How Yield Strength and Fracture Toughness Considerations Can Influence Fatigue Design Procedures for Structural Steels
The K,c /cfys ratio provides significant information relevant to fatigue design procedures, and use of NRL Ratio Analysis Diagrams permits determinations based on two relatively simple engineering tests
BY T. W. CROOKER AND E. A. LANGE
ABSTRACT. The failure of welded fatigue failures in welded structures probable mode of failure, the impor structures by fatigue can be viewed are primarily caused by crack propa tant factors affecting both fatigue and as a crack growth process operating gation. Numerous laboratory studies fracture, and fatigue design between fixed boundary conditions. The and service failure analyses have re procedures are discussed for each lower boundary condition assumes a peatedly shown that fatigue cracks group. rapidly initiated or pre-existing flaw, were initiated very early in the life of and the upper boundary condition the structure or propagated to failure is terminal failure. This paper outlines 1-6 Basic Aspects the modes of failure and methods of from pre-existing flaws. Recognition of this problem has The influence of yield strength and analysis for fatigue design for structural fracture toughness on the fatigue fail steels ranging in yield strength from 30 lead to the development of analytical to 300 ksi. It is shown that fatigue crack techniques for fatigue crack propaga ure process is largely a manifestation propagation characteristics of steels sel tion and the gathering of a sizeable of the role of plasticity in influencing dom vary significantly in response to amount of crack propagation data on the behavior of sharp cracks in struc broad changes in yield strength and frac a wide variety of structural materials. tural metals. The degree of plasticity ture toughness. However, strength and However, a degree of organization involved in a given situation influences toughness do vary widely among struc and perspective has been lacking in both the mechanical behavior of tural steels, and thereby exert profound much of this work. Preoccupation by cracks and the methods of analysis effects on the fatigue failure process. By which can be applied to the problem. applying the principles of the NRL Ratio researchers with crack propagation Analysis Diagram, an analysis is pre laws has obscured the fact that crack The degree of plasticity associated propagation is only a portion of the with a crack can be categorized into sented whereby the plane strain fracture 7 toughness to yield strength ratio (Ku/ fatigue failure process. Furthermore, one of three situations. First, there as.) can be employed to determine fa it is probably the portion of the fa can be gross ductile yielding where the tigue design procedures. The full spec tigue failure process least amenable to fully plastic state extends beyond the trum of structural steels is divided into a substantial improvement. vicinity of the crack and dominates three groups on the basis of fracture be The failure of welded structures by the behavior of the entire cracked havior: fatigue can be viewed as a crack body. Second, there can be large scale 1. Ductile plastic fracture mode (K,„/ growth process operating between localized yielding where the nominal a,. > 1.5). fixed boundary conditions—Fig. 1. deformations away from the vicinity 2. Elastic-plastic fracture mode (1.5 The lower boundary condition as of the crack are elastic. However, a > K,Joys > 0.5). sumes the existence of a rapidly initi plastic zone exists at the crack tip 3. Brittle elastic instability fracture ated or pre-existing flaw, and the up which, although contained in an elas mode C&0/
488-s I OCTOBER 1970 (K), the yield strength of the material increases and the crack tip plastic curve is limited by upper and lower (o-g,s), the stress state at the crack tip zone becomes large in relation to inflection points. The lower inflection (plane stress or plane strain), and the physical dimensions, successive cor point is indicative of nonpropagating mode of load application (monotonic rections must be applied and empiri fatigue cracks and occurs under ex or cyclic) :7 cism creeps into the fracture mechan ceedingly low stress intensities, where ics analysis. Nevertheless, the remain the crack growth rates are of the 2r = — (X/V )'--plane stress (1) der of this paper is presented using magnitude of the atomic spacing in y ys fracture mechanics terminology be the crystal lattice (~ 10~7 to 10-8 cause, despite analytical limitations, it in./cycle). The upper inflection point
