8
......
f4.. ......
'Speed of testing I .... ~/min I I

2
. . 'N
.........
~~
1000
.....
102
Cycles to failure
..... ...........
4 6 8
10 3
':).04
0.2
The fatigue strength of soldered joints is a complex and difficult subject to examine. Because solder alloys are strain-rate
sensitive and have large elongation capabilities, the performance of fatigue tests under constant stress causes progressive
and rapid relaxation of the joint, and conversely, tests under
constant strain do not reflect a practical application situation.
The influence of the rate of stress cycling in terms of rate of
straining on the fatigue life of copper soldered joints with
60%Sn-40%Pb alloy is presented in the above graph.
Source: Metals Handbook, 9th Edition. Volume 6, Welding, Brazing, and Soldering, American Society for Metals, Metals Park
OH, 1983, P 1095
411
412
16-2. Babbitt: Variation of Bearing Life With Babbitt Thickness

Babbitt thickness, in.
0.030
0.040
o Bimetal
I
Trimetal
~
OJ
:::l
......
..
200
Cl
',J
OJ
.~
Q;
100
a:
0.25
0.50
0.75
1.00
Babbitt thickness, mm
Variation of bearing life with babbitt thickness for

lead or tin babbitt bearings. Bearing load was 14
MPa (2000 psi) for all tests.
One of the most useful concepts in bearing-material design

came with the recognition that the effective load capacities and
fatigue strengths of lead and tin alloys were sharply increased
when these alloys were used as thin layers intimately bonded to
strong bearing backs of bronze or steel. Use is made of this principle (see graph above), in both two-layer and three-layer constructions, in which the surface layer is composed of a lead or tin
alloy, usually no more than 0.13 mm (0.005 in.) thick.
Source: Metals Handbook, 9th Edition, Volume 3. Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals, American Society [or Metals, Metals Park OH, 1980, P 806
16-3. SAE 12 Bearing Alloy: Effect of Temperature on Fatigue Life

Bearing temperature, F
300
en
.;;
200
300
r---,c-r.----~-----,
200 I------'\-t---+------t
.:?
'"
>
.;;
'" 100
Gi
I----~r_--t-----f
II:
OL-_ _L--_ _L - - _ - - - - l
200
100
150
50
Bearing temperature,
Varlanon of bearing life with temperature for

SAE12 bimetal bearings.
The alloy lining was 0.05 to 0.13 mm (0.002 to 0.005 in.) thick,
on steel backing. Bearing load: 14 MPa (2000 psi). As indicated,
operating temperature markedly influences fatigue life.
413
414
17-1. Unalloyed Titanium, Grade 3: S-N Curves for Annealed vs Cold

Rolled
500
"'
400
a..
:2
'" 300
~'"
'"
E
6,
- 70
r-;
.~~
- 60
Cold rolled
.-. .... -
Annealed
. 200
)(
"'
:2
100
.;;;
50
.><
::i
40 ~
'"
E
J
30 E
'x
20 :2"'
10
SON curve for unalloyed grade 3 titanium. Data were obtained by rotating-beam testing at room temperature.
Source: Metals Handbook, 9th Edition. Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
17-2. Unalloyed Titanium, Grade 4: S-N Curves for Three Testing

Ternperatu res
600
500
'"
Q.
400
'"
'"
E
::>
.Sx
'"
300
-------
- 80
_40C (-40 F)
- 60
20C (6SoF)
- 50
.;;;
..><
:i
E
::>
I
315C (600 of)
- 40 .Sx
30
200
100
70
'"
20
104
S-N curves for unalloyed titanium, grade 4, at subzero, room, and elevated temperatures. Data were obtained by rotating-beam testing of unnotched, polished specimens
machined from annealed bar stock.
Metals, American Society for Metals, Metals Park OH. 1980.0378
415
416
17-3. Ti-24V and Ti-32V: Stress Amplitude vs Cycles to Failure
STRESS
CI
a..
::lE
CONTROLLED
600
ILl
0
::>
I-
:J
a..
Ti-24%V, A.Q.
::lE
<l:
400
lI)
lI)
ILl
It:
l-
ll)
200
10
CYCLES
TO
FAILURE
Stress-life curves of two Ti-V alloys that undergo cyclic hardening (Ti- 24%V) and
cyclic softening (Ti-32%V).
Cyclic-response curves indicate that the Ti-24%V alloy undergoes extensive cyclic
hardening, whereas Ti-32%V undergoes cyclic softening, as indicated above. Hardening is caused by incomplete reversibility of twinning.
American Society for Metals, Metals Park OH. 1979, p 236
17-4. Ti-5AI-2.5Sn: Effects of Notches and Types of Surface Finish
...
800
a::2: 600
~
~l>
~o
t: 400
rt:-
:I
e~
60 t:
Ground
Hand finished
.>I.
80
100 .;;;
Shot peened
40
.
200
)(
';(
...
20 ~
::2:
8
10
Lifetime, stress cycles
...
a::2:
l:/"
e
t:
800
Ultrasonic machined
100
Slab milled
600
80
Chem milled and annealed
Chem mille
400
E
E 200
';(
60
40
:I
::2:
0
104
...
105
..
106
107
800
100 ]
....... 600
t: 400
E
:I
...
)(
::2:
20 ::2:...
0
108
a::2:
E
:I
E
';(
...
,:;
...
......
~
.>I.
200
~are
- 80
notch (K r = 2.4)
~~
l>Al>A
V-notch (K r = 3.2)
'I
16
~
~l>l>
60
40
E
E
:I
'x
- 20 ::2:...
108
SON curves for annealed titanium alloy Ti-5AI-2.5Sn (rotating-beam

tests). Top and center graphs show fatigue strength for different types of
surface finish. Bottom graph shows fatigue strength as affected by type
of notch.
Melals, American Society for Metals, Metals Park OH, 1980, p 382
417
418
17-5. Ti-SAI-2.SSn and Ti-6AI-4V: Fatigue Crack Growth Rates

Stress intensity factor range, /iK, ksi Viii.
5
10
10-4
Ti-6AI-4V
NI24 to -269 0C - - ---f

c:l 10-31---- (75 to - 452 of) ----=",...j~t;f--+_--__l
0C(75 0F)
~
{24
E
ELI -196C (-32F)
E
-269C (-452 of)
~'
NI
ELI
24 to -269C
(75 to -452 of)
b 10-41-----+--~~--I~----4-----=1
Cl
OJ
.,
:J
.Cl
'"
u.
Stress intensity factor range, s, MPa
.,;m
Fatigue crack growth rates for Ti-5AI-2.5Sn and Ti-6AI-4V.
Data on fatigue crack growth rates for Ti-5AI-2.5Sn and Ti6AI-4V alloys are plotted above. These data indicate that the
exposure temperature has no effect on the fatigue crack growth
rates for Ti-5AI-2.5Sn and Ti-6AI-4V(NI). However, over part
of the 11K range, the fatigue crack growth rates for Ti-6Al4V(ELI) are higher at cryogenic temperatures than at room
temperature at the same 11Kvalues.
17-6. Ti-6AI-6V-2Sn: Effects of Machining and Grinding

Ti-bAI-bV-lSn (STA, 42 Rcl
Surface
R aug h ne s s , AA
SURFACE
GRIND + PEEN
Ah ...
b5
43
44
70
lO
20
END MILL
PERIPHERAL CUT
Gentle
Abusive
150
J67
67
14
14
73
28
39
12
II
145
I 45
I",
Off- Sta nda rd
147
'".
85
I
20
48
120
~5
iOff-S'.n""."
43
55
)83
,v.
Gentle
Abusive
ECM +
PEEN
Gentle
HAND
GRIND
ECM
Gentle
Cony.
Abusive
SURFACE
GRIND
40
60
80
ENDURANCE LIMIT, KSI
Bar chart presentation showing effects of various machining and grinding operations
on fatigue characteristics of titanium alloy Ti-6AI-6V-2Sn.
Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 355
419
420
17-7. Ti-6AI-6V-2Sn (HIP): S-N Curves for Titanium Alloy Powder

Consolidated by HIP
iii 120
ANNEALED PLATE
(MINI
"
vi
en
w
a:
l-
ll)
10('
:2
:>
:2
X
<{
:2
80
60 -'--
---,
"T"
"T'"
--,
10
NUM8ER OF CYCLES
OJ
HIP RUN 1. AS MACHINED (SPEC W "M" SU8SCRIPT WAS EVALUATED 8Y METALLOGRAPHY)
HIP RUN 2. VAC ANN AT 1300FI24 HR
mHIP RUN 2. VAC ANN AT 1300FI2 HR

rn HIP RUN 2. VAC ANN AT 1300FI16 HR
iii
HIP RUN 4. VAC ANN AT 1300"FI24 HR
S-N curves showing endurance limits for titanium alloy powder consolidated by HIP at
900C (1650 OF).
Note that most data points obtained in this phase fell within the representative data
band for annealed forgings. In the specimen designated with an "M"subscript, low fatigue
endurance was apparently associated with failure initiation at an inclusion, This shows
that a clean powder is required for parts that are fatigue-critical and must operate with the
equivalent of fully forged properties.
Source: R. H. Witt and W. T. Highberger, "Experience With Net-Shape Processes for Titanium Alloys," in Production to Near Net
Shape: Source Book. C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals. Metals Park OH. 1982. P 277
17-8. Ti-6AI-6V-2Sn (HIP): S-N Curves for Annealed Plate vs HIP
ANN PLATf'
. / eeo F-16KeI-2 /lR
120
tOO
tgTReeS,
KSI
80
JO
<,
22-5'0~-fOl<'Sl-I/-IR
'7------:
RAfIIG[ O~ PROGRAM l?ATA
40
20
~4
~~
~~
~7
CVClP8 TO {"AILURf"
S-N curves showing that HIP Ti-6AI-6V-2Sn is equivalent to annealed plate of
the same composition.
Source: W. Theodore Highberger. "Manufacture of Titanium Components by Hot Isostatic Pressing," in Production to Near Net
Shape: Source Book. C. J. Van Tyne and B. Avitzur, Eds.. American Society for Metals. Metals Park OH, 1983, P 304
421
422
17-9. Ti-6AI-2Sn-4Zr-2Mo: Bar Chart Presentation on Effects of

Machining and Grinding
Ti-bAl-lSn-4/'.r-lMo (STA, 36 Re!
Surface
Roughness, AA
SU RFACE
GRIND
END MILL-
39
41
-, 68
I-l'G~e.:.:n~tl~e_....,.."....
Cony
I 17
~IO
lZO
I-l'G~en~ltI.:i-e
"T""
PERIPHERAL CUT ...

A"'lb~u~.,:.:v~e
'"
8Z
36
77
...1 47
I
ZO
40
I
60
80
Bar chart presentation showing the effects of specific machining and grinding operations on fatigue characteristics of titanium alloy Ti-6AI-2Sn-4Zr-2Mo.
Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys,"in Titanium and Titanium
17-10. Ti-6AI-2Sn-4Zr-2Mo: Constant-Life Fatigue Diagram

Minimumstress. ksi
-100
-50
+50
100
150
1200
150
1000
.;;;
800
If.
.><
~
t:
E
:>
E
:;;
10D ~
t:
600
E
E
:>
'x
'x
~ 400
~
50
10 7 cvcles lifetime
200
0
-800
-600
--400
-200
+200
400
600
800
1000
0
1200
Minimumstress, MPa
Constant-life fatigue diagram for duplex annealed Ti-6AI-2Sn-4Zr-2Mo sheet, 1 mm

(0.04 ln.) thick.
423
424
17-11. Ti-6AI-2Sn-4Zr-6Mo: Low-Cycle Axial Fatigue Curves
1600
'"
Q.
:2
220
---
1400
STA
200
~.
'"
E
::J
E
1200
'"
:2
1000
'x
180
160
140
800
1
103
'"
.><
~.
E
::J
E
'x
'"
:2
120
105
Number of cycles
Low-cycle axial fatigue curves for Ti-6AI-2Sn-4Zr-6Mo. STA
(solution treated and aged) condition: 1 hat 870 C (1600 OF),
water quench, age 8 h at 595C (1100 OF) and air cool. DA
(duplex annealed) condition: 15 min at 870C, air cool, then 8h
at 540 C (1000 F) and air cool. All fatigue tests conducted at a
stress ratio of R = 0.1. Open symbols indicate fatigue tests; solid
symbols, tension tests.
Source: Metals Handbook, 9th Edition, Volume 3. Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
Metals. American Society for Metals, Metals Park OH, 1980, P 395
17-12. Ti-8Mo-2Fe-3AI: S-N Curves; Solution Treated and Aged

Condition
600
500
'"
e,
:::E
400
'"
'"
~
'" 300
E
:>
E
'x
:::E'"
200
80
Tension-tension tests
, R = 0,25; K r = 3.5
'.
'in
60
'"
'"
I'----
.~
- 40
E
:>
E
'x
,....
'"
:::E
- 20
100
Stress cycles
SoN curve for Ti-8Mo-2Fe-3AI titanium alloy in the solution treated

and aged condition. Data are for 1.S-mm (0.060-in.) thick sheet solution
treated 10 min at 790 0 C (14S0 0 F), air cooled, and aged 8 h at 480 0 C
(900 OF).
Source: Metals Handbook. 9th Edition, Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
425
426
17-13. Ti-10V-2Fe-3AI: S-N Curves; Notched vs Unnotched

Specimens in Axial Fatigue
1100
1000 ~o
'" 900
Q.
:!:
~
...e
800
'"
E
:::>
700
)(
'" 600
:!:
.---- --
Unnotched
lili
140
r---
'u;
.:,t.
120
A
100
<,
",'
'"
~
'"
E
:::>
E
'x
~ to-..
500
'"
:!:
----
80
I--
Stress cycles
500
'"
Q.
400
\~
:!:
o RT
~2050C (400F)
'" 100
:!:
o 425 c (800 of)
40
~
'u;
.:,t.
r-,
E 200
:::>
E
60
Kr = 3.0
\0
'"
'" 300
'x
Notched
'"
'"
E
:::>
E
20 'x
'"
:!:
I
Stress cycles
Axial fatigue of Ti-l0V-2Fe-3AI bar stock in the STOA (solution

treated and overaged) condition. Specimens were taken from round
bars 7Smm (3 in.) in diameter that had been solution treated 1 h at 760 C
(1400 OF), furnace cooled, overaged 8 h at 565C (1050 OF) and air
cooled. Tests were conducted at a stress ratio ofR= 0.1 and a frequency
of 20 Hz. Top: results of unnotched bars tested at room temperature.
Bottom: fatigue characteristics of notched specimens tested at elevated
temperature.
Source: Metals Handbook, 9th Edition.Volume 3, Properties and Selection: Stainless Steels, Tool Materials and Special-Purpose
17-14. Ti-10V-2Fe-3AI and Ti-6AI-4V: Comparison of Fatigue

Crack Growth Rates
do/d~ in/cycle (pm/cycle)
10-4-r-----r---...."7T""-------,
(25)
MA Ti'-6A/-4V
iI
i
i
10-5
(025)
Ti"-IOV2Fe-3A/
70.6
(0025)/i'...
O--.....L.L---
20
(II)
LlK, ksi
(22)
vm.
J,
...J
O'---4..J.O - - - - - - - - '
(33) (44)
(MPoW;;)
Comparison of fatigue crack growth rates. Data for Ti10V-2Fe-3AI,R = 0.05,F= 1 to 30 Hz; for MA Ti-6AI4V,R= 0.08,F= 1 to 25 Hz; for RA Ti-6AI-4V,R= 0.08,
F= 6Hz.
Fatigue crack growth rates in air have been found to lie in the
scatter band for mill annealed (MA) Ti-6AI-4V, as shown
above. At high ~Kvalues, Ti-IOV-2Fe-3Al approaches the performance of Ti-6AI-4V in the recrystallized annealed (RA)
condition.
Source: Wayne A. Reinsch and Harry W. Rosenberg, "Three Recent Developments in Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 375
427
428
17-15. Ti-1 OV-2Fe-3AI: S-N Curve; Notched Bar Fatigue Life for a
Series of Forgings Compared With Ti-6AI-4V Plate
rrmox,IOJpsi(MPo)
70-r-----~----------____..,
(48J)
60
(414)
50
(345)
40
STA Ti6AI-4V plate
(216)
30
(201)/i.l-O-4----------'-----"':::OO"------.J
Cycles tofaIlure
Comparison of notched fatigue lives for Ti-lOV-2Fe-3AI forgings and Ti-6AI-4V plate. Data for Ti-lOV-2Fe-3AI,R = O.OS,F
=KT= 2.9; for STA Ti-6AI-4V plate,R = O.l,K T= 3.
Fatigue characteristics of Ti-lOV-2Fe-3Al are equal to or superior to those of Ti-6AI-4V. Notched fatigue results are shown
above. Data from a series of die forgings have shown that the
mean value for fracture toughness is 49.1 ksi ~. (54 MPa
Jill), with a standard deviation of 2.3 ksi~. (2.5 MPa Jill).
K/scc in 3.5% NaCl is typically about 90% of the x;
Source: Wayne A. Reinsch and Harry W. Rosenberg, "Three Recent Developments in Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 375
429
17-16. Ti-13V-11 Cr-3AI: Constant-Life Fatigue Diagrams

