457 October 197 6
ma0aL4Jmcsoo ma!P@mu DEP.A.R'I"NIEN'I" OIF 'I"R.A.NSPOR'I".A.'I"ION
ACOUSTIC EMISSION, FATIGUE, A..l\!D CRACK PROPAGATION
Theodore Hopwood II
111VIIIflll fJI RIIIARtll I.IXIIItiTflll JULIAN CARROLL JOHN C. ROBERTS GOVERNOR SECRETARY OF TRANSPORTATION
COMMONWEALTH OF KENTUCKY
DEPARTMENT OF TRANSPORTATION
BUREAU OF H/GHWA YS
DIVISION OF RESEARCH 5.33 SOUTH LIMESTONE LEXINGTON, KENTUCKY 40508 COMMONWEALTH OF KENTUCKY
DEPARTMENT OF TRANSPORTATION JULIAN M. CARROLL JOHN C. ROBERTS BUREAU OF HIGHWAYS GovERNOR SECRETARY JOHN C. ROBERTS COMMISSIONER
Division of Resea.rch 533 South Limestone - Lexington, KY 40508
November 5, 1976 H.3.33
MEMO TO: G. F. Kemper State Highway Engineer Chairman, Research Committee
11 SUBJECT: Research Report No. 457; "Acoustic Emission, Fatigue, and Crack Propagationj KYP·72-33; HPR·PL·l(l2), Part lii·B
The state of knowledge pertaining to fatigue and the nucleation and growth of cracks in new steels is given in design textbooks and treatises. The state of knowledge pertaining to the detection or discovery of flaws, defects, cracks, or Lnsidious, fatigue damage Ln aged, steel bridges remaLns more formative and less practicable. It would be desirable to have tell·tale lights or alarms on bridges to forewarn of weakening 11 in any part. Presently, the closest approach to a warning system is acoustical monitoring. Steels do cry out in pain 11 when over~stressed; they do not cry very loudly. listening is done through contacting wJcrophones and amplifiers. A crack grows because of high stress at the apex. Isolating a noise artd locating the point of origin is much more complicated. Apparently, fatigue damage remaiilS insidious and not measurable until a crack develops. This is our conclusion from our investigations thus far. We have reported previously on acoustic monitoring following welding (Report 393; June 1974) and on fatigue analyses (Reports 251,275, 318, 323, and 411). An additional report will cover acoustic monitorLng of constrained, welded joints.
Respectfully~ o_:_ ,.e; .d!! /"7:~. Havens Director of Research.
sh Enclosure cc 1s: Research Committee Technical Report Documentation Page
No. 1. Report No. 2. Government Accession No. 3. Recipient's Cotalog
4. I itle and Subtitle 5. Report Date October 1976 6. Performing Organization Code Acoustic Emission, Fatigue, and Crack Propagation
8. Performing Organiz:ation Report No. 7. Authorl s) Theodore Hopwood II 457 10. (TRAIS) 9. Performing Orgonizotion Nome and Address Work Unit No. Division of Research No. Kentucky Bureau of Highways 11. Contract or Grant KYP-72-33 533 South Limestone I.e xington Kentucky 40508 13. Type of Report and Period Covered 12. Sponsoring Agency Nome end Address
lr'tterim 14. Sponsoring Agency Code
1s. Supplementary Note$
Study Title: Evaluation of the Fatigue Life of Critical Members of Major Bridges
16. Abstract I
Acoustic emlSSlOTI was used in conjunction with tensile tests to evaluate the condition of structural steel specimens subject to various tensile fatigue lives. The results indicate that the acoustic emissions detected were the result of plastic deformation. There was no apparent relationship between fatigue history of the steel specimens and the amount of plastic deformation they can accommodate. Further tests revealed that acoustic emission has the physical capability of detecting cracks on large structural steel members. This may prove beneficial for the comprehensive testing of steel bridges.
17. Key Words 18. Distribution Stctemenf acoustic emission pearlitic steels brittle fracture stress·corrosion cracking fatigue s.tress·intensity martensitic steels nondestructive evaluation 21. No, of P cges 22. Price 19. Security C!ossif. (of this report) 20. Security Clcssif. (of this pcge)
Form DOT F 1700.7 18-721 Reproduction of completed page authorized Research Report 457
ACOUSTIC EMISSION, FATIGUE, AND CRACK PROPAGATION
Interim Report KYP-72-33, HPR-PL-1(12), Part Ili-B
By
Theodore Hopwood II Research Engineer
Division of Research Bureau of Highways DEPARTMENT OF TRANSPORTATION Commonwealth of Kentucky
The contents of this report reflect the views of the author who is responsible for the facts and the accuracy of the dat-a presented herein, The contents do not necessarilY reflect the official views or po-licies of the Kentucky Bureau of Highways, This report does not constitute a standard, specification, or regulation.