2ry = — CKV/o-j^pIane strain (2) provides a common language basis. is caused by the onset of rapid un stable crack extension prior to termi Fatigue Crack Propagation nal fracture and places a critical limit 2ry = - (A^KJMatigue (3) on the fatigue resistance of structural HIT Crack Propagation Laws and Mechanisms metals. In actuality these equations merely With the recognition that crack The application of fracture provide rough estimates of the scale propagation is more critical than mechanics to fatigue set the trend for of yielding, mainly for the purposes of crack initiation for analyzing fatigue nearly all subsequent fatigue crack calculating an "effective" crack failure in many structures, numerous propagation studies. It is a successful length, based upon the actual crack efforts to develop engineering ap approach to crack propagation to the length plus a plastic zone size correc proaches to the problem were under extent that the stress-intensity factor taken. Much of this work was aimed tion (ry) for fracture mechanics describes the crack tip plasticity. Fa analyses.8 at revealing crack propagation laws tigue crack propagation occurs which would remain valid over a through cyclic plastic deformation oc Mechanically, large degrees of plas sufficiently broad range of experimen curring at the crack tip, regardless of ticity tend to inhibit fracture. High tal conditions. In their critical review how far plastic deformations exist be fracture toughness is associated with of crack propagation laws, Paris and yond the crack tip region.12'13 The the ability to develop large plastic 9 Erdogan showed that seemingly con relationship between gross cyclic plas zones or fully plastic yielding before tradictory results could be normalized tic strain and fatigue failure was rec fracture. However, fatigue crack prop to fit a single power-law curve by ognized a number of years ago and is agation properties do not seem to be relating the fatigue crack growth rate treated in detail by Manson.14 It was as sensitive to such distinctions, which (da/dN) to the stress-intensity factor used as the basis for a hypothesis for will be discussed in a later section. range (AK). The resulting equation fatigue crack propagation by Weiss.15 Analytically, the degree of accuracy was of the power-law form shown in More recently, Lehr and Liu16 have available for treating the mechanics of equation (4): shown that fatigue crack propagation sharp cracks in structural metals de m 7 8 da/dN = C(AK) (4) is caused by damage accumulation creases with increasing plasticity. ' due to strain cycling at the crack tip Fatigue and fracture under fully plas where C is a material constant and m and that the crack growth rate can be tic loading must be approached large is the slope of the power-law curve. 10 u calculated using strain cycling proper ly on an empirical basis. Linear-elastic More recent research ' has shown ties. fracture mechanics can be applied that the entire fatigue crack growth Consequently, both the proposed where the plastic zone resides within rate curve for most materials is actu mechanisms for fatigue crack propa an elastic stress field. However, rigor ally of sigmoidal form when plotted gation and the crack propagation laws ous analyses are limited to small scale on log-log coordinates, as shown in which seek to describe the phenom yielding. As the degree of yielding Fig. 3. The power-law portion of the enon in engineering terms are based upon cyclic crack tip plasticity. Spe cifically, fatigue crack propagation is primarily related to the range of strain CRITICAL FLAW SIZE FOR FAILURE cycling rather than the extent of the plastic zone. Crack growth rates have been shown to correlate with AK but are relatively insensitive to changes in stress state (plane strain vs plane stress) and are frequently unaffected by changes in yield strength.17 There fore, stress intensity provides a com mon basis for comparing fatigue crack propagation data from many sources. Crack Propagation Data A fairly extensive amount of fa tigue crack propagation data has been generated for steels.18"33 Much of this data is summarized in Figs. 4-6, which are log-log plots of da/dN vs. AK for zero-tension loading in ambi ent room air environments. These data have been separated into three plots on the basis of yield strength. 0 CYCLES OF REPEATED LOAD, N Nf The scatterband limits shown for each Fig. 1—Schematic illustration of the fatigue crack propagation failure plot attempt to encompass virtually process all of the data cited and are not
WELDING RESEARCH SUPPLEMENT | 489-s APPLIED
REGION |: REGION 3; NON-PROPAGAT RAPID ING FATIGUE UNSTABLE CRACKS CRACK GROWTH
Fig. 2—The fracture mechanics model for localized yielding STRESS-INTENSITY FACTOR RANGE, AK at the crack tip Fig. 3—The sigmoidal fatigue crack propagation curve intended to represent the curve for point appears more frequently in data ing in yield strength from 50 to 200 any individual steel. from progressively higher yield ksi, and for these materials, a gradual The striking feature of these three strength steels. The upper inflection increase in the crack growth rate ex plots is the fact that despite the broad point is associated with impending ponent (m) was observed with in differences included in the steels con fracture and occurs at progressively creasing yield strength.35 sidered, the scatterbands overlap one lower stress intensities with increasing Brothers and Yukawa report an another to a large degree. This fact is yield strength. opposite effect in low-alloy heat- even more