Minimum stress, MPa
-100
-80
-60
-40
-20
+20
40
60
80
100
120
140
160
180
200
1400 r-_,_----,~-_,_-_,___,...---.-_r-_r""__,-__,-~""__.--t-er____,"'_,,,._-r____,, 200
180
160
140
120
100
"~" 600
~
80
~~
:;;
E
"K
At room
temperature
400
60
40
200
- - Unnotched
-600
-400
-200
+200
400
600
Edgenotched, Kr = 3.0
800
1000
20
1200
Minimum stress,ksi
Minimum stress,ksl
1400
-100
-80
-60
-40
-20
+20
40
60
80
100
120
140
160
180
160
140
~
0;;;
:;;
800
120 ""
~~
~E
~
"K
100
600
80
"E
o~
:;;
60
400
Al315
-c (600 OF)
40
200
- - Unnotched
- - Edgenotched, Kr=3.0
-400
-200
+200
400
600
800
1000
20
1200
Minimum stress, MPa
Constant-life fatigue diagrams for Ti-13V-llCr-3AI, STA (solution treated and aged) condition, longitudinal orientation. Data arefor axial fatigue of edge-polished sheet specimens of material solution treated and
aged to room-temperature tensile strength of 1203 MPa (174.5 ksi), Corresponding yield strength was 1080
MPa (156.7 ksi); at 315C (600 OF), the tensile strength was 1078 MPa (156.3 ksi) and the yield strength was
876 MPa (127.0 ksi), Tests were conducted at a speed of 60 Hz.
Source: Metals Handbook, 9th Edition, Volume 3, Properties and Selection: Stainless Steels. Tool Materials and Special-Purpose
430
17-17. Ti-6AI-4V: Effect of Condition and Notches on Fatigue

Characteristics
800
co
e,
::E 600
::i
~on 400
E
:::l
E
'j( 200
co
.a...a..
~ Smooth
--...:
--
bar, STA stock

-r:"---~
Smooth bar, annealedrock
Notched bar (K t = 3.5). STA stock
"" Notched bar. annealed stock
100 ';;;
80
60
on
40
::E
106
Number of cycles
S-N curves for titanium alloy Ti-6AI-4V (rotating beam) showing effects
of STA (solution treated and aged) versus annealed conditions, and effect of notches.
17-18. Ti-6AI-4V: Effect of Direction on Endurance
700
650
m
a..
~
CIl
CIl
Q)
;:,
CIl
0)
c:
550
.~
'"c:
L.
2:! 500
~
450
106
Endurance (cycles)
Rotating-cantilever fatigue (S-N) curves for three testing directions in
57 mm thick, forged and annealed Ti-6AI-4V bar.
These curves show that fatigue properties are lowest in the long
transverse direction. This result has been attributed to the fact that
Poisson's ratios are also sensitive to crystal orientation, these ratios
being higher in the longitudinal and short transverse directions because stressing occurs parallel to the basal planes. Higher ratios imply
greater constraint, which means that the levels of strain will be reduced
and the fatigue strength enhanced in these two directions. The differences observed in fatigue strengths in the longitudinal and short transverse directions have been attributed to relative changes in grain
shapes that also occur during processing.
Source: I. J. Polmear, Light Alloys. Edward Arnold Ltd, London, England, and American Society for Metals. Metals Park OH,
1981, p 193
431
432
17-19. Ti-6AI-4V: Effect of Isothermally Rolled vs Extruded Material

on Cycles to Failure
100 , - - - - - - - - - - - - - - - - - - - - - - - - - -
80
o!';.
iii
'" 60
.,;
w
'"
....a:
'"
::;
:::l
~
x
::;
40
AXIAL FATIGUE. RoO. I

ROOM TEMPERATURE
K,' 2.8 FLAT SPECIMEN WITH
0.050IN. HOLE
20
6.
ISR TEE
EXTRUSION
OL-_ _'--_---JL...--l----JL.......I_ _---L_ _---L_-L.........-..L _ _......._ _......._ ...L.............L _ _-'-_ _...... _'--..L-J
103
105
CYCLES TO FAILURE
S-Nfatigue data for isothermally rolled tees versus extruded material. The notched fatigue behavior of the
ISR tees is as good as or slightly better than that of the extrusion.
Source: W. T. (Ted) Highberger, Govind R. Chanani and Gregory V. Scarich, "Advanced Titanium Metallic Materials and Processes for Application to Naval Aircraft Structures," in Production to Near Net Shape: Source Book, C. J. Van Tyne and B. Avitzur,
17-20. Ti-6AI-4V: Comparison of Wrought vs Isostatically Pressed

Material for Cycles to Failure
1200
I 1111
co
o,
~
w'"".""~
grade Ti-6AI-4V
~
1000
I
~
"""
~
~~ ~ ~ ~ ~
'"

U;
~
125
800
~~
~
~ ~~
~ ~"
E
:J
"~
X
co
150
~~ ~
600
I I
Engine mount supports
400
10'
o Witness blocks
1lJ1I Radius failures
~~
~ ~i'-
~ ~ t:S ~"
100
10'
E
:J
E
"xco
:\I'ci
~ :\"
~~
~ ~t'- ~ ~ ~ ~"
105
10'
'"

1il
75
10'
Cycles to failure
900
co 700
n,
~
e'"

1il 500
E
I. I I I
I
o Engme mount supports
Wrought
standard
grade
Ti-6AI-4V
:J
"xco
~ 300
100
10'
~ ~~
~ ~"
~
C'\
125
~
~ ~ ~ -<: ~
100
e'"

75
~ r-;~ -:~" ~ ~ ~~
~ C'\~~ ~
~ ~ ~ ~\ ~ ~
~ ~ ~ ~~ ~ ~~
r~ ~ 88 ~ ~~
':':";lo,;
10'
"iii
-""
105
10'
1il
E
:J
50
~ r-,~ ~"
'x
co
~
25
10'
Cycles to failure
S-N curves for titanium alloy engine mount supports. Top: Data are for the standard wrought
grade;R = O.I,K,= 1.0, load controlled smooth specimens. Bottom: Data are for isostatically
pressed alloy powder, notched specimens;R = 0.1, K,= 3.
Source: Metals Handbook. 9th Edition, Volume 7. Powder Metallurgy, American Society for Metals. Metals Park OH. 1984.P 654
433
434
17-21. Ti-6AI-4V: Effect of Fretting and Temperature on Cycles to

Failure
~
z
~
<Jl
<Jl
560
a:
In 420
Fatigue (all test
,om,,,":,,,,
R.T
200C
o 340C
z
~280
a:
w
~ 140
<l:
1LO"':'~
---..L.::~-----~------.L:;------'
1(f
CYCLES
10~
TO
10~
FAILURE
Effect of simultaneous fretting and fatigue ofshot-peened Ti-6AI-4V at room temperature, 200, and 340 C at a mean tensile stress of 140 MNm2
Room temperature fretting was found to have little effect on the fatigue strength
at 107 cycles. Fretting at 200 C lowered the fatigue strength by approximately IS%;
furthermore, the fretting fatigue life in the overstress region (70 MNm- 2 above the
nonfretted run out stress) was lowered by two orders of magnitude compared with
results in the absence of fretting. At 340 C, fretting similarly reduced specimen life
7
at overstress levels; however, more importantly, the fretting fatigue strength at 10
cycles was reduced to approximately 40% of that found under room temperature
conditions.
The gross result of fretting normally is fatigue failure brought about by surface
damage in conjunction with normal or transient high stresses in a component. It
should be said at the outset that visual assessment of fretting mildness or severity is
inconclusive by itself, in that the presence of more or less fretting debris on a
microscopic examination is not necessarily relevant to the loss of surface integrity.
It may in fact be misleading and should not be relied upon for assessing the severity
offatigue life degradation. It is the stress state acting in concert with stress raisers
(e.g., pits, tears, cracks) which determines the actual fatigue propensity.
Source: Practical Observations of Fretting Fatigue Cracks. p 180
17-22. Ti-6AI-4V (Beta Rolled): Effect of Finishing Operations on

Cycles to Failure
FATIGUE CHARACTERISTICS OF BETA ROLLED TITANIUM 6AI-4V,
az R c
METAL REMOVAL PROCESSES: SURFACE GRINDING. PERIPHERAL END MILLING.

CHEMICAL MILLING
MODE: CANTILEVER BENDING. ZERO MEAN STRESS
TEMPERATURE: 75' F
U;
80
:<:
en
lol
'"
0:
f.<
'"
60
~
~I
-..;
I I I III
............. .... :040
f.<
...l
"-
..;
zo
10 5
III
""'-- I
\J
Z
E=:
..;
Z
0:
lol
II
ENDUR.
LIMIT
SURF.
FINISH
GENTLE MILL
GENTLE GRIND
66
6l
41
35
CHEM. MILL
51
zo
ABUSIVE MILL
az
59
ABUSIVE GRIND
13
65
CONDITION
r-, .....
----
10,6
I<SI -;v;:-
I
I
II
107
CYCLES TO FAILURE
S-N curves for beta-rolled titanium alloy Ti-6AI-4V. Curves show the effects of the various
finishing operations on fatigue.
Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982,p 354
435
436
17-23. Ti-6AI-4V: Effect of Yield Strength on Stress-Life Behavior

700
Ti - 6AI-4V, 24h 500C
-..
tJ
a..
:::E
650
0"0.2
(J"02
910 MPo o eooppm 02

990 MPo 1900ppm 02
ILl
0
::>
I...J
600
Q..
:::E
550
UI
UI
ILl
a:
I-
500
UI
450 4
10
CYCLES
TO FAILURE
Effect of yield strength on the stress-life behavior of two Ti-6AI4Valloys.
In order to establish microstructural effects on fatigue behavior, comparisons should be made on materials having the same
yield stress, especially for stress-controlled tests. This is illustrated above where it is shown that two titanium samples having
different yield strengths have different stress-life behavior when
tested at 500C (930 OF).
Source: Edgar A. Starke, Jr., and Gerd LOtjering, "Cyclic Plastic Deformation and Microstructure," in Fatigue and Microstructure,
American Society For Metals, Metals Park OH, 1979, P 237
17-24. Ti-6AI-4V: Effect of Stress Relief on Cycles to Failure

60
II II
0-AS RECEIVED I
e- H T 400"r. 2 HRS.
.~
"'" 40
1'.,
iii
II:
~
I
\.~
~
30
Ul
-,
~
Iii
20
kr
- BlOKE IN GRIPS
CYCLES
Flexural fatigue tests of titanium sheet (123,000 T.S.).
Flexural tests of the sheet specimen were made at 1725 CPM. Results are indicated by the
open points in the above SoN diagram. The endurance limit was not reached at stresses as low as
20,000 psi.lfan estimated limit of 19,000 psi is chosen, the endurance ratio would be only 0.155,
a value considerably lower than for any other known metal or alloy. Most investigators have
obtained normal values around 0.3 in similar tests.
Several of the sheet fatigue specimens developed fatigue cracks away from the milled specimen
edges. The cracks did not appear to be associated with any visible surface imperfection. For these
reasons, it was assumed that the sample was abnormal, rather than the test procedure. Very
careful oil-powder and fluorescent powder tests, supplemented by metallographic examination,
failed to reveal any surface cracks, even when the sheet was flexed to open any incipient hairline
defects.
It was considered possible, though not probable, that residual stresses from cold rolling were
acting in a deleterious manner. If so, a moderate temperature stress relief might help. Brief
experiments soon disclosed that temperatures at least as high as 400 OF(205C) did not lower
the hardness; in fact, the hardness may have increased very slightly. Knowing this, a set of sheet
fatigue specimens was stress relieved for two hours at 400 OF(205C). The solid points in the
graph above represent the results obtained with these specimens. The endurance limit was not
altered significantly. A definite shift to the left in the upper portion of the curve was evident,
although the direction of shift was opposite to that, had the heat treatment released undesirable
stresses.
Source: Titanium Symposium, Office of Naval Research, p 97
437
438
17-25. Ti-6AI-4V: Interrelationship of Machining Practice and

Cutting Fluids on Cycles to Failure
lee.
E 7EUI
~--
. e __
~
_
---==='~II
--
~~
e.
.Iee
.28e
--_. _--
..
C
~_==
--
--.~--
----
....
--..-----... --= ----=
.60e
I.ee
2.e8
CYCLES TO FAILURE (IN MILLIONS)
le.ee
6.00
Alternating stress vs cycles to failure in high cycle fatigue of machined titanium

surfaces using neutral, chlorinated, and sulfurized soluble cutting fluids. A =
abusive grinding; B = low stress grinding; C = end milling.
Influence of Chlorinated and Sulfurized Cutting Fluids

On High Cycle Fatigue Properties of Ti-6Al-4V
Machined Surfaces at 75F
Sol. Cutting
Fluids
10 7 Cycle Fatigue Strength (ksi)

Low Stress
Abusive
End Milling
Grinding
Grinding
Neutral
75
62.5
12.5
Chlorinated
65
57.5
12.5
67.5
12.5
Sulfurized
Source: V. A. Tipnis and J. D. Christopher, "Machinability Testing for Industry," in Machinability Testing and Utilization of
Machining Data, American Society for Metals, Metals Park OH, 1979, p 26
17-26. Ti-6AI-4V: Relative Effects of Machining and Grinding

Operations on Endurance Limit
Tio 6AI-4V B ETA ROLLED, l2 R c
Surfac e
Roughness. AA
SURFACE
GRIND
I 62
Gentle
~13
HAND
GRIND
r;~ntl~
END MILLEND CUT
Gentle
80
80
157
I 30
Ab .. etve
164
r;~ntl~
I 66
Ah ... lu~
~""n",
CHEMICAL
MILLING
nu_~
67
84
177
u ........
END MILLPERIPHERAL CUT
35
65
'II
59
I l2
...
20
165
51
14 5
I
20
40
60
80
Bar chart presentation showing relative effects of various machining and grinding operations on fatigue characteristics of titanium alloy Ti-6AI-4V.
Source: Norman Zlatin and Michael Field, "Procedures and Precautions in Machining Titanium Alloys,"in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr . Ed., American Society for Metals, Metals Park OH, 1982, P 354
439
440
17-27. Ti-6AI-4V: Effects of Various Metal Removal Operations on

Endurance Limit
Ti- 6Al-4V ANNEALED, 35 R c
Surface
Rouglme9s, AA
Gentle
SURFACE GRIND
14
67
Il
160
ECM FRONTAL
161
I 40
ECM TREPAN
I
lO
J
40
60
80
Bar chart presentation showing effects of various metal removal operations on the fatigue
characteristics of titanium alloy Ti-6AI-4V.
Source: Norman Zlatin and Michael Field, "Procedures and Precautions in MachiningTitanium Alloys."in Titanium and Titanium
17-28. Ti-6AI-4V: Effect of Texture on Fatigue Strength