October 197 6 INTRODUCTION 25, 1940. It had been in service for 5 years. The temperature was about 7 F (-14 C). The bridge did not In the past 30 years, at least 26 major steel bridges collapse upon cracking. throughout the world have suffered insidious fractures While the trusses were being fabricated, distortion resulting in their partial or complete collapse. When this occurred due to weld contraction. The welders corrected study was initiated in 1972, five years had passed since the alignment as the truss progressed across the canal. the Silver Bridge collapse at Point Pleasa.L1.t, West The distortion was probably caused by poor weld Virginia. During the course of this study, the Osage detailing. The steel used in several of the bridges was River Bridge at Warsaw, Missouri, collapsed June 9, found to have high sulphur and phosphorous contents. !975; and the US 75 bridge over the South Canadian Notched impact tests revealed that the steel had low River in Oklahoma collapsed May 22, 1976. impact toughness at the temperatures at which the The first wrought-iron bridge was built in England bridges failed. Cleavage cracks were not restricted to over 200 years ago and is still in use. Wrought iron was points through or adjacent to welds. The failures were used predoiTlinently in early metal bridges and continued not attributable to poor steel weldability. to be widely used for this purpose through the end of The Duplessis Bridge (see Figure 2) at Quebec, the 19th century. The Roebling Suspension Bridge over Canada, was completed in 1947. The bridge was of the Ohio River at Covington, constructed of wrought welded plate-girder construction with plate thicknesses iron, masonry, and timber, has served since 1865. to 2 1/2 inches (64 mm). The bridge had six, 180-foot Pneumatic steelmaking, jointly discovered in the (59-m) and two, !50-foot (49-m) spans. Wilen the bridge mid-19th century by Henry Bessemer in England and was 27 months old, two cracks were found on flange William Kelly at Eddyville, Kentucky, led to the use plates near butt welds, traveling toward the web. of steel in long bridge spans .. However, steel was not Inspection revealed the cracks had probably been in the used in an American bridge until 1880. At the turn of girders prior to erection. Repairs were made by riveting, the century, open-hearth steelmaking began to supplant and all tension joints were reinforced with riveted plates. the Bessemer-Kelly process. Now, the open-hearth A year after repairs, traffic was suspended, and a l 0-day process is gradually being replaced by the basic oxygen inspection was performed on the bridge. The repairs process. were reported to be satisfactory. However, the west - Most metal bridges in service today are made of portion of the bridge collapsed under its own weig..ht steeL \Vh.ile most of these bridges have performed 2 weeks later. The temperature at the time of failure satisfactorily, a few recent failures are worthy of was -30 F (-37 C). Investigation after the accident discussion. Prior to World War II, some 50 Vierendeel revealed that improper material selection was responsible truss bridges were built over the Albert Canal i11. for the failure. The thick sections (ordered to ASTM Belgium. The Vierendeel truss is a welded, through-type A 7) were found to be of rimming quality, unsuitable truss, without diagonals; stiff posts and knees connect for welding, and showed extensive segregation of carbon the lower and upper chords. Structural members and sulphur. Charpy tests revealed low impact toughness consisted of !-beams, plates, or mixed !-beams and plates (3-6 ft-lb (4-8 J) at 100 F (37 C)) (1). of a Belgium, standard Bessemer steeL Plate thicknesses The Kings Bridge at Melbourne, Australia, was built varied up to 2 1/2 inches (64 mrn). The first failure in 1961. The structure consisted of four lanes of occurred at Hassell, March 14, 1938 (see Figure 1). The plate-web girders 100 feet (30.5 m) long and 5 feet (1.5 bridge had been in service 14 months. It had a span m) deep and a reinforced concrete deck. In July 1962, of 244 feet (74.5 m). The bridge collapsed under the one span of the bridge collapsed under the weight of load of a tramcar and some pedestrians about 6 minutes a 45-ton (41-Mg) truck. The bridge sagged 1 foot (0.3 after ·a loud report of the first crack. Witnesses said the m); further collapse was prevented by the concrete deck. failure occurre-d in the lower chord, causing the arch Inspection revealed that all girders had brittle fractures to absorb the load and eventually collapse. The in nearly identical locations. The cracks ran from the temperature at the time of the fracture was -4 F (-20 heat-affected zone of the weld at the upper flange and C). A second truss, at Herenthals-Oolen, failed on March traversed through the parent metal to the lower flange. 19, 1940, after 3 years of .service. Cracking was It was also evident that failures had occurred accompanied by three loud reports. Five hours later, a sequentially over the 15-month period since the bridge train passed over the bridge without causin_g it to opened. The final fracture caused the bridge to collapse. collapse. One crack was found to be 7 feet (2.1 m) long This failure was the result of improper material and open l·inch (25·mm). The temperature at failure specification and inspection. To reduce the dead weight was 7 F (-14 C). A bridge at Kaulille failed January of the bridge, a high-strength steel was ordered to an Belgium (9). Figure 1. Failure of Vierendeel Truss Bridge at Hasselt,
Canada. 2. Failure of Duplessis Bridge at Quebec, Figure Wiley, 1957, (Parker, E. R.; Brittle Behavior of Engineering Structures; p 260)