striking if we examine a This feature of the fatigue curve treated steels over yield strength plot of the crack growth rate ex has been related to a critical percent ranges from 73 to 127 ksi.22 Miller ponent (m) vs. yield strength for age of the plane stress fracture studied several (mainly ultra high these steels—Fig. 7. It can be seen toughness (Kc),27 to observations to 23 strength) steels and concluded that m that, with few exceptions, m varies monotonic crack extension, and to is inversely related to Krc.2* Anctil from values slightly greater than 2 to 29 36 a critical value of crack opening dis and Kula and Throop and Miller values slightly greater than 4 in a 26 placement (S). In all probability conducted systematic investigations to rather consistent scatterband over the this phenomenon is environmentally determine the effect of tempering entire yield strength range from 24 to controlled in ultra high-strength temperature on fatigue crack propa 300 ksi. This does not imply that all steels.34 Nevertheless, it appears to be gation in AISI 4340 steel. Their re steels are essentially the same, since a the one feature of the sigmoidal fa sults, shown in Fig. 8, reveal that variation in the value of m from tigue crack propagation curve which proper tempering can produce an op approximately 2 to 4 is significant. is consistently influenced by yield timum fatigue crack propagation However, there does not appear to be strength and fracture toughness. a basic irresistible trend in fatigue resistance (minimum value of m) in several quenched-and-tempered steels. crack propagation power law curves Factors Which Affect Crack with yield strength such as that which Propagation Characteristics The authors also studied three high- strength steels at the 180-ksi yield exists with fracture toughness vs. yield Although fatigue crack propagation strength. strength level and observed that sig characteristics are not profoundly in nificant variation in the value of m The only consistent and progressive fluenced by strength and toughness, occurred with changes in fracture change with yield strength, that is except for the occurrence of upper toughness.26 So, it appears that, al apparent in the data cited in Figs. inflection points, it is of interest to though a broad trend in fatigue crack 4-6, is the occurrence and location of examine some of the causes for the propagation characteristics does not the upper inflection point. Few low- scatter observed in Figs. 4-7. In a necessarily occur with large changes strength steels exhibit an upper previous paper, the authors presented in yield strength, significant variations inflection point under elastic fatigue an analysis for the fatigue crack prop can occur as a result of heat treat- cycling. However, the upper inflection agation characteristics of steels rang
490-s I OCTOBER 1970 ment and among various alloy com conduct two relatively simple tests to the toughest steels where the KIC/o-ys positions. utilize the method. ratio exceeds 1.5. It is extremely diffi Apparent variations in fatigue A rational approach to the use of cult to obtain valid Kic measurements crack propagation characteristics of this information can be achieved by on steels for toughnesses in excess of ultra high-strength steels can be due to considering fracture performance in this ratio in accordance with ASTM 41 the test environment. One need only terms of the Klc/crys ratio. On this recommended practices. The thick note that the broadest scatterband basis the diagram has been divided ness requirement for plane strain re among Figs. 4-6 occurs for the ultra into three regions. Case 1 consists of straint to prevail at the crack tip is high-strength steels. Dahlberg has shown that fatigue crack propagation 1000 in a heat-treated 4340 steel is sensitive PLAIN CARBON AND to the water vapor content of the room air environment.37 Spitzig, Wei, LOW ALLOY STEELS et al., have also shown that fatigue YIELD Hit crack growth rates in ultrahigh- MATERIAL STRENGTH strength steels respond to water va 500 (KSI) por30' 31 and that growth rates meas MILD STEEL 34 ured in uncontrolled room air corre 304 STAINLESS 34 spond with test results obtained in a 2^- Cr-IMo 35 humid argon (100% relative humidi A2I2-B 45 ty) environment. However, since A302-B 53-61 welded structures normally operate in A508-2 66-69 ambient air (or worse) environments, Ni-Mo-V 82-85 and since virtually all fatigue data are Ni-Cr-Mo-V 73- 117 obtained in air, we shall consider such Cr-Mo-V 85-127 data to adequately represent baseline fatigue crack propagation characteris o tics for the purposes of this paper. >- 100 Fracture Toughness Influence of Yield Strength Fracture toughness is strongly influ gw&t enced by yield strength, as indicated schematically in Fig. 9. Although brit 50 tle steels can be found at any strength UI level, the maximum toughness that < can be obtained tends to remain about rr constant up to a yield strength level of 150 ksi, then falls off rapidly for 5 higher yield strengths. Generally, the o effect of heat treating a given