EFFECTS OF TEXTURE ON SMOOTH FATIGUE LIFE
130
120
.
..
... 110
I
<,
b lOa
Load
axis
.....
II [loio la
........
........................
90
80 L--
Ti-6AI-4V Re-X Anneal

R =0.1
.L..-
........
.....L
--------J
Nt - cycles
SON curves showing the effect of texture on the fatigue strength of Ti6AI-4V. Fatigue strength is greater when the stress axis coincides with
the direction of a high density of basal poles.
Source:J. C. Williams and E. A. Starke.Jr., "The Role of Therrnomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys," in Deformation, Processing, and Structure, George Krauss, Ed., American Society for Metals, Metals Park
OH, 1984, P 334
441
442
17-29. Ti-6AI-4V: Effect of Complex Texture on Cycles to Failure

Ti-6AI-4V
~
900
-o~-
::;:
.,
-[]o-
\
\
800
-0
-,
<,
.-
a.
E
700
II>
II>
'"
Air
...... ....
-- - - - - - - -
600
<fl
10 4
10 5
106
Cycles to Failure
10 7
S-N curves showing the effect of more complex texture on

fatigue strength of Ti-6AI-4V. These data show that a
mixed texture lowers the high-stress end oftheS-N curve
preferentially.
Source: J. C. Williams and E. A. Starke, Jr., "The Role of Thermomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys, "in Deformation, Processing, and Structure, George Krauss, Ed., American Society for Metals, Metals Park
OH, 1984, p 335
17-30. Ti-6AI-4V: Effect of Texture and Environment on Cycles to

Failure
TI-6AI-4V
-,
AIr
-,
<,
............
_-----
lOS
Cycles to Failure
Q.
..
<l:
~
31/2 %
Ti-6AI-4V
900
IJ)
800
\
\
700
Nael Solution
\
\
-,
-,
-,
<,
600
~------
104
105
106
10 7
Cycles to failure
b
O! + ,8- processed Ti6Al4V, showing the
effects oftexture and environment on fatigue strength. (a)
Tested in air. (b) Tested in 3'12% NaC\, These data show
that testing in an aqueous 3'12% NaCI solution reduces
fatigue strength when the stress axis is along [0001].
SoN curves for
Source:J. C. Williams and E. A. Starke, Jr.. "The Role of Thermomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys," in Deformation. Processing, and Structure, George Krauss, Ed .. American Society for Metals, Metals Park
OH, 1984, p 335
443
444
Kmin+lJ.KalK 1c
.::"
1
..,z
.....
..,e
o
'"
10- 10
L-_ _...L<...
log
lJ.K-
Schematic plot showing characteristic shape offatigue crack growth rate (doldN) versus cyclicstress-intensity (~) curves.
It can be seen that at higher growth rates there is a linear

portion of the curve. This linear portion was represented as tiK'"
by Paris and Erdogan and is now frequently referred to as the
Paris law regime of fatigue crack growth. Most structural materials show variations in near-threshold FCP rate and in tiK,,, but
fewer show significant variations in FCP rate in the Paris law
regime. In contrast, Ti alloys show significant variations in FCP
rate over the entire range. At the highest crack growth rates
shown above, the FCP rate curve bends upward. This is controlled by fracture toughness. However, since crack growth
rates are uncontrollably rapid in this latter regime, it is of little
interest and will not be discussed further here. Moreover, since
the majority of the lifetime of a crack component is spent in the
low-FCP-rate regime, factors which control FCP at rates less
than ~ 10-6 tu] cycle are probably most important. These factors
include microstructure and texture.
Source: J. C. Williams and E. A. Starke, Jr., "The Role of Thermomechanical Processing in Tailoring the Properties of Aluminum
and Titanium Alloys," in Deformation, Processing, and Structure, George Krauss, Ed., American Society for Metals, Metals Park
OB, 1984, P 338
17-32. Ti-6AI-4V: Fatigue Crack Growth Rates for ISR Tee, and
Extrusions
10-
rr========::::::;,-----,
Ti-6AI-4V
R = O.l,ROOM TEMPERATURE
TL AND LT ORIENTATION
ISR TEE
o EXTRUSION
W
...J
000
>-
~
:r:
o
z
~
"tl
w
I-
a:
:r:
a:
(9
~
a:
o
w
::J
(9
lL
o
o
-7 ':------:~--........-_:';:-.......-7:........~::_":~
10 10
20
40
60
80 100
STRESS INTENSITY FACTOR RANGE.~K. KSI viN.
Fatigue crack growth results for ISR tee, and extrusions.
Data for both TL and LT orientations are shown above,

along with data for the extrusion. In comparing individual results, no differences were seen between TL and LT. In the chart,
it can be seen that particularly at lower stress intensities the fatigue crack rate for the (lSR) isothermally rolled tee is faster
than that for the extrusions. This is probably due to the extrusions being beta formed while the ISR tees are alpha-beta
formed.
Source: W. T. (Ted) Highberger, Govind R. Chanani and Gregory V. Scarich, "Advanced Titanium Metallic Materials and Processes for Application to Naval Aircraft Structures," in Production to Near Net Shape: Source Book, C. J. Van Tyne and B. Avitzur,
445
446
: : (INCHES/CYCLE)
I.IE
0
~
_11
~
-
,~
'
o
10
100
dK (KSI {INCH)
0- HEAT 3
ENVIRONHENT - LOW-HUHIOITY AIR
b - HEAT 2
ORIENTATION - RW
0- HEAT 1
R FACTOR
-to. 30
HEAT TREATHENT - RECRYSTALLI ZATI ON ANNEAL
Fatigue crack growth rates for three different heats of Ti-6AI-4V

titanium alloy.
The fatigue crack growth rate in the RW orientation for this

alloy, when recrystallization-annealed, behaved similarly with
decreasing oxygen and aluminum. The crack growth rate is
shown as a function of 6.K tested at an R factor of +0.30.
Source: M. J. Harrigan, M. P. Kaplan and A. W. Sommer, "Effect of Chemistry and Heat Treatment on the Fracture Properties of
Ti-6AI-4V Alloy," in Titanium and Titanium Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals,
Metals Park OH, 1982,P 65
17-34. Ti-6AI-4V: Effect of Final Cooling on Fatigue Crack Growth

Rates
~K, MPa
-.jm
100
10
I
CONSTANT AMPLITUDE FATIGUE:

R = 0.1; 10 TO 20 HZ
MAXIMUM SCATTER OF ACTUAL
DATA POINTS FROM ANY CURVE
IS LESS THAN 40%
WQ
AC
FC
WATER-QUENCHED
AIR-COOLED
FURNACE-COOLED
A'A
MILL-ANNEALED (DATA IN- ~

CLUDED FOR COMPARISON)
FC
WQ
k~
'!:/lAC
10.3
'"
o
>..
u
':-
.s
100
:/
10- 4
:'/
:/
1I
~
"U
:/
:,:"
,.":'
10. 5
I
0.1
10
~K
ksi-
100
Jin":
Effects of final cooling rate on fatigue crack growth rate in

duplex-annealed Ti-6AI-4V, I-in. plate, 1775 OF (968.3 0q, 1/2
h, air cooled; and 1450 OF (787.4 "C), I h, cooled as noted.
From the data presented above, it can be seen that air cooling, per se, produced
little or no change in the cyclic crack growth compared to the mill-annealed base
(22
material. The slightly decreased crack growth rates above a IJ.K of 20 ksi
MPa VITi) are, more probably than not, the result of the higher fracture toughness
of the air-cooled material. However, both water quenching and furnace cooling
resulted in fatigue crack growth rates noticeably different from those measured for
the mill-annealed base material. As shown, furnace cooling had a consistently detrimental effect on the crack growth rate while water quenching produced $!eatly
increased crack growth rates above a stress-intensity range of 18 ksi yin. (20
MPa . ~). The accelerated growth rate above 18ksi \!Ill. (20 MPa .
may be
attributed to the proximity of the maximum stress intensity to the critical value.
The critical stress-intensity value for water quenching was an exceptionally low 38
ksi .jll1. (42 MPa . Jill).
Jill.
Jill)
Source: R. E. Lewis, J. G. Bjeletich, T. M. Morton and F. A. Crossley, "Effect of Cooling Rate on Fracture Behavior of MillAnnealed Ti-6AI-4V," in Titanium and Titanium Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for
447
448
17-35. Ti-6AI-4V: Effect of Dwell Time on Fatigue Crack

Growth Rates
10- 4 ,.--
u
>u
10- 5
....
Q>
.s
a.
~
.J::.
0,
...,
I
I
I
I
Q>
'i0
10- 6
-"
u
Sinusoidal
loading
--Dwell at
maximum load
10-7
100
10
Stress intensity
factor range (MPa m 1/ 2 )
A phenomenon which may be unique to certain titanium alloys is the effect of dwell periods at maximum load on rates of
growth of fatigue cracks. This effect is shown schematically
here, and increases in the rate of crack growth of as much as 50
times may occur compared with results obtained in tests on the
same material subjected only to sinusoidal stress cycles. Dwell
effects are maximized in alloys containing substantial amounts
of the a-phase which have a preferred texture such that stressing
is normal to the basal planes, whereas they appear to be insignificant if stressing occurs parallel to the basal planes of the aphase, or if the microstructure is homogeneous and fine grained.
Particular attention has been paid to a / {3 alloys, e.g., Ti-6AI4V, in which dwell effects have also been found to decrease with
increasing amounts of the {3-phase in the microstructure. In all
cases, dwell effects disappear when stressing occurs at temperatures above 75C (165 "Fj.and they are generally considered to
arise from the preferential diffusion of hydrogen, during the
dwell period, to regions of localized hydrostatic tension ahead
of an advancing crack. Such an accumulation of hydrogen
would tend to embrittle this region, and it has even been suggested that brittle plates of TiH 2 may be formed.
1981, P 200
17-36. Ti-6AI-4V: Fatigue Crack Growth Data

0
Annealed 2 hours at 705 C. air-cooled after forging

transus 1005 C
" Axial loading: smooth specimens. K = 1.0
0
alP
I -. ", .'
" ..
Ti-6AI-4V
-SA
10
20
50
100
Left: Fatigue crack growth rates for Ti-6AI-4V rolled

plates in the ,a-annealed (HA) and mill annealed (MA)
conditions. BA = 0.5 h 1038 0 C, air-cool to room temperature. Tests conducted at 5 Hz using compact tension
specimens. Ratio of minimum to maximum load = 0.1.
Above: Branching of fatigue cracks within the Widmanstiitten packets of the a-laths.
Stress-intensity factor range

LlK(MPa m1J2 )
Work on Ti-6Al-4V rolled plate has indicated that the superior fatigue performance with the {3-annealed condition is associated with relatively slower rates of
crack propagation (above graph). This effect, in turn, is attributed to the slower
progress of cracks through the Widmanstatten microstructure, particularly at stress
intensities below a critical value at which desirable crack branching occurs within
packets ofthe a-laths.
1981, P 179
449
450
17-37. Ti-6AI-4V P/M: Comparison of HIP'd Material With

Alpha-Beta Forgings for Cycles to Failure
r:
70
~o
o
~OHIP
------00------
40
Upper and lower limits of
~-_-:--.. alpha-beta-processed forgings
as listed in AFML TR-73-301

30 LJ...Ll..lll_--.J_L-L.l.....cr::ti==I==:=~:::l.-'~
106
10
Cycles to failure
Notched fatigue strength of HIP'd P 1M Ti-6AI-4V compared with fatigue strength ofalpha-beta processed forgings.K, = 3; Hz = 30;R=
0.1.
Source: J. H. Moll, V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications," in Powder MetallurgyApplications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 286
17-38. Ti-6AI-4V P/M: Comparisons of HIP'd Material With

Annealed Plate for Cycles to Failure
140
120~
100
'GM"S'$,
K'S/
80
~ Q I/IPC'tCl.
o
15S0i:-I-5 1t:.'6/-II-IR
40t
''''50.~-1'51t:'8/- 3/-1J:?
22'50~-I0/t:!9I-II-IR
ANN P!AT~
20 1----,-_ _--'--_ _---"I

10~
/0'-
10-5
-'-:-_ _----',
1
/0
/0'
C~ESTO~(L~
SoN curve for HIP'd Ti-6AI-4V and annealed plate. According to the
above data, fatigue results for Ti-6AI-4V are within the required range
for plate properties from MIL-T-9046.
Source: W. Theodore Highberger, "Manufacture of Titanium Components by Hot Isostatic Pressing," in Production to Near Net
Shape: Source Book, C. J. Van Tyne and B. Avitzur, Eds., American Society for Metals, Metals Park OH, 1983, P 304
451
452
17-39. Ti-6AI-4V P/M: Effect of Powder Mesh Size on

Fatigue Properties
70.3 (100)
56.2 (80)
'Vi
~
m
a..
42.2 (60)
:E
VI'
VI
Q.l
.......
en
28.1 (40)
o As received powder (SM 772)

0-80 mesh screened powder (SM 768)
14.1 (20)
0
104
10,
70.3 (100 )
56.2 (80 )
'u;
-:-- r----
42.2 (60 )
li.
--
~li.
Pli.
Ali.
o~l
:E
....~'
28.1 (40 )
O As received powder (SM 772)
li. -80 mesh screened powder (SM 768)
en
14.1 (20 )
io,
10,
Top: Room-temperature properties. Bottom: Properties at 700 OF (370C).
High-cycle fatigue (HCF) data were developed on Ti-6AI-4V (Std) by Williams International in a program to apply near-net-shape HIP technology to
a compressor rotor part for the F-I07 cruise missile engine. In this study, two
size fractions of powder were used: -35 mesh (as-received) and -80 mesh.
There was no difference in HCF test results between the two sizes. Roomtemperature and 700 OF (370C) S/ N curves are shown above.
Source: J. H. Moll. V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications,"in Powder MetallurgyApplications, Advantages and Limitations, Erhard Klar, Ed., American Society for Metals, Metals Park OH, 1983, P 286
17-40. Ti-6AI-4V P/M: Comparison of Blended Elemental,

Prealloyed and Wrought Material for Effect on Cycles to Failure
1200
1000
140
co
Q..
'iii
iii
III
.><
iii
800
III
e
li)
100 ti
E
:J
E
E
:J
E 600
'x
co
'x
co
60
400
200
10'
10'
10'
10'
Cycles to failure
10'
10'
SON curves showing comparison of smooth axial fatigue behavior of Ti-6AI-4V blended elemental and prealloyed P 1M compacts with wrought annealed material. Tested at room
temperature, R = O.
The fatigue behavior of titanium PIM compacts is compared to wrought products in

the graph above. The blended elemental material is inferior to prealloyed compacts and
II M materials. This is caused by residual chlorides and consequent porosity; also, chemcial heterogeneity may lead to areas of similarly aligned alpha plates. Blended elemental
compacts, however, compete well with many titanium alloy castings in fatigue strength.
Prealloyed powder compacts exhibit fatigue behavior equivalent to that of Ij M materials.
This situation is achieved by careful control of cleanliness (powder handling) and microstructure. Cleanliness depends on the environment in which the powder is produced, conditions of subsequent handling, and microstruture developed by compaction. Cleanliness
dictates the amount of contamination contained in the final product; microstructure determines the ability of the compact to accommodate foreign particles and resist crack
initiation.
Source: Metals Handbook. 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,P 753
453
454
17-41. Ti-6AI-4V: P/M Compacts vs 11M Specimens:

Cycles to Failure
co
Q..
1200
'"
1000
en
1ii
800
E
::J
E
600
CO
400
103
'(ij
.>t:
10
106
10
10
f/l
~f/l
100
E
::J
E
60
104
'"
140
'x
CO
~

Fatigue chart presentation showing a comparison of fatigue behavior of Ti-6AI-4V compacts with ingot metallurgy material.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, P 44
17-42. Ti-6AI-4V: Comparison of Specimens Processed by Various