2 OF STEEL old war-time specification intended only for emergency F~4.CTURE BEHAVIOR failure, e.g., buckliilg and use. Impact tests were conducted on the steel prior to Several modes of metal have been effectively dealt erection. Later investigation revealed that the tests were excessive plastic deformation, design rules. improperly performed. As a result of this failure, the with by elastic theory and empirical which were supposedly steel standard was superseded with new alloys having However, a few steel structures in a brittle fashion, greater toughness (2). Proper il'lSpection of the structure ductile have failed unexpectedly knowledge of both would have detected the beam failures months prior to often at very low stresses. A and engineering mechwics are failure. mechanical metallurgy Some insight to The Silver Bridge was built at Point Pleasant, West required to understand this problem. can be gained 'by viewing a Virginia, in 1927. It contained eyebars made of a mechanical metallurgy curve for an unnotched, low·carbon, standard ASTM A 7-24 steeL The steel was basically typical stress-strain as shown by llne 0-A-B-C-D h'1 Figure an AISI 1060 carbon steel, quenched and tempered to tensile specimen of the strain in Region a' is considered elastic produce a martensitic structure. The bridge had been 4. Most Hooke's law applies in this region. As the in service 40 years (see Figure 3) when it suddenly (recoverable). strained further, stress concentrations and collapsed under a load of 28 vehicles, including several specimen is of grain continuity force some grains concrete trucks. The temperature at the time of collapse the requirements plastic deformation before the upper yield was 32 F (0 C). The failure was caused by the brittle to undergo Figure 4. In Region b, further straining fracture of an eyebar due to stress corrosion and(or) point A of yielding. One or two bands of fatigue corrosion. Laboratory examination revealed that results in heterogeneous deformation arise and propagate the crack initiated at corrosion pits in the eye. The localized plastic until the whole gage length has undergone eyebar fracture allowed a paired eyebar to slip from its intermittently Point B. pin, initiating the collapse of the entire structure (3). plastic yielding at of Figure 4, the specimen has Failure must be partially ascribed to design and materials In Region c of cross section along selection. The design allowed the failure of a single undergone a uniform reduction which is the maximum member to initiate failure of the entire bridge in a the gage length. After Point C, the tensile or ultimate catastrophic fashion. Furthermore, the design did not engineering stress (also called occurs in the provide complete access for inspection of these stress), localized deformation or necking with strain members. The eyebar material also lacked sufficient specimen. The engineering stress decreases shown in Region toughness for such a critical application, As a result of as the cross-sectional ar_ea decreases, as region becomes the subsequent investigation, a companion structure, the d. Actually, the material in the necked curve based on St. Maris Bridge, was closed by the State of West strain-hardened, .and a stress·strain would Virginia. instantaneous cross-sectional area at the neck show the stress increasing to the point of fracture. Under None of the bridges discussed above had been in normal test conditions, a mild (low·carbon) steel will service for more than 40 years. Each of the structures fracture in a ductile manner. As the specimen is sustained brittle fractures. Many bridges have been in deformed into Region d of Figure 4, voids form at the service between 80 and 100 years without suffering interface between the metal and hard impurities fatigue failures. Obviously, pure fatigue was not the (inclusions). With further straining, the voids connect prime cause of these past failures but was probably a into a crack at the center of the neck. The crack grows contributory mechanism in several cases. outwardly, in an irregular manner, normal to the tensile Methods of preventing fractures were not widely axis. At a critical size of crack, the fracture mode used at the time the failures occurred. The notched changes to shear at about a 45' angle. The resulting impact test, developed in 1903 by lzod, was not widely fracture at Point D has the classical cup-and·cone applied to structural 'teels until after the Second World appearance. War. The fabrication of bridges by welding was rather An inherently brittle steel, shown by stress-strain new when the Vierendeel failures occurred. Ultrasonic curve O·A' -B' of Figure 4, will not accommodate as testing was limited to the aircraft industry when the much plastic strain (Region c') as the ductile steel Duplessis Bridge was inspected. Little was known about (Regions c '"'d d). However, brittle steels usually have stress corrosion when the Silver Bridge was built. In higher yield and ultimate strengths than ductile steels. contrast, factors contributing to the Kings Bridge Often, brittle steels will not exhibit a definite, collapse were well understood at the time of discontinuous yield region. The lack of significant plastic construction. However, appropriate preventative action deformation prior to fracture indicates low ductility or was not pursued.
3 Figm-e 3. Failure of Silver Bridge at Point Pleasant, West Virginia. (The Courier-Journal and Louisville Times, December 16, 1967)
4 BRITTLE STEEL
DUCTILE STEEL
t D b
rJ) wrJ) et:: 1- U)
(.') ~ et:: w zw (9 z w
ENGINEERING STRAIN < __.,.
Figure 4. Typical Stress-Strain Curves for Ductile and Brittle Steels. that load or test conditions lead to fracture before the The ability to absorb energy in the plastic region specimen can r-edistribute stresses by plastic can be def:i11ed as toughness. It indicates how much work deformation. Brittle fracture is caused by the growth can be done on a material before it fractures. The brittle of a few large cracks or possibly the connection of steel shown in Figure 4 has less area under the curve rnicrocracks over a short time intervaL In either case, in the plastic region ( c') than the more ductile steel research indicates that these cracks are usually not (Regions c and d). Therefore, the ductile material can present in the material before stressing but are produced accommodate more plastic work before failure occurs. by the deformation process. There are three steps in Brittle steels have two distinct disadvantages: they brittle fracture: plastic deformation, crack initiation, cannot accommodate thermal and mechanical shock; and crack propagation ( 4). In the past, engineers but, most importantly, they cannot relieve stress attempted to classify service fractures by comparison to concentrations. A ductile failure of a structure implies the general appearance of tensile specimens fractured in that the designer failed to accommodate the normally laboratory tests. There are several disadvantages to this applied stresses or that the structure was overloaded. approach. An overloaded specimen may fail in a ductile The ductile fracture of a structure can be gradual. A manner and show little elongation or necking. brittle fracture is rapid, usually with disasterous Sometimes it is difficult to classify material failure by consequences. Brittle fractures can occur at normally fracture mode (shear or cleavage) because of complex safe operating stresses as the result of localized design variations in stress during fracture. The best way of and( or) construction errors or defects in materials. Steels classifying a fracture is by microscopic inspection of the usually exhibit intermediate behavior between cracked surface. Due to the void formation and large brittleness and ductility, depending on many factors. amount of local grain deformation, a ductile fracture Because of composition, a steel can be brittle in every will appear dull gray and fibrous. A brittle fracture operating environment. However, the ductility or contains transgrariular cracks which give the surface a brittleness tends to V3IY with atmospheric, geometric, granular and crystalline appearance. and loading conditions. 