Canale L.C.F., Mesquita R.A., Totten G.E. Failure Analysis of Heat Treated Steel Components

Подождите немного. Документ загружается.

16. K.E. Thelning, Chapter 6.6: Induction

Hardening, Steel and Its Heat Treating,

Bofors Handbook, Butterworth, London

and Boston, 1975, p 432–451

17. K.E. Thelning, Chapter 6.2: Quenching and

Tempering of Constructional Steels, Steel

and Its Heat Treating, Bofors Handbook,

Butterworth, London and Boston, 1975,

p 319–325

18. P.G. Simpson, Induction Heating—Coil and

System Design, McGraw-Hill, New York,

Toronto, London, 1960

19. V. Rudnev, D. Loveless, R. Cook, and

M. Black, Handbook of Induction Heating,

Marcel Dekker, Inc., New York, Basel,

2003

20. J. Grum, Induction Hardening, Handbook of

Residual Stress and Deformation of Steel,

G.E. Totten, M.A.H. Howes, and T. Inoue,

Ed., ASM International, 2002, p 220–247

21. V.I. Rudnev, R.L. Cook, D.L. Loveless, and

M.R. Black, Induction Heat Treatment,

Basic Principles, Computation, Coil Con-

struction and Design Considerations,

Chapter 11A, Steel Heat Treatment Hand-

book, G.E. Totten and M.A.H. Howes,

Ed., Marcel Dekker, Inc., New York, 1997,

p 765–767

22. V.I. Rudnev, R.L. Cook, D.L. Loveless, and

M.R. Black, Introduction Heat Treatment,

Modern Power Supplies, Load Matching,

Process Control, and Monitoring, Chapter

11B, Steel Heat Treatment Handbook,

G.E. Totten and M.A.H. Howes, Ed.,

Marcel Dekker, Inc., New York, 1997,

p 873–874

23. G. Pfaffmann, Introduction to Induction

Heating, Fundamentals of Induction Heat-

ing, Society of Manufacturing Engineers,

Anaheim, CA, 1998

24. D.J. Williams, Quench Systems for In-

duction Hardening, Met. Heat Treat., 1995

25. K.E. Thelning, Chapter 6.6.3, Equipment

for Induction Hardening, Steel and Its

Heat Treatment, Bofors Handbook, Butter-

worth, London and Boston, 1975, p 435–

437

26. V.K. Rudnev, D.L. Loveless, M.R. Black,

and P.J. Miller, Progress in Study of

Induction Surface Hardening of Carbon

Steels, Gray Irons and Ductile (Nodular)

Irons, Ind. Heat., March 1996

27. M.G. Lozinskii, Industrial Applications

of Induction Heating, Pergamon Press,

Oxford, 1969

28. R.S. Rufﬁni and V. Nemkov, Induction

Heating Systems Improvement by Appli-

cation of Magnetic Flux Controllers, Proc.

of the Int. Induction Heating Seminar,

Padna, 1998, p 133–139

29. C. Durban, D. Durand, and P. Chevre,

Determination of Austenitic Transforma-

tion During Fast Heat Treatment, 17th ASM

Heat Treating Society Conference Pro-

ceedings, Including the First International

Induction Heat Treating Symposium, D.L.

Milan, D.A. Poteet, G.D. Pfaffmann, V.

Rudnev, A. Muehlbauer, and W.A. Albert,

Ed., ASM International, 1998, p 671–676

30. K.Z. Shepelyakovski and F. Bezmenov,

Advanced Steels for Modern Induction

Through and Surface Hardening, 17th ASM

Heat Treating Society Conference Pro-

ceedings, Including the First International

Induction Heat Treating Symposium, D.L.

Milan, D.A. Poteet, G.D. Pfaffmann, V.

Rudnev, A. Muehlbauer, and W.A. Albert,

Ed., ASM International, 1998, p 651–654

31. E. Ho

hne, Inductionsha

rten, Werk-

stattbu

cher fu

r Betriebsfachlente, Kon-

strukteure und Studierende, Heft 116,

Herausgeber, H. Haake, Ed., Springer-

Verlag, Berlin, Go

ttingen, Heidelberg, 1955

(in German)

32. G. Benkowsky, Induktionserwa

rmung,

VEB Verlag Technik, Berlin, 1985 (in

German)

33. G.D. Pfaffmann, Selective Surface Treat-

ment of Gears by Induction Proﬁle Hard-

ening, Mater. Sci. Forum, Vol 102–104,

1992, p 345–364

34. B. Crique, F. Ru

kcksu

hl, O. Longeot,

C. Delalean, S. Plano, B. Ottosson, G.

Bonzano, and C. Pichard, Multi Frequency

Induction Hardening of Gears to Replace

Carburizing—Methodology of Develop-

ment, 17th ASM Heat Treating Society

Conference Proceedings, Including the

First International Induction Heat Treating

Symposium, D.L. Milan, D.A. Poteet, G.D.