Fabrication Processes for Cycles to Failure
1200
Axial fatigue smooth
Room temperature
R = 0,1
Annealed
ro
a..
~
160
120
800
'ijj
..".
vi
u>
III
III
~
1;;
~
1;;
E
::J
E
80
E
::J
E
'x
ro
'x
ro 400
40
o"10
-'-
...L...-
--'
....L..
---I0
108
Cycles to failure
SON curves (bands) for titanium alloy Ti-6AI-4V processed by various fabrication processes.
The inconsistentfatigue life ofthe hot isostatically pressed product is usually casued by inclusions in the compact.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984,P 439
455
456
17-43. Ti-6AI-4V: Comparison of Fatigue Crack Growth Rate,

P/M vs 11M
Stress intensity, ksiVTrl.
10
10-4
10
11M and P/M Ti-6AI-4V
at room
temperature
(laboratory air),
R = 0.1, at 5 to
30 Hz
102
10-2
Q)
Q)
~
c
>
10- ~
E
E
......ai
i...
ai
III
.s::. 10- 5
Recrystallization
anneal
s:
...0
...001
10- 4 ~
.>0:
(J
(J
...
...
o
III
III
10- 6
10- 5
Stress intensity, MPaViTl

Comparison of fatigue crack growth rate of Ti6AI4V
P 1M compact with 11M material heat treated to various
conditions. The fatigue crack growth rate of blended elemental and pre alloyed compacts is equivalent to 11M
material with the same microstructure.
Source: Metals Handbook, 9th Edition, Volume 7, Powder Metallurgy, American Society for Metals, Metals Park OH, 1984, p 752
17-44. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for

Cycles to Failure
130
<,
<,
110
0---,
<,
l!l'
c:
<,
u;
>l
TEST CONDITIONS:
CONSTANT AMPLITUDE, R ~.1
K ~ 1.0
T
STRESS RELIEVED, FLUSH WELD BEAD
90
a:
t-
IJ)
:E
:::>
:E
"""-
CD
~- ~
70
:E
LEGEND
50
---/:::,.
BASE METAL
--0
SSEB WELDEO
10
NUMBER OF CYCLES (LOG SCALE) TO FAILURE
S-N curve for titanium alloy plate-base metal versus SSEB-welded. Results show that the constantamplitude fatigue life of SSEB weldments in O.440-in.-thick plate equals that of the base metal.
Source: R. H. Will, J. G. Madora and H. P. Ellison, "Sliding-Seal Electron-Beam Welding of Titanium," in Source Book on
Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American SccietyIor Metals, Metals Park OH, 1981, P 87
457
458
17-45. Ti-6AI-4V: Base Metal vs SSEB-Welded Material for

Cycles to Failure
120
TEST CONDITIONS:
CONSTANT AMPLITUDE, R
KT~
~.1
10
STRESS RELIEVED, FLUSH WELD BEAD
1lX'
80
o
~
LEGEND:
40
--- 0
SSEB WELDED
---- 6
BASE METAL
20
SON curves for titanium alloy plate-base metal versus SSEB-welded. Results show that the constantamplitude fatigue life of SSEB weldments in O.940-in.-thick plate equals that of the base metal.
Source: R. H. Witt, J. G. Madora and H. P. Ellison, "Sliding-Seal Electron-Beam Welding of Titanium," in Source Book on
Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, p 87
459
17-46. Ti-6AI-4V EB Weldments: Base Metal Compared With

Flawless Weldments
140
O.OSQ-IN.THICK BASEMETAL
Ti-6AI-4V STOA CURVES
120
ci
s"
100
en
en'
en
w
a:
Ien
::E
SO
60
40
20
103
10 5
CYCLES
SoN curves for EB weldments that were flawless (lower two curves). Upper curve shows scatter band for base
metal (O.080in.-thick Ti-6AI-4V STOA.).
Source: R. Witt, A. Flescher and O. Paul, "Weldability and Quality of Titanium Alloy Weldments," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982,P 313
460
17-47. Ti-6AI-4V EB Weldments: Effects of Porosity on

Cycles to Failure
140
a
a
120
O.oaOIN.THICK
aASEMETAL
Ti6AI4V STOA CURVES
100
II
a:
in
~'
ao
a:
Ii;
X
<l:
:E
60
40
20
103
105
CYCLES
SON curves for ED-welded Ti-6AI-4V titanium alloy showing effects of porosity.
Above are shown experimental data obtained for porosity-containing ED welds which are
superimposed on a set of curves for the base material (0.080-in.-thick Ti-6AI-4V STOA sheet) at
various K, factors. For the points within the boundaries of the band, radiography indicated
scattered porosity (0.003 to 0.005 in. in diameter). For points below the lower boundary of the
band, radiography indicated either linear or heavily scattered porosity.
Source: R. Witt, A. Flescher and O. Paul. "Weldability and Quality of Titanium Alloy Weldments," in Titanium and Titanium
Alloys: Source Book Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, P 312
461
17-48. Ti-6AI-4V Gas Metal-Arc Weldments: Effects of Porosity on

Cycles to Failure
140
0.250-IN.-THICK BASE-METAL
Ti-6AI-4V STOA CURVES
120
0
~ 100
Cii
~
iii
lI)
w
a:
BO
lll)
::E
60
D
40
o
105
CYCLES
SON curves for porosity-containing gas metal-arc welds. In the above graph the experimental fatigue
data for porosity-containing GMA W weldments are superimposed on SoN graphs for Ti-6AI-4V
STOA material (0.25 in. thick) for various K 1 factors.
Source: R. Witt, A. Flescher and O. Paul. "Weld ability and Quality of Titanium Alloy Weldments," in Titanium and Titanium
Alloys: Source Book, Matthew J. Donachie, Jr., Ed., American Society for Metals, Metals Park OH, 1982, p 313
462
17-49. Ti-6AI-4V: Unwelded vs Electron Beam Welded Material for

Cycles to Failure
III
0..
en
III
w
It:
t;
---
100
80
60
-'"
..
'r.I
.. Q=
(!)
5z
40
20
It:
W
C[
f 7500 CPM
Kr
HEAT TREATED TI-6AI-4V
HEAT TREATED
ELECTRON BEAM WELDED
a STRESS RELIEVED TI-6AI-4V
=1.0
I
6
10
CYCLES TO FAILURE
Room temperature rotating-beam fatigue life of unwelded and electron-beam-welded Ti-6AI-4V titanium
alloy in fully heat-treated condition. Decrease in fatigue strength ofthe weldment relative to the parent metal
did not exceed 12%.
Source: S. M. Silverstein, V. Strautman and W. R. Freeman, "Application of Electron Beam Welding to Rotating Gas Turbine
Components," in Source Book on Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals,
Metals Park OH, 1981, P 169
17-50, Ti-6AI-4V: S-N Diagram for Laser-Welded Sheet

120
.....
....
ci
Legend
$1!!~iJfift:
100
II
a:
.....
'iii
.....
'@
80
Ul
Ul
E
E
:;,
oco
.0
-,
Gl
Ul
60
<,
')(
IV
::iE
o 0.230 in, (0.584 cm) sheet

0.140 in. (0.356 cm) sheet
Mean curve for plasma-arc welds 700 co.....
I
0
of mean fatigue strength
....
of all un-welded control
)(
600 N
specimens
40
20
<,
<,
-.....
E
---
----
Lower bound data for pIa welds J

on flat sheet with filler
10 6
4 runouts
.....
l3Gl
Ul
E
:;,
E
400
')(
----
10 7
Cycles to failure
IV
300
::iE
10 8
S-N diagram for laser-welded titanium alloy sheet,
The fatigue properties of welds as shown above indicate that under proper
welding conditions, laser welds can be made in Ti-6AI-4V which exhibit base
metal fatigue characteristics. The best laser weld failures initiated at sites in
the base metal, whereas other weld failures originated at undetected small
pores. Where failures initiated in the base metal, it was concluded that no
porosity or weld defects of sufficient size to preferentially initiate fatigue
fractures was present.
Source: E. M. Breinan, C. M. Banas and M. A. Greenfield. "Laser Welding-The Present State-of-the-Art," in Source Book on
Electron Beam and Laser Welding. Melvin M. Schwartz. Ed., American Society for Metals, Metals Park OH, 1981, P 289
463
464
17-51. Ti-6AI-4V (Cast): S-N Diagram for Notched Specimens

1000
900
800 t- o
..
700
:E
600
Do
~.
...'"
~
E
::J
E
'j(
:E
500
400
300
200
120
oooeJO
- 90
0000y 00
0
0
00'17001:>.61:>.
- 60
~f:>6'17
00
+000#
o 00+-tp.~66 '17 'Ii' '17

o 000
01:>.Q)
+0
x ooo,+oOOx O
+0
0 o-tCo
00
00
'';;
.>t
.n
...e'"'"
E
::J
E
..
'j(
:E
0
0
+o~ooo
30
oc:P
100
No. of stress cycles
Notch fatigue strength of as-cast Ti-6AI-4V. Each symbol represents fatigue data from a different source. Stress ratio,R, typically was +0.1; stress concentration factor, K/, was mostly 3.0,
but a few tests were run at K/ = 1.0.
Metals, American Society for Metals, Metals Park OH, 1980,P 411
18-1. Zirconium 702: Effects of Notches and Testing Temperature on

Cycles to Failure
'"0.....
N
e
U
-,
Cl
-"
to
en
en
w
a::
l-
en
0.5
CYCLES
S-N curves for zirconium grade 702, showing effects of notches
and elevated temperature (400C, or 750 OF) on fatigue
characteristics.
As indicated above, zirconium and its alloys exhibit a fatigue

limit behavior similar to most ferrous alloys.
Source: Donald R. Knittel, "Zirconium." in Corrosion and Corrosion Protection Handbook, Philip A. Schweitzer, Ed., Marcel
Dekker, Inc. New York NY, 1983. p 198
465
466
19-1. Steel Castings (General): Effect of Design and Welding

Practice on Fatigue Characteristics
'W:....
S-N CURVE -CYCLES OF
X "" STRESS VS COMPUTED
"" .... x STRESS