5 of plain pearlitic steels. Uniform Two broad types of steels are used in all steel times less than that the effective load-bearing cross section bridges: the pearlitic type and the quenched and corrosion reduces especially one that has been tempered martensitic type. Pearlitic steels consist of of a structural member, an extended period of time. Uniform layers of a ductile, iron-rich matrix (ferrite) between in service for of eyebars on the thin lamellae of a brittle iron carbide (cementite) (see corrosion led to the reinforcing Ohio River at Newport. Figure 5 ). The ductility of these steels depends on the Central Bndge (7) over the insidious, especially if spacing of the carbide lamellae and dec'reases with a Localized corrosion can be very in the presence of coarsening of the pearlite structure (5). The quenched it occurs as intergranular corrosion This usually occurs and tempered martensitic steels (see Figure 6) have a applied or residual tensile stresses. corrodant (8). fine·grained structure which shows some degree of when a caustic is the .of a structural member spllerodization caused by the tempering operation. An The ductile behavior of stfess. Some multiaxial stress Interconnected film of E·carbides at martensite plate and depends on its state loading components can cause a twinned boundaries can form upon quenching, providing states having tensile stresses and strains below an easy path for crack propagation and leading to brittle structure to fail at applied uniaxial tensile tests. Most behavior. This type of steel, if ineffectually tempered, values determined by simple uniaxial or biaxial can also have a brittle microstructure containing structural designs employ problems encountered with plate-like, internally twinned martensite. stresses to avoid the discontinuities in a Of all the alloying elements, carbon shows the complex loadings. Geometric can also promote greatest increase in strength and the greatest loss of structure, such as holes and notches, multia.xjal stresses which toughness and ductility. Other elements having great brittle behavior by creating t1ow. t:lllbrittling effects are nitrogen, oxygen, hydrogen, restrict plastic raise the temperature at plwsplwruus, and sulphur (6). Sulphur, phosphorous, Notches will appreciably from ductile to brittle behavior. unJ nitrugcn are found in relatively larger concentrations which a steel changes If a round specimen has a circumferential notch and in Bessemer steel compared to open-hearth and basic is loaded in tension, as shown in Figure 8, the remaining (slag) oxygen steels. The cmbrittling tffect of cross section must absorb the tensile load; the material phusplwrous and sulphur depends upon their adjacent to the notch carries rio axial load. Whereas the concentration along gruin boundaries (4). Non~metallic effective cross section tends to constrict, due to inclusiu!lS can have a great intlucnce (usually adverse) Poisson's effect, the unstressed material resists on ductility. Inclusions usually decrease tensile ductility, defom1ation, creating triaxial tensile stresses. The notch but improve low temperature toughness (6). also increases the longitudinal stress required for yielding One of the most important 'lcriables a[Tecting the and localizes plastic flow to material immediately ductility or steel is grai11 size. Both ductility and strength adjacent to the root of the notch (9 ). Whereas good incrc 6 Figure 5. Typical Microstructure of l?eaiitic SteeL Figure 6. Typical Microstructure of Martensitic Steel. 7 80 80 60 60 .,. .,. CLEAVAGE CLEAVAGE 40 40 F" F" Steel. Steel. 20 20 TEMP TEMP SHEAR SHEAR Structural Structural 0 0 Typical Typical for for 0~----~----J-----~----~----~-- 20 20 40 40 80 80 Curves Curves 1 1 ~ ~ w w w w Q_ Q_ z z ~60 ~60 Cfj Cfj I- <( <( 0:: 0:: 80 80 Temperature Temperature Transition Transition 60 60 7. 7. 40 40 --...- 0 Figure Figure F 20 20 / / TEMP TEMP 0 0 ob-----~-----L-----L----~L-----~- 101- 20 20 0 0 140 140 z z w w 0:: 0:: w w >- ro ro .....1 .....1 (!) (!) LL LL C/)3() C/)3() I- C.tJ C.tJ t b...,_-- --0------b • 0w (l)...l (I)<{ o:::o:::w i-W (()~ 5:2 b------b ~------~~!' ~1-...J(I) Wo::: >-LI.. ...Jw ww~0::: ::lEI 9 Later, Orowan and Irwin (cf 4) applied much of As the thickness, t, decreases, a + 0. However, due 1 Griffith 1s analysis to brittle fracture of metals. Irwin s to Poisson's effect, e}Y #: 0. Thls condition is· called two dimensional approach, called linear fracture plane stress. As t mcreases, ayy increases. Plastic analysis, considered that, to propagate an existing crack constraint exists in the Y direction, fi.IJ.d therefore S/y in a structure under a stress a, energy for crack ~ 0, resulting in a condition called plane strain. The extension must come from elastic strain energy ( 11). resulting affect on fracture behavior is shown in Figure The significant resistance to crack propagation comes 12. Above a certain thickness, there will be no decrease from localized plastic work (deformation) at the crack in G , the critical energy release rate. A limiting 0 tip. When the elastic energy exceeds the plastic work, condition of plane strain has been reached, and this the crack will enlarge. He also concluded that, if a value is called the critical plane-strain energy release rate critical rate of strain energy is released per unit of crack and is denoted as Grc- length (termed critical energy release rate G ), the crack The plastic zone preceeding a crack is shown in 0 1 will propagate in an unstable manner. From Griffith s Figure 13. A condition of plastic strain ( 1 1 MATERIAL Gc DUCTILE c Gc BRITTLE B . MATERIAL \G:r SEMI- c G:r EXTENSION------ A/ --~--. MATERIAL CRACK BRITTLE t w C) ~ z u ll:: fi X CJ) ~ z w 0 w ~ u I ll:: u "' ~ Strains. Strains. ' ' • • COND. COND. COND. COND. and and 0; 0; erzz erzz '"xx '"xx ;tO ;tO =0 =0 ~ ~ ' ' yy yy Stresses Stresses Ell' Ell' eryy eryy € € erzz' erzz' t-O t-O t-oo t-oo Ezz Ezz er'f'J=O er'f'J=O STRESS STRESS STRESS STRESS on on ' ' xx' xx' 1£)()(' 1£)()(' erxx erxx er er ~Cyy• ~Cyy• Thickness Thickness PLANE PLANE PLANE PLANE Plate Plate of of eryy eryy Effect Effect xx xx The The er er 11. 