Pfaffmann, V. Rudnev, A. Muehlbauer, and

W.A. Albert, Ed., ASM International, 1998,

p 887–894

35. AGME Standard, American National Stan-

dard, Gear Materials and Heat Treatment

Manual, American Gear Manufacturers

Association, Virginia, 1989

36. G. Parrish and D.W. Ingham, The Sub-

merged Induction Hardening of Gears, Heat

Treat. Met., (No. 2), 1998, p 43–50

498 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_417-501.pdf/Chap_13/ 18/8/2008 4:04PM Plate # 0 pg 498

37. H. Sandinger, The Submerged Induction

Hardening of Gears, Heat Treat. Met.,

Vol 21 (No. 2), 1996, p 111–123

38. W. Amende, Transformation Hardening of

Steel and Cast Iron with High-Power

Lasers, Chapter 3, Industrial Applications

of Lasers, H. Koebner, Ed., John Wiley &

Sons, Chicheste, 1984, p 79–99

39. J. Meijer, R.B. Kuilboer, P.K. Kirner, and

M. Rund, Laser Beam Hardening: Trans-

ferability of Machining Parameters, Pro-

ceedings of the 26th International CIRP

Seminar on Manufacturing Systems—

LANE’94, Laser Assisted Net Shape Engi-

neering, M. Geiger and F. Vollertsen, Ed.,

Erlangen Meisenbach-Verlag, Bamberg,

p 243–252

40. J.F. Libsch, W.P. Chuang, and W.I.

Murphy, The Effect of Alloying Elements

on the Transformation Characteristics of

Induction-Heated Steels, Trans. ASM, Vol

42, 1950, p 121

41. C.R. Brooks, The Metallurgy of Induction

Surface Hardening, Adv. Mater. Process.,

2000, p H19–H23

42. V. Rudnev and R. Cook, Magnetic Flux

Concentrators: Myths, Realities, and Pro-

ﬁts, Met. Heat Treat., March/April 1995

43. W. Brunst, Die Inductive Wa

rmebe-

handlung, Springer-Verlag, Berlin, 1957

44. J. Grum and D. Ferlan, Residual Internal

Stresses after Induction Hardening and

Grinding, 17th ASM Heat Treating Society

Conference Proceedings, Including the

First International Induction Heat Treating

Symposium, D.L. Milan, D.A. Poteet, G.D.

Pfaffmann, V. Rudnev, A. Muehlbauer, and

W.A. Albert, Ed., ASM International, 1998,

p 629–639

45. S.L. Semiatin and D.E. Stutz, Induction

Heat Treatment of Steel, 2nd printing,

American Society for Metals, 1987

46. N. Stevens, Induction Hardening and Tem-

pering, Heat Treating, Vol 4, Metals

Handbook, 9th ed., American Society for

Metals, 1981, p 451–483

47. S. Denis, M. Zandona, A. Mey, and S.A.

Boufoussi, Calculation of Internal Stresses

during Surface Heat Treatment of Steels,

Proceedings of European Conference on

Residual Stresses, V. Hauk, H.P. Hougardy,

E. Macherauch, and H.D. Tietz, Ed.,

(Frankfurt, Germany), 1992, DEM Infor-

mationsgesselschaft mbH, Oberursel, 1993,

p 1011–1020

48. U. Bru

ckner, W. Schuler, and H. Walter,

Untersuchungen zum Eigenspannungszu-

beim inductiven Randschnitha

rten, Eigen-

spannungen: Entstehung-MessungBewer-

tung, Band 1, E. Macherauch and V. Hank,

Ed., Deutsche Gesselschaft fu

r Metallkunde

E.V., Oberursel, 1983, p 293–308 (in

German)

49. M. Melander, Computer Calculations of Re-

sidual Stresses due to Induction Hardening,

Eigenspannungen Entstehung-Messung-

Bewertung, Band 1, E. Macherauch and V.

Hank Ed., Deutsche Gesselschaft fu

r Metal-

lkunde E.V., Oberursel, 1983, p 309–328

50. A.J. Fletcher, Thermal Stress and Strain

Generation in Heat Treatment, Chapter 9.2,

Induction Hardening, Elsevier Applied

Science, London and New York, 1989,

p 182–187

51. G. Seulen and H. Voss, Oberﬂa

chenha

rtung

mil Induktionserhitzung bei mittleren Fre-

quenzen, Stahl Eisen, Vol 63 (No. 51),

1943, p 929–935 (in German)

52. H. Fujio, T. Aida, Y. Masumoto, and T.

Tsuruki, Paper 169-14, Distortion and Re-

sidual Stresses of Gears Caused by Hard-

ening, Third Report, Induction Hardening of

Gears Made of 535



C Carbon Steel, Bull.