....
.... ~ x Xx
00 .....x
~~~
::f ....0.1:-:!-::
o ...............
x~x
10 6
107
NUMBER OF CYCLES
S-N curve for cast box designs.

u,
50
(5.Q.~
~ I 40
Vl
llJ
[3 ~ 30
11::11-0
VI Cl:
olE
llJ
-I
::::I
(3
x,oX,
S-N CURVE -CYCLES OFSTRESS

COMPUTED STRESS
'o'x
VS
'-.""',
'"'ox
lbX~',
"
"'80x.. .
XX
x
-_
--o-J!l<u_x)(""
....x....
20
0",,--
-0_
0_
10
X - 2-BW HAND WELDED, STRESS RELIEVED-o - 2 - BW MACH WELDED. STRESS RELIEVED-
105
106
NUMBER OF CYCLES
SoN curves of box weldments, comparing hand weldments with

machine weldments. All weldments were stress relieved at 1100 F
(593 Qq.
Q
The S-N curves shown above indicate that: (1) the welding
practice is of no great importance; and (2) the cast steel box
design is superior to a weldment design.
Source: Steel Castings Handbook. 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH, 1980,
P 7-6
19-2. Steel Castings (General): Effects of Discontinuities on

1.0
0.9
0.6
0.7
0.6
0.5
:I:
l-
:I:
I-
lI)
lI)
W
.J
e> ~ 0.4
Z
It: It:
l- I-
e>
~
<l: W
u, I-
AVERAGE YIELD STRESS
RANGE
TENSILE STRENGTH
0.3
0.2
0.1
0
0.5
MEAN
- 0.1
0.6
STRESS
TENSILE STRENGTH
- O. 2
- 0.3
-0.4
Goodman diagram for bending fatigue for normalized

and tempered 8630 cast steel. (Machined notch of R. R.
Moore specimen: 60 included angle, 0.0015-in. (0.0381mm) root radius.]
Surface condition has a significant effect on fatigue life and fatigue limit. A
highly polished smooth test specimen can exhibit twice the fatigue strength of
a rough machined sample. A good design approach is to use the notched
fatigue limit asa design value. For cast steels a O.OOIS-in. (O.0381-mm)root
radius circumferential notch in a rotating beam fatigue specimen reduces the
fatigue limit by about 0.7 of the unnotched value. This is sufficient to account
for variations in surface finish and minor surface discontinuities. The above
diagram shows that even severe surface discontinuities, not normally permitted by workmanship standards, do not reduce the fatigue limit by much more
than the 0.7 value.
The above emphasis on surface discontinuities is due to the fact that subsurface discontinuities which do not have a crack-like sharpness and which
do not significantly reduce the load-bearing area of a component generally
have little effect on fatigue performance.
P 7-6
467
468
20-1. Closed-Die Steel Forgings: Effect of Surface Condition on

Fatigue Limit
9oor--....,._~---------,r------------,----------~
125
Ground and polished

8oot-----.::>.,;~~<?'
.......:::_II__-___7'''-------_+---------~
7oo~------~......,;~~-..-==-"""'-"""",------,,L----+----------d
100
6oo~--------___I--=~:::""....,--=_=-=""'l-"",""----------l
500 ~--------___I---___;.,L_-=....,.-=.......; ::_t--"'-..=_-:=.""""=__--=l 75 ~
~ 4oot-----------1------------'''''f.,=---'''......, -----'''''-..---=c.j
Ul
Ii
l'
50 Ul
3oot-----------1----------+----"'-,--=---~c.j
2oot-----------1----------+---------''''''''',----~c.j
25
loot-----------1----------+------------''-l
0'-------------'------------'------------' 0
103
106
50
100
150
1000
I II
800
MaChine[5h
D..
::;;
600
-:V
]
'!l
.~ 400
200
125
~
100
75
/
--- >-_L --1,....--t.>
......
"....
400
600
800
1000
1200
50
----1---
1400
]
~
.~
As-forgedor
decarburized
u.
200
250
200
1600
-...
1800
.;;;
u.
25
2000
Application of small-scale laboratory fatigue testing to the analysis of components or assemblies introduces additional variables. One is the effect of surface condition. The curves in the top
curve above demonstrate that the fatigue strength of steel specimens varies markedly, depending
on whether the surface is polished, machined, hot rolled, or as-forged. The steel tested was an
unidentified wrought low-alloy steel heat treated to 269 to 285 HB, equivalent to a tensile
strength of 876 MPa (127 ksi) and a yield strength of 696 MPa (101 ksi). Sample preparation
required that the specimens be machined and polished after heat treatment and that rolling or
forging precede heat treatment. For a fatigue life of one million cycles, the fatigue limit was 393
MPa (57 ksi) for the ground specimens, 317 MPa (46 ksi) for the machined specimens, 207 MPa
(30 ksi) for the as-rolled specimens, and only 152 MPa (22 ksi) for the as-forged specimens.
The curves in the bottom graph apply to steels with tensile strength ranging from 345 to 2070
MPa (50 to 300 ksi) and are approximations from several independent investigations. Sample
preparation for "as-forged or decarburized" specimens at the 965 MPa (140 ksi) tensile-strength
level include 4140-type steels rough machined from bar stock, heated to approximately 900C
(1650 OF) in a gas-fired muffle for 20 to 30 min, very lightly swaged from an original 7,47-mm
(0.294-in.) diameter to a final diameter of 7.16 mm (0.282 in.), and air cooled. Heat treatment
consisted of austenitizing in a salt bath at approximately 830C (1525 OF)for 45 min, oil quenching, tempering in air for 1 h at approximately 620C (1150 OF), and water quenching. Forging
and heat treating produced a surface decarburized to a depth of about 0.06 mm (0.0025 in.).
These specimens exhibited a fatigue strength, at 106 cycles, of about 310 MPa (45 ksi), compared
with 470 MPa (68 ksi) for samples that were not forged but were machined or polished and free
of decarburization. Decarburization lowers the strength levels obtained by heat treatment.
Park OH, 1978, P 355
21-1. P/M: Relation of Density to Fatigue Limit and Fatigue Ratio
'" 50
~ (345)
tJl
40
0.. (276)
o
g 30
_ (207)
'E
:.:::i
20
~ (138)
Cl
10
(69)
.50
0
tx:
Q)
~
.45
.40
Cl
u..
.35
6.4
6.6
6.8
7.0
Density, gr/cucm
7.2
The relationship of fatigue strength to density is shown above.

Fatigue strength is best at high densities. For similar P / M and
wrought parts, the ultimate tensile strength to fatigue strength
ratios are the same. However, fatigue strengths of P/M parts
generally are more stable and uniform than for wrought parts.
Parts containing nickel show improved fatigue resistance compared to iron-carbon steels, and high-density nickel steel parts
can be case hardened to improve wear and fatigue properties.
Source: Kurt H. Miska, "Powder Metal Parts," in Source Book on Powder Metallurgy, Samuel Bradbury, Ed., American Society
469
470
21-2. P/M: Relation of Fatigue Limit to Tensile Strength for

Sintered Steels
N/mm
150
00
100
o
o
200
400
Tensile strength)
500
600
N / mm
Fatigue limit of different sintered steels as a function of tensile strength.

Triangles are values for materials without phosphorus; open circles correspond to PNC materials, closed circles to PASC materials.
Source: Per Lindskog, "The Effect of Phosphorus Additions on the Tensile, Fatigue, and Impact Strength ofSintered Steels Based
on the Sponge Iron Powder and High-Purity Atomized Iron Powder," in Source Book on Powder Metallurgy, Samuel Bradbury,
Ed., American Society for Metals, Metals Park OH. 1979, P 46
21-3. P/M (Nickel Steels): As-Sintered vs Quenched and Tempered

for Cycles to Failure
40
0,1 Quenched and Tempered

_ Tensile Strength 105,000 psi
As - Sintered, Tensile
Strength 67,OOOpsi
10
106
10
Cycles to Fuilure , N.
SON diagrams representing fatigue behavior of7.0 g/ cm 3 density,
4 Ni-0.48 C steels, and effect of quenching and tempering on tensile and fatigue strength.
One of the characteristics of the fatigue behavior of wrought

steels is that the S- N curve usually shows a distinct fatigue limit.
This is most marked in wrought plain carbon steels and usually
occurs between 105 and 107 cycles. A typical SoN curve for an
as-sintered nickel steel is shown above. As-sintered nickel steels
possess distinct fatigue limits occurring between 106 and 108
cycles.
Source: A. F. Kravic and D. L. Pasquine, "Fatigue Properties of Sintered Nickel Steels," in Source Book on Powder Metallurgy,
Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979, P 28
471
472
21-4. P/M (Nickel Steels): Relation Between Fatigue Limit and

Tensile Strength for Sintered Steels
60
r---,...-----,---,----...------,---,---....,.--=__--,
LEGEND
50
Smooth As-Sinfered
Smooth Quenched
Tempered
.... 2.2 Kt Notch As-Sinlered

t:J. 2.2 Kt Notch Ouenched-B Tempered
Ui
40
(L
o
o
o
I
~
E 30
--'
.,
:::J
0>
" 20
LL
10
20
40
60
80
100
120
t40
160
Tensile Strength-IOOO PSI
Relation between fatigue limit and tensile strength (fatigue ratio) of sintered nickel
steels.
A plot ofthe fatigue ratio (above) indicates an average smooth value of 0.4 up to
150,000 psi tensile strength. Thus the averagefatigue ratio for sintered nickel steel is
0.4 which is apparently independent of density level, alloy content, and state of heat
treatment and therefore can be used to predict the fatigue behavior of other sintered
nickel steels.
Source: A. F. Kravic and D. L. Pasquine, "Fatigue Properties ofSintered Nickel Steels," in Source Book on Powder Metallurgy,
Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979, P 30
21-5. P1M (Nickel Steels): Effect of Notches on Cycles to Failure for

the As-Sintered Condition
50 r--"-'-"'-''"T'"l"'T'T''r---''-'-'''-''"T'''l''T''Mr---r--''T''""T''"T'''''''''TT''I---'''''
40
Smooth
2.2 Kt Notched
10
106
107
Cycles to Fa.ilure IN.
SoN curves for 0.48% carbon-4.0% nickel alloy steel in the as-sintered condition (7.0
g/ em! density). The two curves demonstrate the effect of a notch on fatigue
characteristics.
Samuel Bradbury, Ed., American Society for Metals, Metals Park DB, 1979, P 33
473
474
21-6. P1M (Nickel Steels): Effect of Notches on Cycles to Failure for

the Quenched and Tempered Condition
0-0-'in
0--
a.
o 30
o
o
2.2 Kt Notched
If)
20
+
If)
10
106
107
Cycles to Failure, N.
SN curves for 0.48% carbon-4.0% nickel alloy steel in the quenched and tempered
condition (7.0 g/cm! density). The two curves demonstrate the effect of a notch on
fatigue characteristics.
Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1979,P 34
21-7. P1M (Low-Carbon, 1-5%Cu): Effects of Notches and Nitriding

40
'"
a..
'"
'"
e
l""- I'--.
Nilrided
Nitriderl
30
=
=
=-
en
--.........
r-
to-.
Not nitrided
r--
20
r---..r--
10
lQ5
Smooth (K=1)
Notched (K=2)
I I III
I IIII
Not nitrided
10 8 lQ5
Cycles to Failure
S-N curves for sintered powders (low-carbon; 1 to 5% copper, 7.1 g/cm 3 density). As
shown above, notches greatly lower fatigue strength, particularly of those that were
not nitrided.
Source: "Nitriding Improves Fatigue Resistance of P 1M Parts," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 292
475
476
21-8. P1M (Sintered Iron, Low-Carbon, No Copper): Effect of Density

and Nitriding on Cycles to Failure
<;
I I III I I II I III r
30
Density - 6.4 g per eu em

"iii
D.-
I II
Density - 7.0 g per eu em
Nitrided
20
r-..... r--...
0
0_
""" t--.....
<11<II
en 10
Not nilrided
Nilrided
Not nitrided
11
10 6
10 8 !OS
10 6
eye les to Fa ilure

SoN curves for P/M parts. As shown above, the greater the density, the higher the
fatigue strength of sintered iron powder (low carbon; no copper; notched; K = 2).
Nitriding in a salt bath is especially beneficial, it will be noted. Bath temperature was
565C (1050 OF); nitriding time was two hours.
Source: "Nitriding Improves Fatigue Resistance of PI M Parts." in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, P 292
21-9. P/M: Effect of Nitriding on Ductile Iron and Sintered Iron

(3%Cu) for Cycles to Failure
60
Ductile Iron
I"":iZ ~
~'''':': :::
'Vi
a.. 40
=
=. ~
=
Sintered Iron
::fit k:t
Nitrided
.:.:.:.:.:.: ;:::;:::::; :::;::;::
;:;:; ::: :: :.:.:.:.:.:.
::::; :::::'~
00
::::::;:;::: :.:.:.:.:.:
~:::::: ;~~~;~~~~~~ U~
Not nihided
~ ~ ~ mmmm ~~t~r ~~r m= ~
~'" .....:....
_:
20
10 6
10 7
00
10 8 10 5
.....':':":
~ ::::;:
Nilrided
:;:;:;:;:::::
::: :~ .........
:.:~:~:~:~:~: ~:~:~ :~:~ ~:
m
it ~ : mIt ~~tt :~t~ ~~~~
"""" <' t
106
Not nilrided
::: ~~
~~~~~~~~~~~~~ ~~~~~~~~~ ~~~~;~ .':'
:::::::::
t~~ jtm~~ :::::;:;: ~ ~ ~: ~m ~
10 7
108
Cycles to Failure
Left: Effect of salt bath nitriding on fatigue strength of ductile iron. Right: Effect ofsalt
bath nitriding on P 1M parts. Specimens were made from a 3% copper sintered iron
ranging from 6.2 to 7.0 g/cm l in density. All specimens were unnotched and were
heated in a nitriding salt at 565C (1050 OF) for two hours.
Source: "Nitriding Improves Fatigue Resistance ofP I M Parts," in Source Book on Nitriding, American Society for Metals, Metals
Park OH, 1977, p 291
477
478
22-1. Brass/Mild Steel Composite: Comparison of Brass-Clad Mild

Steel With Brass and Mild Steel for Cycles to Failure
r----------r-------,.---------,15
20
125
Brass clad mild sleel
a Mild sleel
o Brass
~0~4;:---------:<----------,:-:':T------~IOV
5
6
10
Cycles 10 failure,
10
log scale
SoN curves for composite of mild steel clad (by the explosion technique) with
brass.
Initially, the composite has greater fatigue strength than either brass or
mild steel alone, but above about 106 cycles the values for the composite drop
to about that of mild steel but still remain substantially higher than for brass
alone.
Source: S. K. Banerjee and B. Crossland, "Mechanical Properties of Explosively-CJadded Plates," in Source Book on Innovative
Welding Processes, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, P 148
22-2. Stainless Steel/Mild Steel Composite: Comparison of

Stainless-Clad Mild Steel With Stainless Steel and Mild Steel for
Cycles to Failure
...-----------r----------,---------,15
20
staintess clad mild steel

steet
0 Mild sleel
0____~
.
' Slainl e s s
2 /5
-e
12-5
'"
0
o~~-'====t
10 '1:
.e
0_
OJ'
"l::I
75 ~
~
5
'"
:::
~
/05
Cycles 10 failure,
/0 6
log scale.
SoN curves for composite of mild steel clad (by the explosion technique) with
austenitic stainless steel. Here it is seen that fatigue characteristics of the composite are nearly the same as for stainless steel, and substantially higher than the
fatigue strength of the unclad mild steel.
Source: S. K. Banerjee and B. Crossland, "Mechanical Properties of Explosively-Cladded Plates," in Source Book on Innovative
Welding Processes, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, P 148
479
480
23-1. Carbon and Alloy Steels (Seven Grades): Effects of

Nitrocarburizing on Fatigue Strength
~ Normalized
c::J
Normalized and treated in cyanide-base salt bath (treatment 1).

90 mm (3.5 in.), 570C (1060 of), water quenched
600,------------------------,
~
500 1 - - - - - - -
:2
s: 400
a,
c
1-------
300
Cl>
.~ 200
'"
u,
100
SAE 1000
SAE 1015
SAE 1035
SAE 1045
SAE 1060
600
'"
500
0-
:2
-S' 400
'"
300
Cl>
::l
.g 200
'"
u,
100
0
SAE 1035
SAE 5134
Nitralloy
Bar charts showing increases in fatigue limit that may be obtained by nitrocarburizing (gas or liquid processes).
The amount of improvement in fatigue strength of nitrocarburized materials, as determined with unnotched Wohler test
specimens, depends on the hardness and depth of the diffusion
zone. The potential for improvement in fatigue strength lessens
with increasing carbon and alloy content.
Source: Metals Handbook, 9th Edition, Volume 4, Heat Treating, American Society for Metals, Metals Park OH, 1981, p 269
23-2. Carbon and Alloy Steels (Seven Grades): Effects of

Tufftriding on Fatigue Characteristics
481
Normalized and tufflrided,
Tufftrided, 90 min, 1050 F, water-quenched
90 min, 1050 F, water-quenched

"" 1---, ~ I------'r---.....:..;'--T''''''-=..:~..:...;.;.'''''''':....:l.::..:..;,;.:..;;;.;.'__,
~
~
80
"is~
en
60 J----,---------t
.".
t-----t
+ I---l~-~----t
'"
c,
g 401-----1
=-
20 1-----11_
5134
SAE
Basic fatigue characteristics are directly related to carbon content, as indicated in the
above bar charts for carbon and alloy steels (unnotched test bars). Tufftriding these steels
shows results which prove that fatigue strength increases inversely with carbon content;
that is, the lower the carbon, the greater percentage increase in fatigue strength by
Tufftriding.
Source: Edward Taylor, "Tufftride: Only Skin Deep"," in Source Book on Nitriding, American Society for Metals, Metals Park
OH, 1977, P 280
482
23-3. Carbon and Alloy Steels (Six Grades): Effects of Nitriding on

Fatigue Strength
60
After Atmoaphere NitridJng
After Hardening
111111/
I
I
55
....
III
.c
..10:
11111
NltraUoy
1015
1141
8.620
4620
4340
Il5
Atmosphere nitriding studies showing the interrelationships of steel composition and nitriding by the gas process, and the effect on fatigue strength from nitriding.
Park OH, 1977, P 286
23-4. Carbon-Manganese Steel: Effects of Nickel Coating on

Fatigue Strength
Thickness, mils
0.4
340
320
:2
300
...en
c
~
'"
Q)
280
''::;
'"
0.8
1,2
1.6
45
'iii
.:.l
-,
J::."
-,
'"
u,
260
1
c
1;;
Q)
<,
::s
en
40
r-, .....
'fJ
u.
240
10
20
3D
40
35
50
Thickness, /.1m
Effect of coating thickness on the fatigue strength of a carbonmanganese steel.
The reduction in fatigue strength produced by electro less

nickel deposits is affected by the thickness of the coating.
Thicker deposits have the greatest effect on fatigue strength,
This is illustrated in the above graph, which shows the reduction
in strength of a carbon-manganese steel (Werkstoff St52) produced by different thicknesses of a 5% boron-nickel.
Source: Metals Handbook, 9th Edition, Volume 5, Surface Cleaning, Finishing, and Coating, American Society for Metals, Metals
Park OH, 1982, P 232
483
484
24-1. Coil Springs, Music Wire (Six Sizes): Data Presented

Means of a Goodman Diagram
by
40
200 1-----,,(---+----+----+----+--_=
20
o
Minimum stress,MPa
Wire
diam
Spring
on
Spring
No.
mm
in.
mm
1
2
3
4
5
6
0.81
0.81
1.22
2.59
3.07
4.50
0.032
0.032
0.048
0.102
0.121
0.177
9.52
6.35
15.88
22.22
22.22
22.22
Free
length
in.
Spring
index
mm
in.
Total
turns
Active
tums
Total
tested
0.375
0.250
0.625
0.875
0.875
0.875
10.7
6.8
12.0
7.6
6.2
4.9
22.10
26.97
44.45
60.20
57.15
57.15
0.87
1.062
1.75
2.37
2.25
2.25
6.0
7.0
7.0
7.0
7.5
7.5
4.2
5.2
5.2
5.2
5.7
5.7
16
28
38
43
35
25
Data are average fatigue Iimita from S-Ncurv.. for 185 unpeened springs of various wire diameters run to 10 million cycles oCetrese.Allstresees were
corrected for curvature using the Wahl correction factor. The springs were automatically coiled, with one tum squared on each end, then baked at 260'C
(500 'F) for 1 h, after which the ends were ground perpendicular to the spring axis. The test load was applied statically to each spring and a check made for
set three times before fatigue testing. The springs were all tested in groups of six on the same fatigue testing machine at ten cycles per second. After
testing. the unbroken springs were again checked for set and recorded. Number 4 springs, tested at 1070 MPa (166 ksi) max strese, had undergone about
2,",%set after 10 million strese cycl.., but the stresees were not recalculated to take this into account. None of the other springs showed appreciable set.
The tensile strengths of the wires were according to ASTM A228.
By means of the Goodman diagram many fatigue-limit test results can be shown on the same
diagram as indicated above. In this diagram, line OM represents the minimum stress for the
cycle; the plotted points represent fatigue limits for the respective minimum stresses used. The
vertical distances between these points and the minimum stress reference line represent the stress
ranges. Some scatter may be expected, at least partly attributed to normal changes of tensile
strength with wire diameter. Line UT is usually drawn to intersect line OM at the average ultimate shear strength of the various sizes of wire.
Source: Metals Handbook, 9th Edition. Volume I, Properties and Selection: Irons and Steels. American Society for Metals. Metals
Park OH. 1978, P 293
24-2. Coil Springs: S-N Data for Oil-Tempered and

Music Wire Grades
1 5 0 0 , - - - - - - - - - - - - - - , - - - - - - --r-r-- - - - - - - - - ,
200
1250t---------+--------+------------j
175
~
~
::i 1000f-------""....,,-=-::l-::--------+f
00
---l150
125
750t---------+----.::::!!oo""""'::-l----------l
500L ;--104
- - - - - '-;-- - - - - - --'-::-- - - - - - ___==!