11. / / '>~-~-·~------, '>~-~-·~------, Figure Figure erzz erzz ', ', erzz erzz " " ', ', ern' ern' ern ern l l ...... ', ', I I -,, -,, // // r-- - / / ~, ~, " " / / / / / / / / 't 't ~------"~C>'- w w .. I· t '"' ~ u (!) I I I I u I (!) I I I I (.) (!) I I I I u ! (!) ---- 38l::lO::l NOISN31X3 >i8V'tl:::l 14 PLANE PLANE STRAIN STRAIN STRESS STRESS PLANE PLANE SURFACE SURFACE MIDPOINT MIDPOINT AT AT Plate. Plate. a a in in X-SECTION X-SECTION Through-Crack Through-Crack a a of of Front Front the the al al Zone Zone Plastic Plastic 13. 13. Figure Figure ~ ~ - o-L --- o-L FOR YIELD UNDER PLAN'E 1 STRAIN CONDITIONS I I I I I I I -I--.::--~-.::.-- o-L = o-YS FOR YIELD UNDER PLANE STRESS I CONDITIONS I I I I I rY PLANE STRAINH ! I I I rY PLANE STRESS figure 14. Gruph Showing the Result of a Higher aL at Yield. 16 Different categories of stress cycling are shown iJ: SUBCR!TICAL CRACK GROWTH displayed on a Unless the stress intensity Kc decreases, a steel Figure 17. Fatigue behavior is usually· should not fail under stress-vs-number-of-cycles (S-N) diagram as shown by 51 n.Jcture with very small flaws to approximately l x stresses less than Up (Equation 1) or ays· However, the Figure 18. There is a range up 6 (Jet that failures occur in low-stress environments 10 cycles where failure is expected at a finite number fatigue or indicates that subcritical crack growth is involved. There of C)''cles. Below a given stress called the Jre two types of subcritical crack growth: stress endurance limit, a seemingly infinite fatigue life is corrosion cracking and fatigue cracking. Stress corrosion expected. Data used in S-N diagrams are subject to a has two unfortunate results. It can both form and wide range of scatter, and the curves usually represent a crack. The only requirement for stress average values. 1_;ropagate corrosion cracking is a local tensile stress. Usually, an Many factors affect the fatigue properties of steels e_c 17 KISC KISC c c 1 K -- __ __ __ __ FRACTURE FRACTURE f f FAILURE FAILURE BY BY NO NO 3 3 Environment Environment 10 FAILURE FAILURE Corrosive Corrosive a a in in Time Time HOURS HOURS 2 2 with with IN IN 10 K K of of TIME TIME Variation Variation 15. 15. 1 1 10 Figure Figure ------.-- ------~ ------~ I I In In ~ ~ oc oc z z ~ ~ ~ ~ >- ~ ~ ~ ~ 1- :><: :><: 00 00 REGION I REGION IT REGION ill 10 1 THE PRESENCE OF MICRO DEFECTS I MICRO CRACK GROWTH ON COALESCENCE INTO IMACRO CRACK GROWTH TO :.e. DEBONDING AT SMALL INCLU-~ MACRO CRACK -DEFECTS SIZE> I in.(2.5 mm) ISTRUCTURAL FAILURE. CRACK SIONS, SLIP EXTRUSiON 5 I NTRU- I READILY OBSERVABLE SION ETC. DEFECT SiZE >.1 in I I ( 2.5 mm) I I I I I I I 10 (f) I I w I I I u I I ;f; I ;f; I I w N U) I FAILURE 51 lo-' ::.:: I u I Figure 16. Prucric~1l Rc.·l~1tion hcl\\l'L"n Crack Initiation, Propagation, and Failure. 19 STRESS STRESS MIN---- O" O" RANDOM RANDOM (4). (4). + + ~~------~~------ (/) (/) (/) (/) (/) (/) 1- Cycles Cycles TIME- Stress Stress STRESS STRESS Fatigue Fatigue Typical Typical REPEATED REPEATED 17. 17. + + (/) (/) (/) (/) ~~~------ (/) (/) 1- Figure Figure \ \ I I STRESS STRESS --O"MJN --O"MJN \ \ REVERSED REVERSED STRESS STRESS I I I I i i 1-- i i -COMPRESSION -COMPRESSION +TENSION +TENSION ~II ~II (1) (1) 1- 0 0 wll: CD :2: ::l z 0 0 0 0 0 0 0 0 0 0 0. o_ o. o. o_ @ 0 ...,0 0 0 '¢ "' Xii'IIIJ .D 'SS31d1S li'Jnv.!IXii'IIIJ lii'Niv.!ON 21 ~i II.. I Considering unflawed steel subject to cyclic as nucleation sites (see Figure 20). Inclusions can also stresses, two stages of fatigue crack growth, nucleation act as sites for initiation. The subsequent mechanism and propagation, usually occur before failure. A material of Stage I crack growth is not well understood. II· with good resistance to nucleation does not have to be Striations in Stage I are usually not definable. Stage I ductile or tough. The fatigue limit of steel in reversed cracks propagate along common slip planes, which loading has been approximated as one-half of the tensile explains their 45° orientation to the applied stress. strength; a better basis is the proportiOnal limit. The Stage II crack growth can be deduced from the stronger steel, considered brittle in Figure 4, has a higher fracture surface containing striations. While slightly proportional limit than the ductile steeL Therefore, in different striations can be generated by different the absence of a geometrical flaw at a given stress stresses, a single mechanism can be used to describe .II I; between points A and A' in Figure 4, brittle steel is growth (see Figure 21 ). On application of a tensile stress, more resistant to fatigue crack nucleation than ductile a double-notched crack opening occurs -M with slip at steel (20). the notches being 45° to the crack front. At maximum Three size effects must be considered in the fatigue strain, the crack is blunted by plastic flow preventing of steels. One is a statistical effect. A larger specimen further crack growth. On closure (compressive load), the has more surface area, which is the preferred site of slip at crack notches is blunted and a new area at the rnicrocrack formation. Because of this, a large spech11en crack tip is folded into the plane of the crack, creating has a statistically greater probability of forming a fatigue another pair of notches (22). Each striation represents crack. Another factor is related to the replication of crack grmvth per load application, and a correlation the same microstructure in larger and smaller structural between fatigue crack growth (Stage II) and the stress sections. Large sections have lower cooling rates which intensity K is possible (23): can affect the microstructure of the steel. The third factor is notch size, which is related to the stress·strain dc/dN = C(I'IK)m 5 gradient at the tip of a notch (21). A surface is a favored location for rnicrocrack where ilK = increment of stress intensity, initiation. Fatigue life decreases greatly with increased c one-half crack length, surface roughness. Even tightly-held mill scale