JSME, Vol 22 (No. 169), 1979, p 1001–1008

53. G.H. Ledl, Why Quenching Is Important

in Induction Heating, Met. Prog., 1967,

p 200–204

54. J. Grum, How to Select Induction Surface

Hardening and Finished Grinding Condi-

tions in Order to Ensure High Compressive

Residual Stresses on Machine Parts Surface,

Mater. Sci. Forum, Vol 426–432, 2003,

p 2599–2604

55. G. Parrish, D.W. Ingham, and M. Chaney,

The Submerged Induction Hardening of

Gears, Part 1: Processing and Application

Aspects, Heat Treat. Met., Vol 25 (No. 1),

1998, p 1–8

56. G. Parrish, D.W. Ingham, and M. Chaney,

The Submerged Induction Hardening of

Gears, Part 2: Properties Imparted and

Problems Solved, Heat Treat. Met., Vol 25

(No. 2), 1998, p 43–50

57. T. Inoue, H. Inoue, F. Ikuda, and T. Horino,

Simulation of Dual Frequency Induction

Hardening Process of a Gear Wheel, Proc.

of Third International Conf. on Quenching

and Control of Distortion, G.E. Totten,

B. Lisicic, and H.M. Tensi, Ed., ASM

International, 1999, p 243–250

Induction Hardening / 499

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_417-501.pdf/Chap_13/ 18/8/2008 4:04PM Plate # 0 pg 499

58. M. Field, J.F. Kahles, and J.T. Cammet,

Review of Measuring Method for Surface

Integrity, Ann. CIRP, Vol 21 (No. 2), 1971,

p 219–237

59. E. Brinksmeier, A Model for the Develop-

ment of Residual Stresses in Grinding,

Advanced In-Surface Treatments, Tech-

nology—Applications—Effects, Vol 5, A.

Niku-Lari, Ed., Published in Cooperation

with the Institute for Industrial Technology

Transfer International, Pergamon Press,

Oxford, 1987, p 173–189

60. R. Snoeys, M. Maris, and J. Peters, Ther-

mally Induced Damage in Grinding, Key

Note Papers, Ann. CIRP, Vol 27 (No. 2),

1978, p 571–581

61. P.G. Althaus, Residual Stresses in Internal

Grinding, Ind. Diamond Rev., (No. 3), 1985,

p 124–127

62. M. Moris and R. Snoeys, Heat Affected

Zone in Grinding Operations, Proceed-

ings of the 14th International Machine

Tool Design and Research Conference,

F. Koenigsberger and S.A. Tobias, Ed.,

Manchester, 1973, p 569–669

63. J. Triemel, Grinding with Cubic Boron

Nitride, Proceedings of the 14th Inter-

national Machine Tool Design and

Research Conference, F. Koenigsberger

and S.A. Tobias, Ed., Manchester, 1973,

p 671–676

64. J. Grum, A Review of the Inﬂuence of

Grinding Conditions on Resulting Residual

Stresses after Induction Surface Hardening

and Grinding, J. Mater. Process. Technol.,

Vol 114 (No. 3), 2001, p 212–226

65. J. Grum, How to Select Induction Surface

Hardening and Finished Grinding Condi-

tions in Order to Ensure High Compressive

Residual Stresses on Machine Parts Surface,

Mater. Sci. Forum, 2002, p 404, 407, 623,

628

66. J. Grum, Measuring and Analysis of Resi-

dual Stresses after Induction Hardening and

Grinding, Mater. Sci. Forum, 2002, p 347,

349, 453–458

67. F. Kosel and L. Kosec, Internal Stresses in a

Surface Hardened and Ground Steel, Mech.

Eng. J., Vol 31 (No. 9–10), 1985, p 225–230

68. H. Staudinger, Induktive Ober-

ﬂa

chenha

rung und Eigenspannungen,

Untersuchungen u

ber Stahlauswahl und

geeignete Wa

rmevorbehandlung, Ha

rt.-

Tech. Mitt., Vol 2 (No. 21), 1966, p 111–123

(in German)

69. V. Nemkov, Role of Computer Simulation

in Induction Heating Technique, Proc. of

the Int. Induction Heating Seminar (Padua),

1998, p 301–308

70. P.K. Braisch, The Inﬂuence of Tempering

and Surface Conditions on the Fatigue

Behaviour of Surface Induction Hardened

Parts, Mater. Sci. Forum, Vol 102–104,

1992, p 319–334

71. H.C. Child, Residual Stress in Heat-Treated

Components, Heat Treat. Met., Vol 4, 1981,

p 89–94

72. H.C. Child, The Heat Treatment of Tools

and Dies—A Review of Present Status and

Future Trends, Tools and Dies for Industry

Book, The Metals Society, p 185, 291–321

73. Electromagnetic Induction and Electric

Conduction in Industry, Chapter 9, Heat

Treatments by Induction, Centre Francais de

I’Electricite, 1997

74. R. Jo

nsson, Toleranzen bei der Wa

rmebe-

handlung von Sta

hlen in Ha

rterein, TZ Met.

bearb., Vol 80 (No. 5), 1986 (in German)

75. K.E. Thelning, Chapter 4: Hardenability,

Steel and Its Heat Treatment, Bofors

Handbook, Butterworths, London and

Boston, 1975, p 127–181

76. H. Arend and W. Neuhaus, Die Ha

rtbarkeit

des Stahles, Verlag W. Girardet, Essen,

1955 (in German)