105
106
la)
Type of wire
Number
of
springs
Oil tempered
Music wire
Avg
~
104
105

(b)
(a) Springs were made of minimum quality music wire 0.59 mm (0.022 in.) in diameter. Spring
diameter was 5.21 mm (0.205 in.);D/dwas 8.32. Minimum stress was zero. Stresses corrected
by Wahl factor. (b) Life of springs used in a hydraulic transmission. They were made of
oil-tempered wire (ASTM A229) and music wire (A228). Wire diameter was 4.75 mm (0.187
in.), outside diameter of spring was 44.45 mm (1.750 ln.), with 15 active coils in each spring.
The springs were fatigue tested in a fixture at a stress of 605 MPa (88 ksi), corrected by the
Wahl factor.
The upper graph is a typical S- N diagram showing results of compression testing coil
springs, where the minimum stress is zero while maximum stress is shown by points on the
chart (see spring and testing details given in caption). The lower graph shows an alternate
method of presenting fatigue data for steel springs.
Park OH, 1978, p 292
485
486
24-3. Coil Springs: Effects of Shot Peening on Cycles to Failure

100
90
80
<:
e
"
70
'in
<:
111
60
50 ~
40
~~ r-...
<,~tressed
-----
/peened
in bending
""-Not peened
I-
/Peened
~Stressed
in torsion
'\:"Not peened
Shot peening is often used to improve fatigue strength of springs by prestressing the surface in compression as indicated in the chart above. Shot
peening can be applied to wire 1.6 mm (1/16 in.) or more in diameter, and
slightly smaller wire using special techniques. The kind of shot used is important; better results are obtained with carefully graded shot having only a few
broken, angular particles. Shot size may be optimum at roughly 20% of the
wire diameter. However, for larger wire, it has been found that excessive
roughening during peening with coarse shot lessens the benefits of peening,
apparently by causing minute fissures. Also, peening too deeply leaves little
material in residual tension in the core; this negates the beneficial effect of
peening, which requires internal tensile stress to balance the surface
compression.
Shot peening is effective in largely overcoming the stress-raising effects of
shallow pits and seams. Proper peening intensity is an important factor, but
more important is the need for both the inside and outside surfaces of the
spring to be thoroughly covered. An Almen test strip necessarily receives the
same exposure as the outside of the spring, but to reach the inside, the shot
must pass between the coils and is thereby much restricted. Thus, for springs
with closely spaced coils, a coverage of 400% on the outside may be required
to achieve 90% coverage on the inside.
Cold wound steel springs normally are stress relieved after peening to restore the yield point. A temperature of 230C (450 OF) is common because
higher temperatures degrade or eliminate the improvement in fatigue
strength.
The extent of improvement in fatigue strength to be gained by shot peening, according to one prominent manufacturer of cold wound springs, is
shown in the above diagram.
Park OH, 1978, P 297
24-4. Coil Springs, 8650 and 8660 Steels: Relation of Design

Stresses and Probability of Failure
800
9J", p'~b'blllty of f.i1u,~
100
<,
600
-.......
.
<,
o Not peened
500
r-.. l"-
Peened
110
100
/Peened
sn
BO
..---- fl--Not peened
400
~-
10
I::
60
50
106
BO0
50% p,oti.bility 01 f.i1~,e
I 10
100
100
.....
",0
600
<,
500
r-,
400
90
r--
r---- r-rr---- r-----
BO
10
I-
60
50
106
Number of cycles
(0
failure
BOO
J", p,lb.b!lIty 011,"u,l

100
600
500
400
0""
110
100
90
:ii
~~
80
"" f'--.. I'-- tr--o r--..

0
.........
10
.
o'
il
60
60
106
Number01 cycles to failure
Design stresses. Springs were made from 15.9 to 27.0-mm (% to 1-1/16-in.) diam 8650 and
8660 hot rolled steel and heat treated to between 429 and 444 HB. Springs were shot peened to
an average arc height of 0.008 in. on the type C almen strip at 90% visual coverage.
The desirability of conservative design in cyclical service is

illustrated in the three charts above, in which the minimum
stress used was low. Such data on springs hot wound from bars
with as-rolled surfaces are limited, and interpretation is therefore difficult. The value of peening, however, is made quite apparent. Pertinent test data are given above.
Park OH. 1978, p 304
487
488
24-5. Coil Springs, HSLA Steels: Effects of Corrosion on

Cycles to Failure
~ 540
:<:
Mean
stress:
637 MPa
(92 ksi)
.; 490 -
-o
::s
~ 440
0&-
0.
(J)
- 64
..>:
:r:
C
'tl
0=, ,
:r:
48
Cycles to Failure
,
,
\
H
111
106
~ 50
.\.
p::
- 57
10 5
104
, \
52 -
.,-l
(/)
(/)
o SUP7
_ SUP7-Nb-V
390
(/)
(/)
Q)
H
+J
- 71
ocj-
stress: 490 340 t~Pa

(71 50 ksi)
- 78
e:tJ-
o SUP7
SUP7-Nb-V
I
I
I
5x104 10 5 2x10 5
2x10 4
Cycles to Failure
Fatigue life of coil springs: (left) not corroded and (right) corroded,
Compositions of HSLA Springs Tested
SUP7
SUP7-Nb-V "
Si
Mn
Cu
Cr
AI
0.58
0.56
2.09
1.94
0.83
0.79
0.014
0.014
0.008
0.008
0.09
0.09
0.14
0.09
0.025
0.021
Nb
0.15
0.18
Fatigue tests on coil springs at a hardness of 50 HRC were performed to examine the feasibility of S UP7-Nb-V to the actual suspension coil springs. When the coil springs were free of
corrosion, the result was as shown above (left), in which SUP7-Nb-V has comparable fatigue life
to that of SUP7 in any stress amplitude. When the coil springs were corroded, on the other hand,
the result was rather different. The corroding condition was as follows: an exposure in a chamber
filled with saltwater mist for 1.08 X 104 s (3 h) and a keeping in the atmosphere for 7.56 X 104 s
(21 h). After 10and 20 cycles ofthe corroding, the coil springs were loaded with the surface stress
of 490 340 MPa (71 50 ksi). The fatigue life of the coil springs subjected to 20 cycles of the
corroding are shown above (right). This time, different from the case in the graph at left, there
appears a remarkable difference between SUP7 and SUP7-Nb-V. Measurement ofthe surface
corrosion depth of the two steels showed no difference.
Source: Toshiro Yamamoto, Ryohei Kobayashi, Toshio Ozone and Mamoru Kurimoto, "Precipitation Strengthened Spring Steel
for Automotive Suspensions,"in HSLA Steels-Technology & Applications, American Society for Metals, Metals Park OH, 1984,
P 1022
24-6. Leaf Springs, 5160 Steel: Maximum Applied Stress vs

Cycles to Failure
2070/ 300
250
U)
U)
ILl
lX:
E-<
U) '.-1
1380/200
III
Q::<:
ILl <,
>-<<1l
...:Il:lo
l:lo;:;:
c,
><
150
. ."'.-... . ...... . . -----....-...

b----_
- - a
"
..... "
,.
'.
......d
,,
-.
690/100
~
~ u
'.
.,
f
"
104
-1070
-950
-565
-207
-69
,
345/5
-------
PEAK
RESIDUAL
STRESS, MPa
105
..
106
_ . -
+400
107
CYCLES
Residual stress and unidirectional bending fatigue data for strain-peened SAE
5160 steel. Applied strain during peening curve a, -0.60%; curve b, -0.30%;
curve c, 0% (conventional peening); curve d, preset only; curve e, +0.30%; and
curve f, +0.60%.
Leaf spring specimens of SAE 5I60 steel quenched and tempered to 48

HRC were shot peened under various conditions of applied strain to introduce a wide range of residual stress; then the S-N curves (see above) were
obtained from the same samples by testing in unidirectional bending. In this
illustration, the endurance limit corresponding to the specimen strain-peened
to produce a residual stress of -565 MPa (-82 ksi) will be used to develop a
stress-free ASR diagram for 5160 spring steel (48 HRC). This stress-free
ASR diagram will be used to predict the endurance limit for the other specimens containing peak stress values of - 1070 (- I55), -950 (- 138), -207
(-30), -69 (- 10), and +400 MPa (+58 ksi). The predicted endurance limit
will be compared with the values determined experimentally.
Source: V. K. Sharma and D. H. Breen, "Some Aspects of Incorporating Residual Stresses in Gear Design." in Residual Stress for
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 82
489
490
24-7. Front Suspension Torsion Bar Springs, 5160H Steel:

Distribution of Fatigue Results for Simulated Service Testing
40,-------------------------------.....,
25 lots, 300 parts
'"
t:
30r--------~A_r/hr7'T_V.l+_-------------____l
'0
.8
20
E
:l
101------r;..."...,1'A -V/WA-f/.HV/l-V.l1-V'A-t/'J--Y./lH/.,-+-r.,1'A - - - - - - - - _ _ _ _ l
Service life, 1000 cycles
Here are results from simulated service fatigue tests offront suspension torsion bar springs of
5160H steel. Size of hexagonal bar section was 32 mm (1.25 in.), Mean service life, 134,000
cycles; standard deviation, 37,000 cycles; coefficient of variations, 0.28. It must always be
considered that results from actual or simulated service testing are likely to vary considerably
from results of laboratory testing as shown above.
Park OH, 1978, P 677
24-8. Gears, Carburized Low-Carbon Steel: Relation of Life Factor to

Required Life
5.0
4.0
II
3.0
u-' 2.0
....
.3
u
'"
u..
Q)
...J
1.0
0.5
10 4
106
10 8
Required Life in Cycles
The life factor depends on the required life in cycles. For a single mesh the number of revolutions and the number of cycles are equal. For a gear which has more than one mating member,
the life must be equal to the required number of revolutions multiplied by the number of mating
gears.
Source: "Bending and Contact Stresses in Hypoid Gear Teeth," in Source Book on Gear Design, Technology and Performance,
491
492
24-9. Gears, Carburized Low-Carbon Steel: Bending Stress vs

Cycles to Failure
(/)lI'l
(/)0
LLl ......
0:::
E-><
(/)
,......
~ '~
6 . 90/ 1. 0
...... 0..
0 .........
Zc;l
LLlo..
a:l::'<:
'-'
3.45/.5
CYCLES
Bending fatigue design curves for carburized gears having different amounts of
circumferential residual stress at the root-fillet surface,
The bending fatigue design curves for case-carburized gears with the circumferential root-fillet residual stress varying from +138 MPa (20 ksi) to
-690 MPa (100 ksi) are shown above. From these curves the residual stress
factors at various life cycles were calculated as the ratio of the allowable
bending stress for gears with -483 MPa (70 ksi) residual compression to the
allowable stress for gears with + 138 (20), 0, -276 (40), and -690 MPa (100
ksi) residual stresses.
Source: V. K. Sharma and D. H. Breen, "Some Aspects of Incorporating Residual Stresses in Gear Design," in Residual Stress for
24-10. Gears, Carburized Low-Carbon Steel: Effect of Shot Peening

L50
L50
L50
3. 45/ 5L....::--
'-;-
----'----;;-
----'
105
CYCLES
'"
....
0
1.2
....
Vl
Vl
1.1
....
'"
Vl
...J
:::>
1.0
SHOT CLEANED
Vl
'"
0.9
o. 8 '--.......4- -L..,,----'----;-----..L-:;--.....L.-:;---I
5
7
6
10
10
10
10
10
CYCLES
Top: Allowable (LSO) bending stress design curves for as-carburized,
shot cleaned, and shot peened carburized gears. Bottom: Residual stress
factor computed from the upper chart. K a for carburized shot cleaned
gears equals 1; that is, the allowable SoN curve for carburized shot
cleaned gears is used for design purposes.
Ka Based on Dynamometer Tests

The dynamometer test data obtained on testing sets of gears with different magnitudes of
residual stresses can be used to develop the S- N curves necessary to calculate the Ka factor. The
L50 design curves for as-carburized, shot cleaned, and shot peened gears are shown in the upper
chart. The data for as-carburized and shot cleaned gears were obtained on testing six-pitch test
pinions on a Four Square Dynamometer. The S-N curve for shot peened gears is derived from
the results published by Alman and Black. The lower chart shows the residual stress factors
calculated from the S-N curves in the upper chart. It is assumed that the S-N curve for shot
cleaned gears is used for design purposes; that is, Ka for shot cleaned gears equals one. According to these results the effective bending stress for shot cleaned gears at 108 cycles is approximately 20% higher than as-carburized gears and approximately 15% lower than shot peened
gears. The value of Ka deviates from unity with increasing cycles, indicating a more significant
effect of residual stress at higher life cycles. At low cycles, the residual stress factor seems to
approach one, which means the residual stress has almost no influence on the fatigue properties
of a material at high loads. This is obviously because ofthe stress relaxation caused by the cyclic
plastic deformation accompanying low-cycle fatigue.
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, pp 77, 78
493
494
24-11. Gears, Carburized Low-Carbon Steel: Probability-Stress-Life

Design Curves
U'l
U) ......
U)
13.80/2.0
L90
Jl.lX
c:<:
E- .......
10.35/1.5
U) 'M
L50
Ul
t:lp..
z .......
'"
6.90/1.0
......
00..
L10
z ......
Jl.l:'<:
!Xl '-'
3. 45/ . 5 L..5
10
...I...-
.1....-
--'
10
CYCLES
Probability-stress-life design curves for shot cleaned carburized steel gears having a root-fillet surface finish of 5Jl in.; l.e., KJl = 1.
Bending Stresses in Gears

In designing gears for a new application, a designer usually begins with a
preliminary selection of the tooth widths and other design parameters based
on past practices and empirical approaches recommended by AGMA. The
applied root-fillet bending stress is then calculated to predict the gear life
from the stress-life design curves such as those shown above. The procedure
is reiterated to optimize the design so that the calculated life is just equal to
the required life with an appropriate level of safety. The allowable stress-life
diagram characteristic for each material, heat treatment, and surface treatment is usually obtained on testing acceptable commercial quality gears on a
dynamometer.
Designers and Metallurgists, Larry J. Vande Walle, Ed., American Society for Metals, Metals Park OH, 1981, p 74
24-12. Gears, 8620H Carburized: Bending or Contact Stress vs

Cycles to Fracture or Pitting
500
BENDING
400
(St)
OR
CONTACT
GlO~
~}s,
300
(Se)
STRESS,
KSI
200
G5~
G90
~~
_-}st
10 6
10 7
CYCLES TO FRACTURE OR PITTING
S-N curves showing the wide difference in cycles to failure between bending and contact stress.
Source; D. H. Breen, "Fundamental Aspects of Gear Strength Requirements," in Source Book on Gear Design, Technology and
Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 66
495
496
24-13. Gears, 8620H Carburized: A Weibull Analysis of

Bending Fatigue Data
5,600
.c
..J
c::
- 5,200
Q>
::I
eo
1-c:
4,800
c:
0...
4,400
10 5
10 6
Cycles
Weibull analysis of bending fatigue data from gear tests

indicates that gears made from either the experimental
CH steel or 8620H have equivalent durability.
Metallurgical data gathered on these gears established the

adequacy of the experimental steel (a CRAT steel-computer
harmonizing by application tailoring). Although the experimental steel had a significantly lower case hardenability, it
quenched out to a 100% martensite plus austenitic structure at
the root-fillet surface. Obviously, it had adequate, though not
excessive, case hardenability, thus representing an efficient use
of alloy hardenability in CRAT steels.
Source: G. H. Walter, "Computer Oriented Gear Steel Design Procedure," in Source Book on Gear Design, Technology and
24-14. Gears, 8620H Carburized: T-N Curves for Six-Pinion,

Four-Square Tests
8
7
M
)(
G90
G50
GIO
'" 5
..0
c:
::::l
CI::
aI-
BENDING
FRACTURE
----
SPAlLlNG
: : : : - - - FATIGUE
~1~~----MIXED
..............
PITTING
2L--
-L..
'-........
'-........G90
1.. . . . .
....
'-.....: G50
GIO
----I
-.l
105
CYCLES
T-N curves for carburized six-pitch pinion, four-square gear tests.
The above fatigue data show torque versus cycles to breakage, pitting and
spalling for a six-pitch pinion test. Note that there is a mixed area where
failure can occur from anyone or a mixture of the three modes.
Source: D. H. Breen, "Fundamental Aspects of Gear Strength Requirements," in Source Book on Gear Design, Technology and
497
24-15. Hypoid Gears, 8620H Carburized: Minimum Confidence

Level; Stress vs Cycles to Rupture
498
C1
Ul 105
0:
....Q.
ai
...J
or
.!!!,.
"
..
.. .. . .
....... .--_ev
.. .. . ..
...
......
<,
<,
.....
.. .
<, ~......
.......
a
0:
fill...
1-..........:
.....
Iii
105
CYCLES
FOR
"
10 5
RUPTURE - (N)
<,
107
Fatigue life data for hypoid gears. Sloping line indicates minimum confidence level.
Source: "Gleason Method for Estimating the Fatigue Life of Bevel Gears and Hypoid Gears." in Source Book on Gear Design,
Technology and Performance, Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, P 386
24-16. Hypoid, Zero I and Spiral Bevel Gears, 8620H Carburized:

S-N Scatter Band and Minimum Confidence Level
-------- ----..........-..........
..........
---- 1;,0--..
060
--
---- -- ............
.a"" . ... . ..........
---
co
00
0..ll........
<,
<,
~<QI o
<, ..... .; ~"'-"'.
:--..
<, ~~~
<,
....'
.. ... ----... . . .. --. ..