alloy to rr obtain higher strength is to drastically *: reduce fracture toughness. Almost in < variably, decreased fracture toughness rr is the price that must be paid for o obtaining higher strength. Some ad to vanced technology steels can balance Z> 10 this trend by maintaining high tough ness up to about the 180 ksi yield mint strength level. However, such steels are rare and expensive, and their use is limited to critical applications. f: The Ratio Analysis Diagram The Ratio Analysis Diagram (RAD) provides a means of measur I ing fracture toughness using a simple, inexpensive test and quantitatively relating the results to structural per formance.38 Basically, the RAD is a cumulative plot of 1 in. Dynamic Tear (DT) test fracture energies vs yield strength for the full spectrum of struc tural steels39—Fig. 10. In addition, the DT energies are correlated with K,c which allows a quantitative inter J I I I I I pretation of the results.40 This meth 10 50 100 200 od possesses several advantages: it is STRESS-INTENSITY FACTOR RANGE, AK (KSlTm) applicable to the full spectrum of Fig. 4—Scatterband limits for fatigue crack propagation data in low- structural steels and one need only strength, plain carbon and low-alloy steels
WELDING RESEARCH S U P P L E M E N T I 491-s expressed by the relationship: fracture-safe use can be assured by elastic fracture mechanics can be ap utilizing the principles of the NRL plied with some degree of success.33 42 B > 2.5 (KIC/oysy (5) Fracture .Analysis Diagram. For steels in this category, thickness is Case 2 consists of an intermediate a very important factor in determin So, in essence, this region consists region where the KIe/o-ya ratio varies ing fracture behavior, since thickness of those steels whose fracture behav from 1.5 to 0.5. This region takes in a will determine whether there is ade ior cannot be analyzed by linear-elastic great many structural steels where quate restraint for plane strain condi fracture mechanics. It includes steels high toughness is either prohibitably tions to prevail. A considerable in up to a maximum yield strength of expensive from an economic stand crease in fracture toughness can be 190 ksi. Steels in the Case 1 category point or is difficult to obtain because assured by avoiding sections thick fracture in a ductile manner requiring of a desire to utilize higher yield enough to permit plane strain condi very large critical flaw sizes and large strengths. Typically, such steels frac tions. This, in turn, can have a very plastic strains. Many of these steels ture in what is termed an elastic- large effect on critical flaw sizes. undergo a fracture behavior transition plastic manner. Some measurable In addition, some distinctions from ductile to brittle with decreasing plasticity is associated with the frac should be made within this region. temperature. Fracture problems with ture process, but it tends to be local The subregion between ratio values of these steels are generally associated ized, i.e., large scale localized yield 0.5 and 1.0 is well defined. However, with low temperature service, and ing.7 Under these conditions, linear- above the ratio value of 1.0, a number
I uuu — / uw HIGH-STRENGTH STEELS '- £&•.•£•£ -ULTRAHIGH-STRENGTH 1 ~ STEELS — / 1
500 YIELD 500 MATERIAL YIELD 1 STRENGTH MATERIAL STRENGTH - (KSI) ~ (KSI) ik HY-80 80 I j H-ll 208-242 HY- 130 130 — D6AC 242 1 |j| 9Ni-4Co-0.20C 180 4340 220-292 W&B&i 9Ni-4Co-0.25C 180 IS Ni 245-308 0, 10 Ni 180-190 ft ' ' *M 12 Ni 180-190 @f llllf PH 13-8 Mo 180 I % jif ; 100 — / 100 — M JH? M # w $k ilp - i V m w 50 / / 50 us f / ' UJ - HI f / / 1 ~~ / / O / / E HI a 5i# M i £$!§t$f i i § Iff M i U If 3 10 O 10 — t- •;. .;.''; i ; ; f "-•' j iliiilf -;V:> / -..••''-;. / (-' / 7 5 mmtmw 5 — w I $ '•j u i / - Ill J
/:v-v:. / 1 ';;;0020£f 1 1 f 1 1 1 111 11 1 $!&$§tW i i i i i i i 1 /.. . 10li, 50 100 200 10 50 100 200 STRESS-INTENSITY FACTOR RANGE, AK(KSIVm) STRESS-INTENSITY FACTOR RANGE, AK(KSiyfn) Fig 5—Scatterband limits for fatigue crack propagation data Fig. 6—Scatterband limits for fatigue crack propagation data in high-strength steels in ultrahigh-strength steels
492-s I OCTOBER 1970 10 —[ 0 | | 1 1 STRUCTURAL STEELS 1 m LU QUENCHED - ANDJ-TEMNERED I I- — £-=C(AK) — -z. 2 8 dN o STRUCTURAL STEELS LU 0_ o X CL 0 ° o LU 6 — LU da/dN=C(AK) m < N CC G o •