77. D.H. Breen and G.H. Walter, Computer-

Based System Selects Optimum Cost

Steel—I, Source Book on Materials Selec-

tion, Vol I, American Society for Metals,

1977, p 182–185

78. D.H. Breen, G.H. Walter, and C.J. Keith,

Computer-Based System Selected Opti-

mum Cost Steels—II, Source Book on

Materials Selection, Vol II, American

Society for Metals, 1977, p 186–192

79. D.H. Breen, G.H. Walter, and C.J. Keith,

Computer-Based System Selected Opti-

mum Cost Steels—III, Source Book on

Materials Selection, Vol III, American

Society for Metals, 1977, p 193–197

80. D.H. Breen, G.H. Walter, and J.T. Sponzilli,

Computer-Based System Selected Opti-

mum Cost Steels—IV, Source Book on

Materials Selection, Vol IV, American

Society for Metals, 1977, p 198–201

81. D.H. Breen, G.H. Walter, and J.T. Sponzilli,

Computer-Based System Selected Opti-

mum Cost Steels—V, Source Book on

Materials Selection, Vol V, American

Society for Metals, 1977, p 202–206

500 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_417-501.pdf/Chap_13/ 18/8/2008 4:04PM Plate # 0 pg 500

82. H. Matthes, Novel Process of Quality Con-

trol during Inductive Hardening Process,

17th ASM Heat Treating Society Con-

ference Proceedings, Including the First

International Induction Heat Treating

Symposium, D.L. Milan, D.A. Poteet, G.D.

Pfaffmann, V. Rudnev, A. Muehlbauer, and

W.A. Albert, Ed., ASM International, 1989,

p 727–733

83. J. Grum, Input and Output Control of Steel

Intended for Surface Hardening, 17th ASM

Heat Treating Society Conference Pro-

ceedings, Including the First International

Induction Heat Treating Symposium,

D.L. Milan, D.A. Poteet, G.D. Pfaffmann,

V. Rudnev, A. Muehlbauer, and W.A.

Albert, Ed., ASM International, 1989,

p 763–768

84. P. Birk, Ha

rtenscnutzmassen bei der Ein-

stazha

rtung, Ha

rt.-Tech. Mitt., Herausgeg-

eben von Reibensahm, Vol 10 (No. 3),

1957, p 9–19 (in German)

Induction Hardening / 501

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_417-501.pdf/Chap_13/ 18/8/2008 4:04PM Plate # 0 pg 501

Failure Analysis of Steel Welds

J.H. Devletian, Portland State University

D. Van Dyke, MEI-Charlton, Inc.

FAILURE ANALYSIS OF STEEL WELDS

may be divided into three categories. They in-

clude failures due to: design deﬁciencies, weld-

related defects usually found during inspection,

and failures in ﬁeld service. Failures due to

design deﬁciencies usually result in a ductile

overload due to overstressing the component or

exposing the component to service conditions

well beyond expected performance require-

ments. Weld-related defects, which fall outside

the quality-acceptance criteria speciﬁed by the

applicable welding code, contribute to pre-

mature failures upon testing as well as in service.

Failures in the ﬁeld may be insidious, because an

unanticipated fracture mechanism has been

causing problems without the designer’s knowl-

edge. In this chapter, the failures due to various

discontinuities in the steel weldment are em-

phasized. These include poor workmanship, a

variety of hydrogen-assisted cracking (HAC)

failures, stress-corrosion cracking, fatigue, and

solidiﬁcation cracking in steel welds.

Discontinuities in Steel Welds

By far, the most common steel welding prob-

lems are associated with poor workmanship,

such as lack of fusion, lack of penetration, por-

osity, undercut, arc strikes, and others. The num-

ber and size of such discontinuities may be cause

for rejection by most codes and can act as stress-

concentration sites to reduce weld joint strength

and promote cracking, such as fatigue failures. A

description of the wide variety of discontinuities

that can occur in steel weld metal and base metal

is provided by many welding codes, one of

which is the American Association of State

Highway Transportation Ofﬁcials/American

Welding Society (AASHTO/AWS) D1.5 Bridge

Welding Code (Ref 1). From this code, reject-

able discontinuities that can occur in common

butt welds are listed in Table 1 and illustrated in

Fig. 1. In this ﬁgure, discontinuities 1 through 7

are caused by poor workmanship. These include

(1) porosity, (2) inclusions, (3) lack of fusion,

(4) lack of penetration, (5) undercut, (6) under-

ﬁll, and (7) overlap. Discontinuities 8, 9, and 10

are associated with the rolling operation in the

steel mill. These include (8) laminations, (9)

delaminations, and (10) seams and laps in the

steel plate.