.. . . . .
~
......
8~
I~
PROPOSED DESIGN LIMIT(NOT OVER 5 PER CENT FAILURES)

1
'l,
.............
"
n
no
lI)
lI)
a:::
l-
ll)
105
106
CYCLES FOR RUPTURE - (N)
Fatigue data-composite for results obtained by testing various gear designs,
Source: "Gleason Method for Estimating the Fatigue Life of Bevel Gears and Hypoid Gears," in Source Book on Gear Design,
499
500
24-17. Spiral Bevel and Zerol Bevel Gears, 8620H Carburized:

S-N Scatter Band and Minimum Confidence Level
....,..
~
g.
i.
1'---.",-
III
a:
w
CL
ai
.J
~
I
III
:la:
10
.-
<,
.. ..- ..........- .-. ....

... . .'.
<, ...
. ... . . . . . ...- ..
<,
-----...... . .... ,... . . . . . . .
... ..
1'--- ...
<;
,..,..~"
a..
<,
... -.........:
IIII
.........
10 5
C.YCLE:S
10&
fOR
107
RUPTURE -(N)
Graph offatigue life for spiral bevel and Zerol bevel gears. Sloping line indicates the minimum
confidence level.
Source: "Gleason Method for Estimating the Fatigue Life of Bevel Gears and Hypoid Gears," in Source Book on Gear Design,
501
24-18. Gears, 8620H Case Hardened: Relation of Life Factor to

Cycles to Rupture
5
4
--~
3
~
I
5
5%
n0
50%
...
Gl
....
:..:i
1'000.
.......
Gl
......
C
Gl
95%
0.9
0.8
-0
c;:
C
0.7
0.6
10J

.....Gl
105
106
Cycles for Rupture
Both strength and durability are fatigue phenomena and therefore display a relationship between stress and life. The life factor for strength may be obtained from the above
data.
Source: "Bending and Contact Stresses in Hypoid Gear Teeth," in Source Book on Gear Design, Technology and Performance.
Maurice A. H. Howes, Ed., American Society for Metals, Metals Park OH, 1980, p 127
502
24-19. Bevel Gears, Low-Carbon Steel Case Hardened: Relation of

Life Factor to Cycles to Rupture for Various Confidence Levels
5
...
5%
I
l>
50%
0
u,
...::;
po".
I>
.......
I
I
.......
,
I
1
0.9
0.8
95%
I
I
)]
0.6
103
1O~
106
108
10 7
Cycles for Rupture
The life factor is obtained from the graph above. This depends upon the required life in cycles.
For a single mesh the number of revolutions and the number of cycles are the same. For a gear
which has more than one mating member the life must be equal to the required number of
revolutions multiplied by the number of mating gears. When the required life is less than
6,000,000 cycles on the pinion, the life factors will be different on gear and mating pinion.
In cases where the load varies, the designer may wish to determine the equivalent life at maximum torque. One suggested method is as follows:
L cp=60L H
T~
[k lnpl+ k n n ( T)
2
5.68
+k)npJ
(Tf; )
5.68
+ ..
+kl/npl/
T)
i;
5.68 ]
where L cp = required equivalent life in pinion cycles at maximum torque.

L H = required total life in hours.
k),k 2,k)
k,,=proportion of time at torque loads T" Tb T) . . . T"respectively.
n PI' n n. n P)
n PI/= pinion rpm corresponding to torque loads T" T 2, T) . . . TI/
respectively.
T 1, T 2, T) . . . T" = torque loads where T I is maximum torque and T" is minimum
torque which will produce a stress above the endurance limit.
The required equivalent life in gear cycles at maximum torque may be obtained by multiplying
the life in pinion cycles by the gear ratio:
n,
L CG=L cp NG
Source: "Bending Stresses in Bevel Gear Teeth," in Source Book on Gear Design. Technology and Performance. Maurice A. H.
Howes, Ed., American Society for Metals, Metals Park OH, 1980, p 149
503
24-20. Gears, AMS 6265: S-N Data for Cut vs Forged

35
.c
.....J
=
=
=-
30
.6.
",,-
,9
25
""
ll...
.8
"C
,~
20
CL
CL
<I:
"C
co
0
.....J
E
15
6 CuI gears
Forged gears
:::I
E
';(
co
::;;:
10
5
lQ3
104
10 5
10 6
1Q7
Cycles 10 Failure
Fatigue data shown in this chart proved that teeth on precision forged AMS 6265
helicopter pinions have a higher fatigue limit than cut teeth. Loads shown are applied
actuator loads. Tooth loads are approximately 33% greater.
Source: "How Gearmaking Methods Compare," in Source Book on Powder Metallurgy. Samuel Bradbury, Ed" American Society
504
24-21. Spur Gears, 8620H: S-N Data for Cut vs Forged

120
110
100
"u;
a..
90
0
0
0_
80
"0
0
0:::
co 70
e
lJ)
lJ)
en 60
1% failure
m
c
'j
0
z:
50
40
30
- - - Cui gears
Forged gears
10 5
Cycles 10 Failure
S-N curves for spur gears forged and cut from 8620H steel.
As shown above, results of beam fatigue tests indicate that precision flow-forged
gear teeth are about 20% higher in fatigue strength than cut teeth.
Source: "How Gearmaking Methods Compare," in Source Book on Powder Metallurgy, Samuel Bradbury, Ed., American Society
24-22. Gears and Pinions: P/M 4600Vvs4615; Weibull Distributions

LEGEND
~.
4
5.
6
7
8
3
99._
_5.0
SINTERING
TEMP TIME
ALLOY
2~F I MIN
4600V
4600Y
2100-F 3MIN
2350F 3MIN
2350F 6MIN
4600 V
4600V
4600V
4615
2100-F 6 MIN
(BAR STOCK)
80.0
::>
400
300
...J
u.
20.
z
w
a:
w
Q.
60.0
w
a:
'00
8
60
40
20
/
1.0
.1
.2
--(
.3
TIME (HR.)
.-
LEGEND
~.
,
2
3
!!::ill
2000
2000
4615
SINTERING
TEMP TIME
2350F "'6MiN
2350F 3 MIN
(BAR STOCK)
95.0
80.0
600
w
a:
40.0
300
u.
20.0
a:
w
Q.
10.0
80
60
40
2.0
1.0
.1
.2
.8
1.0
34~6810
TIME (HR.)
Top: Weibull distribution charts for fatigue testing of actual gears and
pinions made from 4600V alloy with various sintering times and
temperatures as shown, compared with cut pinions (4615 bar stock).
Bottom: Similar to graph at top exceptfor alloy 2000 and 4615bar stock.
Source: P. C. Eloff and L. E. Wilcox, "Fatigue Behavior of Hot Formed Powder Differential Pinions." in Source Book on Powder
Metallurgy, Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1980, p 308
505
506
24-23. Gears and Pinions: P/M Grades 4600V and 2000 vs 4615;
Percent Failure vs Time
99.9
LEGEND
NO.
950
2
3
80.0
1-1
ALLOY
4600V
2000
4615 8AR STOCK
95'/. CONFIDENCE
LIMITS
40.0
w
0::
60.0
'i
30.0
--l
<l:
20.0
lJ...
IZ
0::
W
(L
10.0
8.0
6.0
4.0
2.0
1.0
.1
5 6
10
TIME (HR.)
Fatigue data for actual gears tested in specially designed machines. Presented
here are Weibull distributions for the three types of alloys tested.
Since the data from the two powder alloys fell into two groups, it was decided to
fit one Weibull curve to all of the data points from each alloy. This was done to
obtain more data points for each curve. The results are shown in the graph above,
which also graphically indicates the 95%confidence limits on the BID lives. It isplain
that the 4600V pinions have superior fatigue life at the stress level of 92,400 psi, and
the slope ofthe Weibull curve indicates uniform deoxidation of preforms and therefore less scatter (steeper Weibull slope) in the fatigue data. In the case ofthe 4600V
alloy, the sintering temperature should have little effect on deoxidation, since the
major alloying constituents, nickel and molybdenum, are readily reducible by CO
at temperatures even below 1150 C (2100 OF).
Source: P. C. Eloff and L. E. Wilcox, "Fatigue Behavior of Hot Formed Powder Differential Pinions." in Source Book on Powder
Metallurgy, Samuel Bradbury, Ed., American Society for Metals, Metals Park OH, 1980, P 313
24-24. Gear Steel AMS 6265: Parent Metal vs Electron Beam Welded
2
100
4 68l
4 681
4 681
4 681
I
900
10
10
i~
50'
SON diagram for AMS 6265-parent metal versus electron beam

welded.
The welded specimens failed in the weld zone at 86% joint

efficiency.
In the weld evaluations made, excellent mechanical properties were found. Other gear materials tested resulted in compara ble weld joint efficiencies. In general, it was demonstrated that
electron beam mechanical properties were comparable or better
than welds made with other fusion welding processes such as gas
tungsten arc and metallic arc welding.
Source: N. F. Bratkovich, W. L. Mcintire and Robert E. Purdy. "Electron Beam Welding-Applications and Design Considerations for Aircraft Turbine Engine Gears.t' in Source Book on Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals, Metals Park OH, 1981, P 199
507
508
24-25. Gears, 42 CrMo4 (German Specification): S-N Curves for