X rr H G — OS© ^" T" " o G 0 G— 4- O G CC V Gg % 0 o o GG °°G G Q o O O 9 ° GG rr c^o-o o ce_ja Q___ _ _G QQ _ . _S o < — o LT. o o o < 0 _1 1 1 1 1 0 400 800 1200 0 50 100 150 200 250 300 TEMPERING TEMPERATURE (°F) YIELD STRENGTH (KSI) Fig. 8—Variation in the crack growth rate exponent value Fig. 7—Summary plot of crack growth rate exponent values with tempering temperature for AISI 4340, Cr-Mo-V, vs. yield strength for a broad sample of structural steels is~a& Ni-Cr-Mo-V, and XAR-30 quenched-and-tempered structural steels22.2». 29 of severe limitations begin to take welded structures frequently tend to plane strain, only to have further effect, thereby limiting our knowledge contain flaws, and the significance of crack movement arrested by localized of plane strain fracture behavior in these flaws can only be judged in plasticity. Therefore, it must be em this subregion. Relatively few high- terms of their potential to cause fail phasized that fracture in this category strength steels, which can generate ure. of steels is highly dependent on thick valid Kic values, are sufficiently tough Critical flaw sizes can be related to ness. As yield strength decreases, plane strain conditions only prevail in to be above the 1.0 ratio line. There the KTc/o-ys ratio by applying the fore, one must turn to lower strength principles of linear-elastic fracture very heavy sections. steels to define this subregion. But, as mechanics. Fig. 11 illustrates such a Finally, for ratios greater than 1.5, yield strength decreases, thickness re relationship for the case of the long, this plane strain, linear-elastic analysis quirements increase rapidly, thereby shallow flaw in plane strain. A family generally does not apply. Plane strain imposing experimental difficulties for of curves is shown corresponding to elastic fracture can only occur in ex conducting such tests. The upper limit nominal stress levels expressed as tremely heavy sections or at low tem ratio of 1.5 for this region is arbitrary fractional values of the yield strength peratures.44 Therefore, the critical but appears to be well justified on the stress. flaw sizes indicated, large as they are, basis of current technology. However, Relating this figure to the three are only extremely conservative esti it must be emphasized that this region cases presented in the previous sec mates. is not well defined, and as the ductile tion, several observations can be fracture region is approached, a close made. First, for ratios less than 0.5, Comments on Design Procedures ly spaced family of ratio lines ranging critical flaw sizes are very small, ex The following comments are from values slightly greater than 1.0 cept under very low stresses. Further, to infinity is likely to emerge.43 offered as guidelines as to the manner except in thin sections, plane strain in which a broad knowledge of crack conditions apply for these steels and Finally, Case 3 consists of steels propagation and fracture characteris little can be done mechanically to gain with the lowest level of fracture tics can influence fatigue design added toughness. toughness, where the KIe/o-ys ratio is procedures. less than 0.5. Generally this condition For ratios between 0.5 and 1.5 flaw is avoided wherever possible, and only sizes remain fairly small, but for the Case 1: Kic/
WELDING RESEARCH SUPPLEMENT I 493-s inherently affected by yield strength, STRUCTURAL STEELS ratory specimens where plastic strains lower working stresses in lower only develop in the final stages of strength steels will almost invariably failure. For low-cycle fatigue crack result in slower growth rates. propagation involving gross plastic de
The lower strength steels which re METALLURGICAL formation over the entire fatigue life, quire heavier sections to maintain low QUALITY empirical approaches based on gross stresses present the fewest fatigue strain are required. The authors con problem areas. Here, there is a large ducted strain cycle crack propagation amount of material to contain crack tests on several pressure vessel steels growth. Growth rates tend to be slow, and showed that an arbitrary index of and high toughness, which permits fatigue life based on their studies cor large flaw sizes, is easily attainable. YIELD STRENGTH (KSI) responded reasonably well with the Also, these steels present the fewest Fig. 9—Schematic illustration of the ef fatigue life of full-scale model vessels welding problems to complicate this fects of yield strength and metallurgical of the same materials.46 McEvily47 favorable picture. quality on fracture toughness for struc has proposed a strain-intensity factor tural steels Service experience justifies these approach to fatigue crack propagation observations. Barring unusual circum which is not limited to elastic loading, stances, low-strength welded struc and are based on plastic fatigue crack and this approach seems well worth tures seldom fail by fatigue. Design initiation. The fact that this approach further investigation. procedures for common, low-strength works as well as it does simply sub welded steel structures are well stantiates the fact that fatigue is not a Case 2: 1.5 > Ktc/
TECHNOLOGICAL LIMIT STRUCTURAL STEELS 8000
CASE I: K , /o- > 1.5 7000 I( ys DUCTILE PLASTIC FRACTURE LARGE CRITICAL FLAWS 6000 TRANSITION TEMPERATURE IMPORTANT