The most insidious discontinuities are those

that cause brittle cracking due to metallurgical

origins, such as HAC, solidiﬁcation cracking,

or stress-corrosion cracking. Such brittle fail-

ures are characterized by fracture surfaces

exhibiting intergranular, interdendritic, or cleav-

age modes of fracture, which occur at stress

levels well below the yield stress. Addition-

ally, even failures by quasi-cleavage and/or

Table 1 Common types of weld discontinuities

illustrated in Fig. 1

No. Discontinuity Location(a)

1 Porosity(b) W

Cluster(1b in Fig. 1)

Piping(1d in Fig. 1)

2 Inclusions W

Non-metallic slag(2b)

3 Lack of fusion(b) W

4 Lack of joint penetration(b) W

5 Undercut(b) BM

6 Underﬁll(b) W

7 Overlap(b) W

8 Laminations(b) BM

9 Delaminations(b) BM

10 Seams and laps(b) BM

11 Lamellar tears BM

12 Cracks(b)

a. Longitudinal(12a) W, HAZ, BM

b. Transverse(12b) W, HAZ, BM

c. Crater(12c) W

d. Throat(12d) W

e. Toe(12e) HAZ

f. Root(12f)

g. Underbead and HAZ(12g)

(a) W, weld; HAZ, heat-affected zone; BM, base metal. (b) See Fig. 1. Source:

Ref 1

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_503-519.pdf/Chap_14/ 18/8/2008 4:05PM Plate # 0 pg 503

Failure Analysis of Heat Treated Steel Components

L.C.F. Canale, R.A. Mesquita, and G.E. Totten, editors, p 503-519

DOI: 10.1361/faht2008p503

www.asminternational.org

microvoid coalescence can occur at reduced

levels of ductility. Such cracking is illustrated in

discontinuities 11 and 12a through 12g in Fig. 1.

These types of brittle failures are especially

insidious, because the fracture occurs at stress

levels and ductility levels well below the design

parameters.

Two examples of brittle failure in steel welds

are hydrogen-assisted cracking and solidiﬁ-

cation cracking. Among the discontinuities illus-

trated in Fig. 1, HAC can commonly take the

form of lamellar tears, longitudinal cracks in the

heat-affected zone (HAZ) (12a), transverse

cracks (12b), toe cracks in the HAZ (12e), and

underbead cracks in the HAZ (12g). Solidiﬁ-

cation cracking can commonly take the form

of longitudinal cracks along the weld centerline

(12a), crater cracks (12c), and ﬁssures and

microﬁssures in the weld metal.

Example 1: Porosity in Weld Metal. Two-

pass submerged arc welding was used to butt

weld two 19 mm (

in.) thick A709-grade 250

plates. The joint design was a simple square

groove butt joint with 1.6 mm (

in.) root

opening. It was designed to be welded in two

passes, with a single pass on each side of the

plate. After the ﬁrst side was welded, the weld

bead exhibited an unusually large amount of

reinforcement. Subsequent magnetic particle

inspection indicated extensive discontinuities in

the weld bead. The ﬁrst pass of the weld was

sectioned, polished, and etched in 5% nital.

The macrograph of the welded plate is shown

in Fig. 2. Extensive rejectable porosity (in ac-

cordance with AWS D1.1 Structural Welding

Code) developed within the weld metal but was

not readily visible on the surface.

The welding system was inspected for possi-

ble sources of porosity. Inspection revealed that

the mating ﬂanges were covered with thick mill

scale and were neither ground clean nor pre-

heated prior to welding. The welding electrode

and ﬂux combination was F7A2-EM12K. Nor-

mally, this ﬂux/wire combination is designed to

be tolerant of light rust. However, in this case,

the porous mill scale was so thick that it acted as

a sink for moisture, oil, and other volatile con-

taminants.

Excessive porosity was caused by welding on

heavily rusted steel. Simply rough grinding the

faying surfaces to bare metal prior to welding

completely eliminated the porosity problem.

If rough grinding is not possible, the welding

engineer can choose a more active ﬂux that is

designed for welding rusted plate. Active ﬂuxes

contain strong deoxidizers, such as titanium

and aluminum, but should only be used in a

single pass.

Fig. 1

Schematic illustration of common types of discontinuities in welds. Reproduced with permission of the American Welding

Society. See also Table 1. Source: Ref 1

504 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_503-519.pdf/Chap_14/ 18/8/2008 4:06PM Plate # 0 pg 504

Example 2: Lack of Penetration. A36

structural steel was butt welded by automatic

ﬂux-cored arc welding. A double V-groove joint

preparation was used for a two-pass welding

operation on 6.4 mm (

in.) thick plate. By

visual examination, the weld appeared satis-

factory. Full penetration of the butt joint was

required. Transverse-to-weld tensile testing

revealed inadequate yield strength. The tensile

test failure occurred in the weld metal. Normally,

previous tensile tests always failed in the unaf-

fected base metal.

Radiographic examination of the weld joint

and the mandatory transverse-to-weld tensile

test showed the weld to have inadequate pene-

tration. Inspection of the fracture surface of the

tensile specimen revealed unfused metal at the

plate midthickness. The fractured steel exhibited

a ductile slant fracture. By measuring the ap-

proximate area of fused cross section and area of

unfused cross section, it was determined that

approximately 80% of the joint thickness carried

the load.