can act C,m = material constants which vary with as a surface notch. The free surface allows slip bands atmospheric and stress application to extrude or intrude under repeated stresses. These conditions, and bands act as sites for microcrack initiation and growth. N number of cycles of the applied If the cyclic stresses are small, a smooth flat crack will stress a. initiate at the surface. In the interior, the crack will steels (ASTM A 514/ A 517), remain flat but parallel lines will become visible, For martensitic structural increasing in prominence until they tenninate into the dc/dN = 66 x 10"8(1'1Kil·25 6 region of fmal fracture. In fatigue fracture areas, little deformation will be visible. The fmal fracture will have (in inches per cycle) a fibrous or granular appearance, depending on the fracture mode. For pearlitic structural steels (ASThl A 7, A 36, A 440, The first two stages of fatigue failure, crack etc.) (24), nucleation and propagation, are inter·related. A 10 schematic enlargement of the fatigue zone of the dc/dN 3.6 x 10" (1'1Kl 7 fracture profJ.le near the surface illustrates two stages (in inches per cycle) (see Figure 19). The initial fatigue crack growth (Stage I) usually lies at 45° angles to the stress axis with slight Equations 6 and 7 apply to room temperatures and orientation changes at grain boundaries. The cracking air environments. Figure 22 shows the typical change mode changes (Stage II) at some distance to a 90° in crack growth per cycle of applied load. As expected, orientation to the stress axis and exhibits striations the sum of the K values will approach the critical stress which run parallel to the crack front. Rates of growth intensity Kc and failure will eventually occur. If the Kc 8 u y; the uncracked area of the in Stage I are in the order of angstroms (lo- em) per value is large, ac > 3 cycle. Rates in Stage II crack growth are in the order member will yield plastically, and ductile fracture will of microns (lOA em) per cycle. occur. Fatigue crack nucleation entails cyclic creep which places the surface in an unstable condition. These instabilities appear as s~ipband intrusions or extrusions caused by slipband motion and( or) cross-slip which act 22 OF OF STRESS STRESS q q CRACK CRACK APPLIED APPLIED ORIENTATION ORIENTATION Propagation. Propagation. 1I 1I CRACKS CRACKS Crack Crack of of I I Stages Stages STAGE STAGE Two Two the the of of Repres~ntation Repres~ntation Schematic Schematic 19. 19. \ \ Figure Figure N N ~' ~' I LJ.. 0 w urt 0:: ~\ '24 SHEAR SHEAR 'LOCALIZED 'LOCALIZED CLOSURE CLOSURE {a) {a) MAXIMUM MAXIMUM CRACK CRACK LOAD LOAD AT AT ON ON (d) (d) CRACK CRACK SHEAR SHEAR Process. Process. COMPRESSIVE COMPRESSIVE Gmwth Gmwth REVERSE REVERSE ADVANCED ADVANCED Crack Crack H H Stage Stage ~ ~ 21. 21. SLI SLI LOCALIZED LOCALIZED Figure Figure :~,~t :~,~t (c) (c) (a) (a) MAXIMUM MAXIMUM AT AT ? ? ~;~ ~;~ STRESS STRESS CRACK CRACK TIP TIP ' ' i i OF OF INITIAL INITIAL TENSILE TENSILE BLUNTING BLUNTING 'J• 'J• 'J 'J -o ~ 0 ..J u ~ :>, u .;:; "00c u0 -= ·~- 0 e. r4 "' " .§= ~ :26 only can welding introduce cracks or crack-like fatigue conditions, Stage Not Fe-r j:.Jg._h. cycle (low stress) embrittle the weld metal and the stress risers, it can also grow1.h may be predomi.flent over most of defects 1 ... ; ;.~c.:K base metal adjacent to the weld. These unseen propagation processes. For high fatigue ·:;_F.',;:.a~ion and brittle failure and, in the presence of a predoroinate aild the can initiate ~.·u;~ses, Stage II crack groVifth will reduce the safe life of large fatigue environment, drastically period will be short. If a sufficiently result from ;:'Jdeation a structure (28). Residual stresses may nucleation period may not take .:r3.ck is present, the promote cracking and contribute to Stage II may welding. They can pla.:e, fu!d crack growth of the type in localized brittle fracture, Welding can also cause IT'0ricte failure. zone. Two welding crack growth corrosion in the heat-affected The effect of corrosion on fatigue of backing processes are subject to criticism. The use Generally) its greatest effect will lS jifflcult to assess. unfavorable grain growth water vapor bars in butt welds creates :.-''-" ::)n Stage II growth. In laboratory tests, bar is which is prone to cracking. Even if a backing to decrease the fatigue lives of steel 1-::.c~·, been shown removed, the grain structure remains found that corrosion products subsequently - ~ · However, Paris that can be criticized is the reducing the oochanged. Another process .::-:)uld lower the cyclic growth rate by a danger grindh1g of weld reinforcements. There exists of crack closure (26). The effect of corrosion especially a :i;c·J.:Junt of embrittling the surface layer of a steel, crack growth rate probably depends on ur. the fatigue (martensitic) steel. If overgrinding occurs, the frequency of load hardenable th~' action of the corrodent, concentrations may of the undesirable flexing and high stress d.pplication, and the mechanical properties occur at the weld. ~:or.rosion products. tests are Standard, Charpy V-Notch (CVN) impact t.here satisfactory for avoiding brittle steels. However, GENERAL VIEW Of THE PRESENT SITUATION as they are are several deficiencies in tough_ness tests though the phenomenon of fatigue was Even today. Standard, notched impact tests measure prior, allowance for fatigue loads applied re::ognized 35 years to form a crack and the energy 1885 by the Phoenix two energies: the energy ~,_.~·re not made for bridges until a crack is propagate a crack. If the energy to form The first standard specification for to Bridge Company. between two materials in their Railway Engineering very large, differences :,t-idges was by the America..T1 will be masked. recognized resistai1Ce to crack propagation ,·\ssociation in 1910. In 1923, a nationally CVN test are not In Toughness values provided by the design specification was issued by AASHO. ~ridge relatable to field service conditions. Experience first standard specifications for readily 1901, ASTivf issued t.