5000 LD CC LU 4000 < 2000 000 500 BRITTLE ELASTIC INSTABILITY FRACTURE SMALL CRITICAL FLAWS ENVIRONMENTAL SENSITIVITY IMPORTANT 80 I00 I20 I40 I60 I80 200 220 YIELD STRENGTH (KSI) Fig. 10—Ratio Analysis Diagram for structural steels. The diagram is divided into three regions on the basis of the Kic/cry, ratio, and the prominent characteristics of steels in each region are listed 494-s I OCTOBER 1970 using higher strength materials is such steels as 4340, H-ll, D6AC, and Summary 18Ni maraging. These steels are in smaller section sizes, therefore less There is an abundance of evidence capable of gross plastic deformation, material is available to contain cracks. to support the premise that fatigue and they fracture elastically under The concomitant penalty of higher failure in large, complex welded struc plane strain conditions in plate thick strength materials is lower toughness tures is predominantly caused by ness sections. However, an analysis of and smaller critical flaw sizes, thereby crack propagation originating at pre their fatigue crack propagation behav severely limiting the upper boundary existing defects which are too small ior is complicated by the fact that condition for fatigue failure. Howev and too numerous to totally eliminate. they are hypersensitive to environ er, many new designs are contemplat It has been proposed that fatigue fail ment. Even the water vapor content ing new steels in the 100 to 180 ksi ure by crack propagation can be of laboratory room air is an important yield strength region, which will in viewed as a growth process operating factor in determining their fatigue be clude a great number of Case 2 steels. between fixed boundary conditions, havior. Successful fatigue applications Fatigue is likely to take on a new i.e., the initial and terminal flaw sizes. of these steels require that they be urgency in the failure-safe design of Fatigue design procedures must take coated or otherwise protected to re such high-strength structures. all three aspects into account. The move their surface from the atmos Linear-elastic fracture mechanics problems of determining initial flaws works rather well for analyzing crack phere. 4 48 come under the scope of quality con propagation and the potential for Tiffany, et al., ' have proposed trol and nondestructive inspection. fracture in Case 2 steels, especially if and extensively applied a sophisticated 33 Crack growth and failure are materi heavy sections are involved. In thin plane strain linear-elastic fracture als problems whose nature and seri ner or tougher steels within this cate mechanics approach to low-cycle fa ousness vary widely with variations in gory, fracture mechanics continues to tigue and fracture in plate section yield strength and fracture toughness. be the best analytical approach for ultrahigh-strength materials. It in It has been shown that the baseline crack propagation, but its application cludes all three elements of the fatigue fatigue crack propagation characteris to fracture becomes less well under failure process, the initial flaw severi tics of steels are not fundamentally stood as a departure from plane strain ty, the conditions for terminal failure, influenced by yield strength, although instability fracture is achieved. Also, and the duration of the growth proc significant variations can be obtained "leak-before-fracture" situations can ess linking those boundary condi at any strength level. Fracture charac be attained for many applications with tions. However, despite the ample an teristics of steels do, however, respond these steels, especially for ratio values alytical tools available to deal with strongly to yield strength with signifi between 1.0 and 1.5. In other words, ultra high-strength steels, the actual cantly decreased toughness at higher large visible flaws and considerable physical problems involved severely yield strength levels. localized plastic deformation will limit their application. precede terminal fracture. These conditions may preclude a 6.0 physical catastrophe. In many in h—2 c —1 i stances, however, such as industrial pressure vessels, a through crack rep resents an economic catastrophe. a/2c 3 0.10 5.0 Crack growth in these steels may pro (LONG, SHALLOW FLAW) ceed at surprisingly rapid rates, espe cially if assisted by corrosion. Further, UJ many steels in this category are prone rr to welding problems. So, even though o 4.0 a high degree of assurance against < brittle fracture can be achieved in rr these steels, failure by fatigue crack U_ rr growth remains a serious problem o which must be dealt with in a design u_ basis and cannot be precluded by at 3.0 - taining improvements in basic fatigue o. crack propagation characteristics of UJ materials. Q Case 3: K,c/a„ < 0.5 2.0 - I 6 6 The failure mode for steels in this / / category will be brittle elastic instabil ity fracture. The critical flaw sizes will < be very small. The application of this £ i.o - category of steels in large welded rr structures poses extreme difficulties o from a fatigue and fracture standpoint and should be avoided except where extreme necessity imposes severe ^i^-^^^^--- i i strength-to-weight ratio requirements. 0.5 .0 1.5 2.0 Although these steels are the most difficult to apply in service, they are v*ys RATIO the easiest to deal with from an an Fig. 11—Fracture mechanics relationships for critical flaw depths of long, alytical point of view. Hence, there is shallow flaws in plane strain vs. the Ki^la^ ratio. The family of curves an abundance of literature on fracture represent stress levels expressed as fractional values of the yield strength mechanics experiments conducted on stress WELDING RESEARCH SUPPLEMENT ! 