Even though the weld fracture was ductile, the

reduced cross-sectional area of fused metal (due

to lack of penetration) did not provide enough

sound metal to meet the required joint strength.

Welding parameters were changed to ensure full

penetration.

Example 3: Lack of Side-Wall Fusion. DH-

36 structural steel for shipbuilding was butt

welded by robotic gas metal arc welding. A

double V-groove joint preparation was used for a

two-pass welding operation. By visual exam-

ination, the weld appeared satisfactory. Dye-

penetrant testing revealed a discontinuity along

one edge of the weld. Face bend testing revealed

a rejectable discontinuity along the same edge of

the weld.

Visual inspection of the discontinuity in the

bend test showed that the base metal and weld

metal were not fused at the toe. Subsequent

metallographic examination showed that the

lack of side-wall fusion existed throughout the

length of the weld.

Clearly, the robot and ﬁxturing were slightly

misaligned. The robot welder directed the

welding heat too far on one side of the joint and

not enough on the other side. The robot was re-

programmed to provide adequate weave to melt

both faying surfaces sufﬁciently to prevent a

recurrence of the problem.

Fatigue of Welded Joints

Fatigue cracking is a result of repetitive

ﬂuctuating stress causing fracture well below the

yield strength of the steel. Factors required for

fatigue fracture include a sufﬁciently high ten-

sile stress, a sufﬁciently large variation in ap-

plied stress, and a sufﬁcient number of cycles of

applied stress. Cracks resulting from fatigue

account for more than half of all failures. Fatigue

fractures are insidious because there is no visible

plastic deformation preceding fracture. Fatigue

fractures are promoted by stress concentrators

such as notches, sharp ﬁllets, corners, holes,

threads, splines, keyways, dents, gouges, laps,

folds, ﬂakes, and delaminations in plates, sheets,

and forgings. Tensile residual stresses arise

from punched holes, heat treatment, welding,

and so on.

Visible beach marks indicate the location of

the fracture initiation site. Beach marks are

produced when oxidation at the crack tip

propagating under an alternating stress is inter-

rupted from time to time, for example, constant-

speed machinery that is turned on and off.

Also, visible beach marks are generated when

non-steady-state loading occurs or with rapid

changes in loading, for example, auto/truck

components and road signs.

Usually, fatigue cracks have multiple initi-

ation sites, particularly in rotating shafts subject

to bending loads with stress concentrations.

Multiple initiation sites are easily recognized be-

cause of the presence of ratchet marks. A ratchet

mark ridge or ligament occurs at the merger of

two adjacent fatigue cracks that propagate on

different planes. When these two fatigue cracks

meet, they are connected by a ridge of deformed

Fig. 2

Wormhole or piping porosity in weld metal deposited

by submerged arc welding. Plate is 19 mm thick.

Failure Analysis of Steel Welds / 505

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_503-519.pdf/Chap_14/ 18/8/2008 4:06PM Plate # 0 pg 505

metal called a ratchet mark. For example, three

adjacent propagating fatigue cracks will have

two ratchet marks connecting them.

When viewing a fatigue fracture surface at

high magniﬁcation of approximately 500 to

5000 · under the scanning electron microscope

(SEM), striations are usually visible. In ductile

structural materials, crack propagation is by

microvoid coalescence, and the crack path be-

comes perpendicular to the tensile axis. The

correlation between striation spacing and tensile

load has been well established, where one stria-

tion represents one fatigue load cycle. Striation

spacing increases with increasing maximum

tensile loading. Secondary cracks sometimes

occur alongside striations, which better resolves

striation spacing.

For brittle materials, crack propagation occurs

in the cleavage or intergranular mode. Since

brittle cracks can grow unstably, the crack path

provides less clear correlation between striation

spacing and loading. Although striations are less

well deﬁned, secondary cracks are helpful in

resolving some striations. An absence of observ-

able striations is possible when metals with

complex metallurgical microstructures obscure

striations, or when the fatigue crack closes and

plastically deforms the striations. The latter case

usually produces a very shiny fracture surface.

Example 4: Fatigue Cracking of Welded

Pipe Flange. A short drain pipe from a heat-

recovery steam generator failed in service by

developing leaks around a circumferential weld

near a ﬂange. The pipe was a 25 mm (1 in.)

schedule 80 pipe (25 mm, or 1 in., internal dia-

meter by 4.8 mm, or

in., wall thickness)

section 127 mm (5 in.) long with ﬂanges welded

onto both ends. The pipe had been in service for

5 years with an operating pressure of 8.3 MPa

(1200 psi). Upon failure, the crack completely

separated the ﬂange from the pipe.

The 127 mm (5 in.) section of pipe with

ﬂanges was cut from the boiler and submitted for

examination, as shown in Fig. 3. The fracture

surface of the pipe is shown in Fig. 4. There are

clearly deﬁned ratchet marks in both the top and

bottom edges of the fracture surface. Beach

marks are present on both the top and bottom

of the fracture surface that run toward the cen-

ter, where there is a brightly colored band.