~e steels exhlbiting CVN Formerly. most has shovm that low-carbon structural steel (ASTM A 7 and A 9). will usually fracture energies of 15-2{) ft-lb (20-27 J) steels used in Kentucky bridges were not st.ructural toughness. Therefore, steels with a content, with the exception of provide adequate specified by chemical based on these CVN energies, properties including transition temperature !)hosphorous, but by material corresponds to tensile tests where the transition temperature tensile and bending tests (27). Generally, safe. anticipated operating conditions, are considered from each heat of steel produced by the '<~:ere made consists of two factors: strength and and full-size tensile tests of However, toughness st.:-elmaker, includin_g half- same CVN energy at elonrration ductility. Two steels having the elastic limit 0 ' eyebars. Tensile strength ' , '-' temperature can differ in ductility. A steel and fracture surface appearance were the transition reduction-in-area, less ductility than another tests. TI1ese older having greater strength and considered relevant criteria for tensile and Irwin steel in may not withstand equivalent strains. Roberts bridges were riveted and used ductile, pearlitic that the CVN test be replaced by a dynamic trusses and built-up beams. recorrunend (29). designed, riveted structure can be better tear test A properly usually performed on the base metal theories than can welded CVN tests are d::alt with by fatigue damage not reflect the of pre-existing of a welded structure. This does structures; this is because the possibility zone a properties of the weld metal or the heat-affected and stress concentrations is low. Therefore, cracks The use of AWS standard electrodes 1 period can be expected in of the base metal. ],~ ng, fatigue nucleation metal toughness any Wilson (cf 14) does not preclude variances in weld nveLed structures. However, in 1944, the use of an ASTM steel satisfies base metal where members of riveted railway brid(J'es more than enumeraTed in the heat-affected zone is develop fatigue cr~cks. Using hls toughness. Toughness were li_l(ely to and composition of several hundred affected by heat input, electrode guidelines, bridge inspectors discovered These few base metal, joh>t design, and welding variables. cracks in railway bridges within the next the fatigue duplicated by laboratory tests. ::ears. cannot be adequately 27 Rolfe (22) has noted that the chances for brittle ACOUSTlC EMlSSlON fracture increase as An acoustic emission (AE) is a transient elastic 1. designs become more complex, wave generated by the rapid release of energy within 2. the use of thick, high-strength, welded steels a material. Consider breaking a stick by slowly bending becomes more common compared to the use it. When the wood fibers fracture, noise can be heard. of thin, low-strength, riveted plates, Most dynamic processes in metal release insufficient 3. the choice of construction practices becomes energy to be detected audibly. Therefore, sensitive more dependent on minimum cost, electronic equipment is required to detect these 4. the magnitude of loadings increases, and processes. 5. actual factors of safety decrease because of For many years, the phenomena of metals releas:ing more precise, computerized designs. audible energy has been recognized but poorly was in the The presence of alien metals in low carbon steel will understood. The first experience with this as 1tin·cry'. cause a major welding problem in the next decade deformation of tin, which became k.n.own because of the increasing scrap content in steels. They This was caused by the twinning of tin. Later, clicking during increase hardness and decrease joint toughness. sounds were emitted from large steel castings Another factor to consider is the probabilistic cooling. These sounds were caused by the formation of aspect of bridge failure in the future. Of the states martensite platelets in the steel. In the early 1920's, reporting cracking in bridges in a recent survey (13), researchers measured the energy output of a deforming the frequency of occurrence was roughly 0.06 percent metal using a piezoelectric crystal. However, defl.Ilitive of all bridges inspected. With the passing of time, tire work on acoustic emission did not occur until the late chances for subcritical growth of undetected cracks by 1940's when Kaiser used audio microphones to study stress corrosion or by fatigue increase. the AE response of metals under stress. He discovered Nondestructive evaluation (NDE) is an indirect way the irreversibility of acoustic emissions with respect to to ascertain the structural reliability of bridges. Widely load; this phenomenon was later termed the Kaiser of used methods of searching or exploring for defects effect. However, he erroneously ascribed the source include ultrasonics, magnetic-particle, dye-penetrant, acoustic emissions to be grain boundary deformation. eddy-current, and X-ray. However, X-ray has been found In the mid-1950's, AE testing gained interest in the unsuitable for detecting tight cracks. Eddy-current, UrtJted States. Schofield, using single crystals of metal, magnetic-particle, and dye-penetrant methods are found that AE signals were caused by elements of limited to surface or near-surface defects; inspection microplasticity and corrected Kaiser's hypothesis. requires direct access to the surface being tested. Schofield and Tatro were also responsible for much illtrasonics can be used to locate subsurface defects, but improvement in the electrical hardware. Dunegan first its use is restricted by the shape of a piece. None of used ultrasonic frequency testing which allowed these NDE methods is capable of directly assessing inspection in the field. Green pioneered the use of AE severity of subcritical flaws. In addition, these methods testing of rocket fuel tanks in the early 1960's. are labor intensive and require a certain degree of Nondestructive evaluation using acoustic emission has operator skill and interpretation. grovvn rapidly in the last decade, especially since it has Large cracks 2 to 6 inches (51 to 162 rrun) long been paired with computers to locate growing defects may be tolerable on some bridge members, and in large structures (33). conceivably be easily located. Due to the large size of Metals produce two types of AE activity. Where the bridge, there also exists the possibility that some the dynamic processes are easily sustained, as in the cracks may be missed by most conventional NDE plastic deformation of a ductile metal with a vary methods. Extensive tests by skilled personnel have face-centered cubic crystal structure, and do not shown that 100-percent detection and location of rapidly 'Nith time, continuous acoustic emissions are or defects is difficult to achieve using standard NTIE tests emitted, YVhere the pr9cesses involve rapid evolution as (30, 31, 32). NTIE tests in the field by semi-skilled large amounts of energy at distinct points in time, a semi·brittle metal with a operators would not be reliable. in the deformation of are produced, Acoustic ·emission testing is the one form of body-centered crystal structure, AE bursts nondestructive testing that may indicate flaw severity and enable reliable, large-scale inspection of bridges. 