495-s As a means of categorizing these um on Metallurgical Aspects of Fatigue Fractographic Investigation of the Effect Fracture, Materials Engineering Congress, of Environment on Fatigue-Crack Propa problems quantitatively, it has been Philadelphia, Pennsylvania, 13-16 October gation in an Ultrahigh-Strength Steel," proposed that the Kl0/o-ys ratio can 1969. ASM Transactions Quarterly, Vol. 60, No. provide significant information rele 12. Grosskreutz, J. c.. "A Theory of 3, September 1967, p. 279. Stage II Fatigue Crack Propagation," Air 31. Wei, R. P., Talda, P. M., and Li, vant to fatigue design procedures. Force Materials Laboratory, Technical Re Che-Yu, "Fatigue Crack Propagation in Further, the use of the NRL Ratio port 64-415, March 1965. Some Ultrahigh-Strength Steels," Fatigue 13. Laird, Campbell, "The Influence of Crack Propagation, ASTM STP 415, Amer Analysis Diagram permits these deter Metallurgical Structure on the Mechanisms ican Society for Testing and Materials, minations to be made on the basis of of Fatigue Crack Propagation." Fatigue 1967, p. 460. two relatively simple engineering Crack Propagation, ASTM STP 415, Amer 32. Crooker, T. W., and Lange, E. A, ican Society for Testing and Materials, "Corrosion-Fatigue Crack Propagation tests. The value of the KIC/o-ys ratio 1967, p. 131. Studies of Some New High-Strength Struc thereby obtained can be used to deter 14. Manson, S. S., Thermal Stress and tural Steels," ASME Paper No. 69-MET-5. Low-Cycle Fatigue, McGraw-Hill, 1966. mine the probable severity and poten 33. Clark, Jr., W. G, and Wessel, E. T, 15. Weiss, Volker. "Analysis of Crack "Application of Fracture Mechanics Tech tial for fatigue failure, the probable Propagation in Strain-Cycling Fatigue," nology to Medium-Strength Steels," (to be Fatigue—An Interdisciplinary Approach, published in ASTM STP 463, pending). mode of terminal failure to be de Proceedings, 10th Sagamore Army Materi signed against, the most applicable als Research Conference, Syracuse Univer 34. Wei, R. P., and Landes, J. D.. "Correlation Between Sustained-Load and analytical approach, and critical flaw sity Press, 1964, p. 179. 16. Lehr, Kenneth R., and Liu, H. W., Fatigue Crack Growth in High-Strength size estimates. "Fatigue Crack Propagation and Strain Steels," Materials Research and Standards, Cycling Properties," International Journal Vol. 9, No. 7, July 1969, p. 25. It is intended that the broad scope of Fracture Mechanics, Vol. 5,. No. 1, 35. Crooker, T. W., and Lange, E. A., of fatigue and fracture information March 1969. p. 45. "Low Cycle Fatigue Crack Propagation brought together in this paper will 17. Johnson, H. H., and Paris, P. C, Resistance of Materials for Large Welded "Sub-critical Flaw Growth" Engineering Structures," Fatigue Crack Propagation, thereby serve to influence fatigue de Fracture Mechanics, Vol. 1, No. 1, June ASTM STP 415, American Society for sign procedures, most especially for 1968, p. 3. Testing and Materials, 1967, p. 94. design situations involving the use of 18. Gurney. T. R., "An Investigation of 36. Throop, J. F., and Miller, G. A., the Rate of Propagation of Fatigue Cracks "Optimum Fatigue-Crack Resistance," newer, higher strength steels. in a Range of Steels," The Welding Insti Symposium on the Achievement of High tute Research Report No. E18/12/68, De Fatigue Resistance in Metals and Alloys, cember 1968. 72nd Annual, ASTM, Atlantic City, New Jersey, 22-27 June 1969. Acknowledgments 19. Schwab, R. C, "Use of Tapered Double Cantilever Beam Specimens for 37. Dahlberg, E. P., "Fatigue-Crack The authors are grateful to the Fatigue Crack Growth Studies," ASME Propagation in High-Strength 4340 Steel in Office of Naval Research for their Paper No. 69-MET-16. Humid Air," ASM Transactions Quarterly, 20. Brothers, A. J., "Fatigue Crack Vol. 58, No. 1, March 1965, p. 46. financial support of this work. Growth in Nuclear Reactor Piping Steels," 38. Pellini, W. S., "Advances in Fracture U.S. Atomic Energy Commission Research Toughness Characterization Procedures in and Development Report GEAP-5607, Quantitative Interpretations to Fracture- References March 1968. Safe Design for Structural Steels," Weld 21. Mowbray, D. F.. Andrews, W. R., ing Research Council Bulletin, No. 130, 1. Signes, E. G., Baker, R. G., Harrison, and Brothers. A. J., "Fatigue-Crack May 1968. J. D.. and Burdekin, F. M., "Factors Growth-Rate Studies of Low-Alloy Press- 39. Puzak, P. P.. and Lange. E. A, Affecting the Fatigue Strength of Welded sure Vessel Steels," ASME Transactions, "Standard Method for the 1-Inch Dynamic High Strength Steels," British Welding Journal of Engineering for Industry, Vol. Tear Test," NRL Report 6851, February Journal, Vol. 14, No. 3, March 1967, p. 108. 90, Series B, No. 4, November 1968, p. 648. 13, 1969. 