Beach marks are not present in the brightly

colored band.

Metallographic sectioning of the weld proﬁle

showed that the crack initiated at the toe of the

weld and proceeded in between the weld ﬁller

metal and HAZ. There was no evidence of weld

defects such as undercut or porosity present in

the section.

Scanning electron microscopy showed ﬁne

striations within the beach mark areas of the

pipe. The SEM showed a ductile dimpling struc-

ture within the brightly colored band of the frac-

ture surface. Energy-dispersive spectroscopy of

several areas of the fracture surface revealed

only iron, manganese, and oxygen.

The pipe fractured due to fatigue cracks that

initiated at several locations at the toe of the

weld. The fatigue cracks propagated until there

was insufﬁcient cross section to support the

service loads. The morphology of the fracture

suggests the cracking propagated under reversed

bending stresses.

Hydrogen-Assisted Cracking Theory

In addition to discontinuities resulting from

poor workmanship, a very common cause of

Fig. 3

Overview of pipe section. Cracking is visible on right

end of the pipe at the toe of the weld. Courtesy of MEI-

Charlton, Inc.

Fig. 4

Overview of pipe end. Ratchet marks and beach marks

are clearly visible. Courtesy of MEI-Charlton, Inc.

506 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_503-519.pdf/Chap_14/ 18/8/2008 4:06PM Plate # 0 pg 506

failure in welded steel plates, castings, and

structural components is HAC. Although HAC

of steel weldments has been studied for over

a half-century, the mechanism of cracking is

still uncertain. Fortunately, judicious “hydrogen

management” has permitted the crack-free

welding of high-strength steels primarily by

preheating the weld to a temperature that must

be increased with increasing:

 Diffusible hydrogen content

 Yield strength

 Hardness

 Carbon equivalent

 Thickness

 Restraint

 Susceptible microstructure

Cracking can take place in either the weld

metal or the HAZ, depending on which is more

susceptible. Generally, the peak hardness,

strength, and carbon equivalent values of the

HAZ are greater than those of the weld metal.

So, cracking usually takes place in the HAZ.

This is the most common form of HAC and is

called underbead cracking. Because HAC is

dependent on time for hydrogen atoms to diffuse

at room temperature to crack nucleation sites, it

is also called delayed cracking. Such cracking

can take place anytime, from minutes to several

days after the weld has cooled. Several other

forms of hydrogen cracking are also possible

and are discussed.

After over a half-century of research on the

deleterious effects of hydrogen in steel, the

fundamental mechanism of HAC is still uncer-

tain. Preheating, however, is the most reliable

method to prevent weld cracking. Unfortu-

nately, preheating is labor-intensive and costly.

More recent studies have shown that preheating

temperatures can be either reduced or eliminated

if the composition of the steel base metal and

ﬁller metal is low in crack-promoting elements.

(Ref 2–19). For example, an empirical carbon

equivalent formula (CEN), developed by Yur-

ioka et al. (Ref 2), clearly reﬂects the fact that

alloying elements such as carbon and boron

produce high CEN values and have a strong

detrimental effect on hydrogen crack resistance:

CEN= C+ A( C) f Si=24+ Mn=6+ Cu=15

+ Ni=20+( Cr+ Mo+ Nb+ V) =5+ 5Bg

where A(C) = 0.75+0.25 tanh{20(C0.12)},

and all elements are in weight percent.

Thus, modern structural steel plate and ﬁller

metals contain reduced carbon for increased

resistance to hydrogen cracking and, coinci-

dently, for increased fracture toughness. Many

low-carbon steels, such as the U.S. Navy’s

HSLA-65, HSLA-80, and HSLA-100, are

designed to be welded without preheating (Ref

20–22).

The presence of hydrogen in welds can cause

failure at levels of strength and plastic strain well

below design yield strength and ductility. The

current models of HAC involve pre-existing

defect sites in the weld metal. These initiation

sites include inclusions, minor phase particles,

microscopic cracks, and other discontinuities.

Many theories attempt to explain the brittleness

caused by the presence of as little as a few parts

per million (ppm) of hydrogen in steel. The

decohesion theory of Oriani (Ref 23–29) states

that dissolved hydrogen migrates to regions of

triaxial tensile stress, and that the cohesive for-

ces between atoms in the iron lattice are reduced

in proportion to the interstitial hydrogen con-

centration. This theory provides for the observed

increase in hydrogen solubility at the tip of a

crack (Ref 30). Diffusible hydrogen has the

additional effect of supplying internal mechan-

ical stresses that produce an apparent softening

of the steel. Hydrogen also causes growth of

microvoids in undeformed specimens and the

initiation of microcracks in specimens deformed

to the critical strain (Ref 29). During cooling and

shrinkage of a weld, these defect sites become

localized regions of triaxial tensile stress con-

centration. Atomic hydrogen diffuses inter-

stitially to these regions of the expanded lattice.