Various Profiles
80
z-,
~. ;.>.
>\0
40
.........~
.~
30
. .......
~
" ~
,....
I~
~:~ ! .,..:.-;:::~ "."!.:'

~~~
-- _..
3.10 4
ll.i
O,I..l,fl
(I ~ (; (I
.::--.
~~.
20
--- -- -
:~~.. :. ;. ........
i\:
.- .
. .-
.. --- ..
O.~o2
0, 'lG r\(W/m
~
1.)C G
-J-L
LW--+
SoN curves for various tooth profiles (50% survival probability in the short life
range).
The Woehler curves shown above are based (in the sloping section) on a
survival probability of 50% at the number of cycles indicated. The horizontal
sections of the curves are based on the highest load that can be carried for a
minimum life of five million cycles.
Source: H. Winter and M. Hirt, "The Measurement of Actual Strains at Gear Teeth, Influence of Fillet Radius on Stresses and
Tooth Strength,"in Source Book on Gear Design, Technology and Performance, Maurice A. H. Howes, Ed .. American Society for
24-26. Gears, 42 CrMo4 (German Specification): Endurance Test

Results in the Wei bull Distribution Diagram
95
I ,
I I
f/
.....
II ~
II'
;-- 70
Of
II
if
50
~ ...
il
71'
If
30
f
,I:
20
J~!'I
10
8
A -1-
l-
-- _..
----f I - - 1-I
6
4
3.,0'< 4
- --.-
8 1.'0:>
. .
t. 5
,-W
(;
. -
1.1
--~
Endurance test results in the Weibull distribution diagram.
The parameters ')I,A are adjusted to the test points. An example is

shown in the graph above.which represents in such a probability grid
the test results for gears of one tooth form. Scale of ordinates is the
failure probability A = 1- W= if (n + I) for test i out of n test results
sorted to the number ofload cycles at which fracture occurred. The test
points are approximated with a straight line. From this curve we are
able to read, respectively, life values L IO , L so, and LJo for 10,50, and 90
percent failure probability, or 90, 50 and 10 percent survival probability.
A more adequate approximation by the theoretical distribution is
achieved by a three-parameter Weibull distribution. This formulation
produces a minimum endurance Lo, which is reached by all test pieces.
Also, the above chart shows the compensating curve which results
from the formulation of the three-parameter Weibull distribution.
Source: H. Winter and M. Hirt, "The Measurement of Actual Strains at Gear Teeth, Influence of Fillet Radius on Stresses and
Tooth Strength,"in Source Book on Gear Design, Technology and Performance, Maurice A. H. Howes, Ed.. American Society for
509
510
24-27. Bolts, 1040 and 4037 Steels: Maximum Bending Stress vs

Number of Stress Cycles
100
0
1040 steel
I
100
95
4037 steel
615
s:
:::iE
~g'
650
"v;
"'".
~
625
0.0
:0
~
)(
600
.n .
Cl
c
90 :0
c
..
.0
)(
:::iE
~o.
of>
~-
~~-
85
515
o
>--1-
The bolts (7'8 by 2 in., 16 threads to the inch) had a hardness of 35 HRC. Tensile properties of
the 1040 steel at three-thread exposure were: yield strength, 1060 MPa (154 ksi); tensile
strength (axial), 1200 MPa (175 ksi); tensile strength (wedge), 1190 MPa (173 ksi). For the
4037 steel: yield strength, 1110 MPa (161 ksi); tensile strength (axial), 1250 MPa (182 ksi);
tensile strength (wedge), 1250 MPa (182 ksi).
In general, if bolts made of two different steels have equivalent hardnesses throughout
identical sections, their fatigue strengths will be similar (see above S- N data), 'as long as
other factors such as mean stress, stress range, and surface condition are the same. If the
results of fatigue tests on standard test specimens were interpreted literally, high-carbon
steels would be selected for bolts. Actually, steels of high carbon content (more than
0.55% carbon) are unsuitable because they are notch sensitive.
The principal design feature of a bolt is the threaded section, which establishes a notch
pattern inherent in the part because of its design. The form of the threads, plus any mechanical or metallurgical condition that also creates a surface notch, is much more important than steel composition in determining the fatigue resistance of a particular lot of
bolts.
Source: ASM Committee on Carbon and Alloy Steels. "Threaded Steel Fasteners." in Quality Control Source Book. A. K. Hingwe,
Ed .. American Society for Metals, Metals Park OH, 1982. P 206
24-28. Bolts: S-N Data for Roll Threading Before and After Heat
Treatment
450 r - - - , - - - - , - - - - - , - - - ,
Roll threaded before
heat treatment
60
375 1 - - - t - - - - - 1 - - - - t - - - - - j
50
3001---t-----1----t-----j
40
225
30 U5
150 f-----11-
20
751----t------I---'
L-
! -_ _-'----_ _
10
~_~
104
450 , - - - , - - - - , - - - - - . - - - ,
60
375
-
50
300
40
'"
:;;:
0-
::i 225 I - - - t - -
1"
U5
150 f - - - - t - - - - t - - - + - - - - - j 20
Roll threaded after

heat treatment
75
1---+---+---+---::::1 10
S-N curves showing fatigue limits for roll-threaded bolts. Upper

graph represents four different lots of bolts that were roll
threaded, then heat treated to average hardness of 22.7, 26.6,
27.6, and 32.6 HRC. Lower graph represents five different lots
that were heat treated to average hardnesses of 23.3, 27.4, 29.6,
31.7, and 33.0 HRC, then roll threaded. Bolts having higher
hardnesses in each category had higher fatigue strengths.
Other factors being equal, a bolt with threads properly rolled

after heat treatment-that is, free from mechanical imperfections-has a higher fatigue limit than one with cut threads. This
is true for any strength category. The cold work of rolling increases the strength at the weakest section (the thread root) and
imparts residual compressive stresses, similar to those imparted
by shot peening.
Source: ASM Committee on Carbon and Alloy Steels, "Threaded Steel Fasteners," in Quality Control Source Book, A. K. Hingwe,
Ed., American Society for Metals, Metals Park OH, 1982, P 202
511
512
24-29. Power Shafts, AMS 6382 and AMS 6260: Electron Beam
Welded vs Silver Brazed Joints
.. ELECTRON BEAM WELDED JOINTS
60
o
o
TEST TEMPERATURE, 500-600 oF
SILVER BRAZED JOINTS
50
III
S/N-2
Z
:l
:I:
..
40
z
+1
w
30
0:
I-
>0:
o
!;i
20
0:
>
--
r-- l I-
.. ----
~~
S/N-2
SIN-I
ROOM TEMP.
i----
r---r--.
..
S/N-3
o0..
....
SIN-I
:---
--. "-
-I---
10
CYCLES TO FAILURE
S-N curves for electron beam welded versus silver brazed power shafts made from AMS 6382 and
6260 alloy steels.
In the welded shafts, failures occurred apart from the weld, while in the brazed units all
failures occurred in the brazed joints.
Source: S. M. Silverstein, V. Strautman and V. R. Freeman, "Application of Electron Beam Welding to Rotating Gas Turbine
Components," in Source Book on Electron Beam and Laser Welding, Melvin M. Schwartz, Ed., American Society for Metals,
Metals Park OH, 1981, p 187
24-30. Axle Shafts, 1046, 1541 and 50854 Steels: S-N Data for
Induction Hardening vs Through Hardening
513
:\, B & C
I1'I1lUCf IO~ HARDENED 1541

1046
E - THROUQI HARDENED 50B54
D-
~on
if.
'""
if.
v: 1 SO
?i
..9."Il!':.? ....
,.....
-,
tr.
".
.... ..........
:~
...
"
.....
u:: ] on
>c:::
<
~
0;
., ............ --
\',
.......
-. -'-."
'-'-'-B
......... ::::.:::-:.:::-::.: ~
"
50
APPLIED
~
STRESS GRWIE~T
200
400
600
DISTA.'JCE FRO~l SURFACES (.001")
A, B & C - rxructrox HARDE\'ED 1541

D 1046
E - TI-JROllGH HARDE\ED 50B46
..... 140
IF.
~
u: 100
u:
..... .......~
.
"
..
.~
'
':-.."
l-
i/;
800
60
.,
-, . "
f:S
u:
20
10 2
]0 3
10 4
"
:"::..,,:,,'~ c
.'-._.-E
----B
lOS
10 6
NO. OF CYCLES TO FAILURE

Top: Axle-shaft strength gradients in terms of shear yield strength. Bottom: Fatigue performances of axle shafts as a function of strength gradient.
Induction hardened 0.40% carbon steel axle shafts were developed to replace through hardened alloy steel
shafts for both product- and cost-improvement purposes. This was accomplished after a rather comprehensive
bench-test program, which examined variables such as surface hardness, core hardness, gradient strength, distortion, composition, and surface-condition effects. The more promising approach was then subjected to chassis, proving-ground, and in-service testing. Some interesting reflections can result from examining some of the
fatigue data that were generated. That the through hardened concept was vulnerable can be surmised by considering the stresses developed in a full-floating splined shaft loaded in torsion. The stresses are a maximum at the
surface and drop linearly to zero at the center. At the spline, the stresses drop more rapidly at the onset due to the
stress concentration caused by the spline. The upper graph shows the stress gradient in the body area when the
shaft is loaded to 110,000-psi shear stress. Also plotted on this graph are the shear yield-strength gradients
(converted from hardness) of the production alloy shaft, along with several experimental induction hardened
shafts. One would expect the through hardened shaft to have a surface-origin failure and to be lower in strength,
since its surface is the lowest hardness. Also, the high strength of the center ofthe shaft is essentially wasted, since
it is lowly stressed. Gradient strengthening by induction hardening provides a means of providing a better
strength match for the stress gradient. The lower graph gives the fatigue curves established for shafts having the
strength gradients shown in the upper graph.
American Society for Metals. Metals Park OH. 1979, P 88
514
24-31. Steel Rollers, 8620H Carburized: Effects of Carburizing

Temperature and Quenching Practice on Surface Fatigue
999
SAE 8b2011 - Reheat
990
Carbo
of
95'0
0 Grou 0
II Group Q
900
Group
800
----
70'0
1750
1800
1900'
Quench
Sliding
..
~I
21
21
Slope
Cor r .
Coef.
.89
.97
.98
2.199
I. 463
2.961
vacuum
------
AJl
600
500
40'0 -
Data corrected to
= 400 ksi
Sc___
____
A= 0.5
30'0
200
~
a:
:>
:;;
.... 10'0
I-
Z
~
a:
~
e,
5'0
4'0
30
"0
Q
1'0
0'5
0'4
0'3
02
105
0"
106
4
6 7 8 9 ,
8
10
7
2
c v c i es
6 7 9 ,
6 7,19
S-N data, Weibull probability plot: Effect of three carburizing temperatures on surface fa-
tigue for carburized 8620" steel. All were slow cooled and reheated for quenching. This
technique improved fatigue characteristics compared with direct-quenched rollers.
Source: S. L Rice. "Pilling Resistance of Some High TemperatureCarburized Cases," in Source Book on Gear Design. Technology
and Performance. Maurice A, H, Howes, Ed" American Society for Metals, Metals Park OH, 1980, p 234
24-32. Steel Rollers, 8620H Carburized: Effects of Carburizing

Temperature and Quenching Practice on Surface Fatigue
99 .
=...,
1-
.'0
SAE 8620H - Direct Quenched

Carbo
OF
9"0
'0' 0 1-0=--Gc..r_o_u....
p_N
Sliding
\
Corr.
Coef.
r-
21
.90
1750
+f. ._ _r----1
.97
21
1800
.97
21
1900'
10.0 ~.u!!J"--"S'---'-"=--=----'-".:.-----=:..:..:.:...:..,~J_------+f
...- r _ - - - - _ j
70.0
Group P
~a"'cull
...umlll-.
_f_.=_"~-----_c,.-_r'--:_---__1
6&:~(--------------_:_
. .+-------__,f.;;;;:"M,./_------_I
'Ool---------------.'f--------r..-f---i~-------__I
4o01---------------;.t----------tl~-'":f_--------__I
Data Corrected to
3001--1>:c-~W9__1<!;i_-----.+------__::~+...,,'-----------_f
A = 0,5
200
. JOOj----------r--='---------if---------------_f
I-
U
0:
.. "0 - - - - - - - - f - : ; . ; o - - - - - - - - - - T - . H........---------------1
"0 t - - - - - r - - - - - - - - + - j ' - - - - - - - - - - - - - - - - - - - _ _ I
N
'-0
D"
0"
D"
03
0'\
,105
345678"
C 't'
c i es
4567."
45.71.
S-N data, Weibull probability factor: Effect ofthe three carburizing temperatures on
surface fatigue for carburized 8620H steel. All were direct quenched from the temperatures shown above.
Source: S. L. Rice, "Pitting Resistance of Some High Temperature Carburized Cases," in Source Book on Gear Design, Technology
and Performance, Maurice A. H. Howes, Ed.. American Society for Metals, Metals Park OH, 1980, P 233
515
516
24-33. Linkage Arm, Cast Low-Carbon Steel: Starting Crack Size vs

Cycles to Failure
III
OJ
s:
u
c .300
E
E
I
7
0
.250
en
:.::
<{
i=
a::
en
<{
.150
:.::
af; 0.63 in. 116 m rn )
a::
u
;:!
iii
5
<.9
Z
(136MPa)
.200
a::
U
<.9
Z
i=
a::
.100
<{
I-
en
.050
1000
10,000
CYCLES TO
100,000
FAILURE. N f -
Fatigue life of a linkage arm as a function of starting crack size.
The variation of the fatigue life, NJ, with the starting flaw size a.; is shown in the
diagram above. The fatigue life increases dramatically at very small a i values. The
far curve shows that in the long life regime the final crack size has only a small effect
on Np This is because fatigue crack growth rates are very low at low ~Kvalues and
hence the greatest fraction offatigue life is spent at the smaller crack sizes. Since the
controlling parameter is ~K, low life for small crack sizes is possible at high cyclic
stresses. The second set of curves shows that doubling the cyclic stress range reduces
the fatigue life by about an order of magnitude. Also, if the starting ~K value is
high, the final crack size has a larger effect on the cyclic life. The above diagram
shows the importance of adjusting both the cyclic stress and starting flaw size to
optimize the fatigue life.
Source: Steel Castings Handbook. 5th Edition, Peter F. Weiser, Ed., Steel Founders' Society of America, Rocky River OH. 1980,
P 4-17
24-34. Notched Links, Hot Rolled Low-Carbon Steel: S-N Data for
Component Test Model
... - Component Tests
I. 0 r - - - - - - - - - - - . : . . . - - - - - - - - - - - ,
~:
~.., _
......
0.7
~~notched
~~~ ~/2Z7!?2Z
.~
0.6
_./Coml
---------_--=--~~~nent Tests
0.5
<, .l
~~
Local-Strain Model
--A
rr .............& ~1
......
04
.
0.3
S-N /
Model
R=-1
~@;/?/!?!Z2Z
0.2L..--------1------.........L~------J
7
loll
lOll
Fatigue Life. Nj cycles
10
COMPONENT TEST MODEL. The most straightforward life prediction model of a component is
developed from fatigue tests of the component itself. The component is cyclically loaded in a manner
that attempts to simulate service and the model is the plot of the test results. The cyclic load or nominal
net-section stress is plotted versus cycles to crack initiation N; or to failure Np
Example: Component fatigue tests were performed on the notched link ofthe previous examples; the
results are listed in the table and shown graphically above. The fatigue strength for crack initiation at
one million cycles is:
iJ.S
= 0.47
Sli
106 cycles
The three models-S-N, local strain, and component tests-are compared above, and the three
predictions for fatigue strength (iJ.S) at one million cycles are, respectively, 28, 42, and 47% of the
tensile strength Sli' The local strain prediction is closer to the behavior observed in the component test
than the S-N prediction.
Component Test Results, Hot-Rolled Carbon Steel Notched Link
Tensile
strength,
Yield
strength,
Cycles to
crack initiation,
Su,MPa
S}",MPa
N;
417
388
366
366
402
417
366
402
388
242
236
239
239
235
242
239
235
236
0.818
0.880
0.936
0.690
0.628
0.604
0.444
0.404
0.418
1.36 X
1.69 X
2.62 X
6.99 X
8.24 X
1.05 X
1.64 X
2.21 X
2.99 X
10'
10'
10'
10'
10'
105
10"
10"
106
Source: Harold S. Reemsnyder, "Constant Amplitude Fatigue Life Assessment Models," in Proceedings of the SAE Fatigue Conference P-I09, Society of Automotive Engineers. Inc .. Warrendale PA, 1982. P 127
517
518
24-35. Fuselage Brace, Ti-6AI-6V-2Sn: Fatigue Endurance of

HIP-Consolidated Powder
.;;;
.:.!
tl' 120
~
:;;
E
::>
E 100
'x
'"
80
60
Annealed plate
(min)
-..;
10'
10'
m
10
10'
Number of cycles
Fatigue endurance of Ti-6AI-6V-2Sn powder consolidated by HIP at 1650 of

(900C). Significance of boxed numbers is as follows: 1: HIP run #1, as machined. 2: HIP run #2, vacuum annealed for 2 h at 1300 of. 3: HIP run #2,
vacuum annealed for 16 h at 1300 F. 4: HIP run #2, vacuum annealed for 24 h
at 1300 OF. 5: HIP run #4, vacuum annealed for 24 hat 1300 OF.
A fuselage brace made from HlP'd Ti-6AI-6V-2Sn powder was used to

establish the flight worthiness of a HlP'd P/M airframe component. The
tensile and toughness properties developed compare well with the average
values for forgings. The fatigue endurance limit of H lP powder developed in
this program is given in the SoN diagram above. Here the HIP data points lie
within the band for annealed forgings and plate.
Source: J. H. Moll. V. C. Petersen and E. J. Dulis, "Powder Metallurgy Parts for Aerospace Applications," in Powder MetallurgyApplications. Advantages and Limitations. Erhard Klar, Ed.. American Society for Metals. Metals Park OH, 1983, P 288

Fatigue

Uploaded by

Document Information

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

Fatigue

Uploaded by

Copyright:

Available Formats

Atlas of

SECTION 1: S-N Curves That Typify Effects of Major Variables

S-NCurves Typical for Steel 27

SECTION 2: Low-Carbon Steels: Flat-Rolled, Weldments and Tubes

Typical S-N Curve for Low-Carbon Steel Under Axial Tension 43

2-26. A514B Steel: Effect of Various Gaseous Environments on Fatigue Crack

SECTION 3: Medium-Carbon Steels, Wrought and Cast

AISI 1030 (Cast) Compared With AISI 1020 (Wrought)

SECTION 4: Alloy Steels: Low- to High-Carbon, Inclusive

Medium-Carbon Alloy Steels, Five Grades: Effect of Martensite Content 83

4625 Steel: P/M vs Ingot Forms

SECTION 5: HSLA Steels

HI-FORM 50 Steel vs 1006 147

SECTION 6: High-Strength Alloy Steels

HY-130 Steel: Effect of Notch Radii

SECTION 7: Heat-Resisting Steels

SECTION 8: Stainless Steels

SECTION 9: Maraging Steels

SECTION 10: Cast Irons

Fatigue of Cast Irons as a Function of Structure-Sensitive Parameters

SECTION 11: Heat-Resisting Alloys

A286: Effect of Environment

Inconel625: Effect of Temperature on Cycles to Failure 286

SECTION 12: Aluminum Alloys

12-1. Corrosion-Fatigue Properties of Aluminum Alloys Compared With Those of

12-10. Alloy 2024-T4: High-Cycle vs Low-Cycle Fatigue

SECTION 13: Copper Alloys

13-1. Copper: Effect of Air and Water Vapor on Cycles to Failure

SECTION 14: Magnesium Alloys

14-1. Magnesium Casting Alloy QE22A-T6: Effects of Notches and Testing

SECTION 15: Molybdenum

15-1. Molybdenum: Fatigue Limit Ratio vs Temperature

SECTION 16: Tin Alloys

16-1. Tin-Lead Soldering Alloy: S-N Data for Soldered Joints

SECTION 17: Titanium and Titanium Alloys

SECTION 18: Zirconium

18-1. Zirconium 702: Effects of Notches and Testing Temperature on Cycles to

SECTION 19: Steel Castings

SECTION 20: Closed-Die Forgings

(See also under specific grades of alloys.)

20-1. Closed-Die Steel Forgings: Effect of Surface Condition on Fatigue Limit

SECTION 21: Powder Metallurgy Parts

(See also under specific alloys.)

SECTION 22: Composites

SECTION 23: Effects of Surface Treatments

SECTION 24: Test Results for Component Parts 484

24-32. Steel Rollers, 8620H Carburized: Effects of Carburizing Temperature and

ence of reversed stresses that exceed the flow

Prediction of Fatigue Life

Fatigue Crack Initiation

The latter is considered a negative tensile stress,

If the stresses are fully reversed, the stress ratio

During a fatigue test, the stress cycle usually is

Fig. 1 Nomenclature to describe test

2340 steel-= 100

Stress ratio (R) = - 1

Number of cycles to fracture, N

ing plot of the data is an S-N curve. Three typical

Fatigue Crack Propagation

L..-_ _.L-_ _..I.--- _ _-L._ _- '

Fatigue Crack Initiation

of these phenomena have a significant influence

Fatigue Testing Regimes

Fatigue Crack Initiation