2. Kooistra, L. F., Lange, E. A., and 40. Freed, C. N. and Goode, R. J., "A Pickett, A. G., "Full-Size Pressure Vessel 22. Brothers, A. J., and Yukawa, S., "Fatigue Crack Propagation in Low-Alloy Correlation Between Fracture Toughness Testing and Its Application to Design," Test Procedures for Ferrous Alloys," Pres ASME Transactions, Journal of Engineer Heat-Treated Steels." ASME Transactions, Journal of Basic Engineering, Vol. 89 sure Vessel Technology, Part II, Materials ing for Power, Vol. 86, Series A, No. 4, and Fabrication, ASME, 1969, p. 723. October 1964, p. 419. Series D, No. 1, March 1967. p. 19. 23. Clark, Jr., W. G., "Subcritical Flaw 41. Brown, Jr., W. F., and Srawley, 3. Srawley, J. E, and Esgar, J. B.. J. E., Plane Strain Crack Toughness Test "Investigation of Hydrotest Failure of Growth and Its Effect Upon the Fatigue Characteristics of Structural Alloys," En ing of High Strength Metallic Materials, Thiokol Chemical Corporation 260-Inch- ASTM STP 410, American Society for Diameter SL-1 Motor Case," NASA TM gineering Fracture Mechanics, Vol. 1, No. X-1194, January 1966. 2, August 1968, P. 385. Testing and Materials, 1966. 4. Tiffany, C. F., and Lorenz, P. M., 24. Clark, Jr., W. G., and Wessel, E. T., 42. Pellini, W. S., and Puzak, P. P, "An Investigation of Low-Cycle Fatigue "Influence of a Synthetic Seawater Envi "Fracture Analysis Diagram Procedures Failures Using Applied Fracture Mechan ronment on the Fracture Behavior of HP for the Fracture-Safe Engineering Design ics," Air Force Materials Laboratory Tech 9-4-25 and HP 9^-20 Alloy Steels," Materi of Steel Structures," NRL Report 5920, nical Documentary Report ML-TDR-64-53, als Performance and the Deep Sea, ASTM March 15, 1963. May 1964. STP 445, American Society for Testing and 43. Pellini, W. S., "Evolution of Engi 5. Week, Richard, "How Dangerous Are Materials, 1969, p. 93. neering Principles for Fracture-Safe De Weld Defects?," Metal Progress, Vol. 92, 25. Crooker, T. W., and Lange, E. A., sign of Steel Structures," NRL Report No. 3, September 1967, p. 93. "The Influence of Salt Water on Fatigue 6957, September 23. 1969. 6. Christopher, P. R., and Crabbe, Crack Growth in High-Strength Structural 44. Loss, F. J.. and Pellini. W. S.. D. R.. "A Discourse on Factors Which Con Steels." Effects of Environment and Com "Dynamic Tear Test Definition of the trol Fatigue Behavior in High Yield Low plex Loading History on Fatigue Life, Temperature Transition from Linear Elas Alloy Steel Structures," Ocean Engineer ASTM STP 462, American Society for tic to Gross Strain Fracture Conditions," ing, Vol. 1, No. 5, July 1969, p. 497. Testing and Materials (pending). ASME Transactions, Journal of Basic En 26. Barsom, J. M., "Investigation of gineering, Series D, Vol. 91, No. 1, March 7. Rice, J. R, "Mechanics of Crack Tip 1969, p. 108. Deformation and Extension by Fatigue," Subcritical Crack Propagation," Ph.D. Fatigue Crack Propagation, ASTM STP Thesis, University of Pittsburgh, 1969. 45. Harrison, J. D., "The Analysis of 415, American Society for Testing and 27. Carman. C. M., and Katlin, J. M., Fatigue Test Results for Butt Welds with Materials, 1967, p. 247. "Low Cycle Fatigue Crack Propagation Lack of Penetration Defects Using a Frac 8. Weiss, V., and Yukawa, S., "Critical Characteristics of High Strength Steels," ture Mechanics Approach," BWRA Report Appraisal of Fracture Mechanics," Frac ASME Transactions, Journal of Basic En No. E/13/67, December 1967. ture Toughness Testing and Its Applicar gineering, Vol. 88, Series D, No. 4, Decem 46. Crooker, T. W., and Lange, E. A., tions, ASTM STP 381, American Society ber 1966. "Low Cycle Fatigue Crack Propagation in for Testing and Materials, 1965, p. 1. 28. Miller, G. A., "The Dependence of A201B, A302B, and A517F Pressure Vessel 9. Paris, P. C, and Erdogan, F., "A Fatigue-Crack Growth Rate on the Stress Steels," WELDINO JOURNAL, 46 (7), Research Critical Analysis of Crack Propagation Intensity Factor and the Mechanical Prop Suppl., 322-s (1967). Laws," Journal of Basic Engineering, erties of Some High-Strength Steels," ASM 47. McEvily, A. J., "Fatigue Crack ASME Transactions, Vol. 85, Series D, No. Transactions Quarterly, Vol. 61, No. 3, Growth and the Strain Intensity Factor," 4, 1963, p. 528. September 1968, p. 442. Air Force Conference on Fatigue and 10. Paris, P. C, and Wei, R. P., "Frac 29. Anctil, A. A, and Kula, E. B., Fracture of Aircraft Structures and Mate ture Mechanics in Fatigue," ASM Symposi "Effect of Tempering Temperature on rials, Miami Beach, Florida, 15-18 Decem um on Engineering Aspects of Fatigue, Fatigue Crack Propagation in 4340 Steel," ber 1969. 1969 Materials Engineering Congress, Phil Effects of Environment and Complex 48. Tiffany, C. F., and Masters, J. N., adelphia, Pennsylvania, 13-16 October 1969. Loading History on Fatigue Life, ASTM "Applied Fracture Mechanics," Fracture 11. Johnson, R. E., "Some Observations STP 462, American Society for Testing and Toughness Testing and Its Applications, of Cyclic Crack Propagation from a Frac Materials, (pending). ASTM STP 381, American Society for ture Mechanics Viewpoint," ASM Symposi 30. Spitzig, W. A., and Wei, R. P., "A Testing and Materials, 1965, p. 249. 496-s OCTOBER 1970