As the concentration of hydrogen increases, the

cohesive strength between iron atoms decreases

below the local intensiﬁed stress level, at which

point cracking occurs. A deﬁnitive reason why

the cohesive strength between iron atoms

decreases in the presence of hydrogen atoms

has not been fully explained. In 1960, Troiano

(Ref 31) suggested that certain sites act to con-

centrate stress to promote cracking, and that

hydrogen reduces the cohesive strength between

iron atoms.

In a classic 1972 paper (Ref 32), C.D. Bea-

chem presented a model suggesting that the

presence of sufﬁciently concentrated hydrogen

dissolved in the lattice just ahead of the crack

tip aids whatever deformation processes the

microstructures will allow. Intergranular, quasi-

cleavage, or microvoid coalescence fracture

modes are dependent on the microstructure, the

Failure Analysis of Steel Welds / 507

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_503-519.pdf/Chap_14/ 18/8/2008 4:06PM Plate # 0 pg 507

crack-tip stress intensity, and the concentration

of hydrogen. That is, hydrogen reduces the stress

intensity needed for cracking regardless of the

mode of fracture, as shown in Fig. 5. Even the

stress intensity needed for HAC by microvoid

coalescence is reduced in the presence of hydro-

gen. Beachem’s model uniﬁes several theories

but shows how the stress-sorption and lattice em-

brittlement models are unnecessary. The model

shows that the planar pressure effects are neces-

sary at low stress intensities and are necessary

only to augment the driving force from the ap-

plied loads. The basic hydrogen-steel interaction

appears to be an easing of dislocation motion or

dislocation generation, or both.

In 2003, Albert et al. (Ref 33) found that

hydrogen traps increased with alloying ele-

ments. Measurements of diffusible hydrogen in

weld metal decreased in steel welds containing

increasing amounts of chromium and molyb-

denum alloying elements. The HAC in steel

is dependent on hydrogen trapping, which is

the interaction between diffusible hydrogen

and lattice defects. Several investigators have

examined hydrogen trapping in steel (Ref 34–

40). These traps include point defects (vacan-

cies, solute atoms), dislocations, interfaces and

surfaces (grain boundaries, particle-matrix inter-

faces, cracks, external surfaces), and volume

defects (voids, second-phase particles).

Gibala and DeMiglio (Ref 34) experimentally

measured binding energies (Table 2) and satur-

abilities of hydrogen traps in AISI 4340 steel,

which led them to predict probable suscept-

ibilities to HAC. Both primary and secondary

crack paths were explained in terms of the types

of microstructural traps that predominate in a

given material. Management of hydrogen traps

can have a profound effect on the susceptibility

to HAC. In theory, decreasing the amount of

diffusible hydrogen by introducing permanent

or irreversible traps has the beneﬁcial possibility

of reducing the susceptibility to HAC.

Olson et al. (Ref 37–40) have shown that

effective hydrogen traps can include micro-

scopic particles of Ce

, TiC, Y

VC, and

NbC. Although these traps would decrease

the amount of diffusible hydrogen in the weld,

large quantities of such particles would have a

deleterious effect on ductility and fracture tough-

ness. Steel can be embrittled by hydrogen if it is

stressed at temperatures high enough to allow

hydrogen to diffuse to potential embrittlement

sites, but also low enough that the hydrogen

is not depleted from embrittlement-producing

traps. With increasing temperature, steel is

expected to exhibit mechanical behavior de-

termined largely by the nature of hydrogen

transport among the populations of various

hydrogen traps (Ref 34).

Increasing the tensile strain or residual tensile

stress also increases the ability of the steel to

retain hydrogen in body-centered cubic (bcc)

iron. Normally, the solubility (s) of hydrogen in

bcc iron is given by:

s=47:66 p

1=2

exp (72:72· 10

=RT)

where s is in ppm by mass, R is the gas constant,

T is the absolute temperature, and p is the pres-

sure in atmospheres. For example, at room

temperature and pressure, the solid solubility of

hydrogen in bcc iron is only 6.7 · 10

4

ppm.

Solubility is increased by the presence of traps.

For example, by increasing tensile strain, the

solid solubility of iron (Ref 30) increased by a

Table 2 Hydrogen trap interactions in steel

Hydrogen trap type

Binding energy

), kJ/mol

Interstitial solutes 3–15

Ti atom 26

Vacancy 46

Screw dislocation core 20–30

Mixed dislocation core 59

Hydrogen vapor/void 29

Grain boundary 59

Free surface 70–95

AlN interface 65

C interface 84

TiC interface 96

Source: Ref 34

Fig. 5

Effect of hydrogen content on hydrogen-a ssisted

cracking (HAC) for microvoid coalescence (MVC),

quasi-cleavage (QC), and intergranular (IG) fracture modes.

Adapted from Beachem. Source: Ref 32

508 / Failure Analysis of Heat Treated Steel Components

Name ///sr-nova/Dclabs_wip/Failure_Analysis/5113_503-519.pdf/Chap_14/ 18/8/2008 4:06PM Plate # 0 pg 508