28 and test piece. This allows the best transmission of sonic A schematic diagram of the Dunegan Model 3000 intensity. Fluid couplants damp shear waves. Therefore, A,:Lmstic Emission Detector used in the study is shown some loss of energy &Ld a distortion of wave fOrm will ln Fi::rure 23. The essential components are transducers, occur. The couplant used in this study was a viscous ;)re:u~plifiers, and totalizers (counters). Transducers polyester resin, Dow DV -9. ~onvert small, high-frequency vibrations into electrical signals. Preamplifiers amplify the signals a..n.d filter out are si~nals with frequencies below 100 kHz, which can ACOUSTIC EMISSION TESTING u~u:tilY noise. The totalizers (counting circuitry) AT THE DMSION OF RESEARCH either. total all the counts for the duration of the test the Dunegan Model ,_ r can work in conjunction with 1 study was to develop a rates. Counting is The original goal of this .:to.:; Reset Clock to provide count of physical means of predicting the remaining life dsuallt' displayed on a panel meter and on a strip-chart. bridges subj~ct to fatigue. Acoustic em1Ss1on 11 c ~plified signal is heard through a Dunegan Model 1 phenomenon had not been fully explored from that 702 Audio Monitor. point of view. Information of this type could confirm Figure 24 shows typical amplified and conditioned fatigueupredicting programs based mainly on traffic loads AE signal packets. Each positive or negative voltage peak (36, 37). exc-eeding the threshold level is recorded as one AE Soon after acquiring an AE device, a field test was l'ount. This type of signal processing is called ring-down conducted on the Central Bridge at Nevtport, Kentucky. counting. It is one of the simplest and most widely used Transducers were attached on paired eyebars (lower measure acoustic emissions. Each of the methods to the middle of one span, yield six counts. chord of a through truss) on sig.nal packets in Figure 24 would 80 and AE monitoring was conducted usi..ng a gain of Ring-down counting gives a weighted value of event and high-pass filtering of 0.1 MHz. Several eyebar and magnitude. In the time spans shown in dB frequency for 15-minute intervals while to distinguish between pairs were monitored Figure 24, it would be difficult up traffic crossed the bridge. The AE device picked any of the signal packets. Ring-down counting does not ultrasonic activity long before vehicles reached the take the fullest advantage of the data accumulated (34). span. As the traffic approached the eyebars being Considering the source of acoustic emissions as a center the activity increased, and subsequently point, the newly generated pressure waves propagate tested, as the traffic passed off the span. Strain gages through the body as a series of expanding spherical decreased the Central Bridge showed that a traffic (load) cycle surfaces. The speed of sound in a solid such as steel on 3 lasted for a 5-second interval. However, the ts constant and equal to about 19.3 x 10 ft/sec (5.9 generally activity lasted approximately twice as long. The x 103 m/sec). AE activity was higher for heavier vehicles and There are several effects wDJch weaken the sound acoustic traffic concentrations. Later, tests revealed that pressure from an AE source as it propagates through greater impact on one link of the eyebar chain the materiaL These are scattering, true absorption, true metal-to-metal be detected by the AE device through connecting attenuation, and retransmission tluoug._l-]_ a different could eyebars for a distance of 100 feet (30 meters). The material. Scattering occurs because transmission t}l_rough ultrasonic activity was probably extraneous noise caused a body is affected by inhomogeneities. Inhomogeneities by motion between eyebars and pins. The Kaiser effect include cementite, inclusions, pores, and grain was not observed. boundaries. True absorption is loss of sonic pressure due to the conversion of mechanical energy (wave oscillations) to heat; this is damping. True attenuation is caused by spreading of the wave as shown in Figure 25. \\'hen a sound wave hits a boundary, if the surface finish is smooth, the wave will be reflected. If the surface is rough, the wave will be partially reflected and a scattered. When the wave contacts the surface of material, it changes from a body wave to a Rayleigh at wave (see Figure 26). Whereas Rayleigh waves travel of a lower velocity, they also have a lower coefficient attentuation. In a large structure, the only important AE wave form detectable would be a Rayleigh wave I 35 ). A fluid couplant is placed between the transducer 29 DRC DRC 702 702 Strip Strip Chort Chort Audic Audic """"'""' """"'""' MonitOt MonitOt I I I I 1 1 I I I I I I I I I I I I j j J J ._...... _..... System. System. D,ll D,ll dl~;~itol dl~;~itol counter counter fillers fillers com.,,., com.,,., D/A D/A 1---- j-.. j-.. 1-- converlfW converlfW DRC DRC DRC DRC TOTALIZER TOTALIZER Detection Detection TOTALIZER TOTALIZER L.. L.. 0-40db 0-40db diQitol diQitol 0-60db 0-60db '"""'"' '"""'"' amplifier amplifier amplifier amplifier ______ 310 310 ______ I I 310 310 L L ' ' I I I I L L I I l l ~-----l ~-----l Emission Emission ' ' r-----1 r-----1 I I 1 1 I I I I Acoustic Acoustic p p SPECIMEN SPECIMEN 23. 23. ORC ORC DRC DRC DRC DRC 402 402 Reset Reset Clock Clock 802P 802P 801 801 Preamp Preamp Preamp Preamp Figure Figure L.. L.. 8 8 TENSILE TENSILE Sl40 Sl40 01408 01408 or or DRC DRC ~~TRANSDUCERS ~~TRANSDUCERS J J u u I I 0 0 0 0 0 0 w w I I I I ~ ~ ~ ~ ~~ ~~ ,'j ,'j --l!ll>- TIME TIME TIME- -THRESHOLD -THRESHOLD t t transducer transducer A. A. the the sin sin ------THRESHOLD ------THRESHOLD for for e-PT e-PT peak peak Vo Vo = = V V constanf constanf Packets. Packets. voltage voltage frequency frequency f('/1 f('/1 \1 \1 IIIlA IIIlA r r ------THRESHOLD ------THRESHOLD \_(c) \_(c) \ \ {\ {\ Signal Signal II\1\IIIIIII\1\HI\111\Mv"•"" II\1\IIIIIII\1\HI\111\Mv"•"" 11 11 I I 1 1 linear linear initial initial tl tl time time ' ' damping damping II II v. v. = = II II = = " " = = II II ------ II II I I f f ------THRESHOLD ------THRESHOLD t t 11.=21Tf 11.=21Tf fJ fJ 1 1 I I Vo Vo I I Conditioned Conditioned ----1-1--1------ Aft•·q Aft•·q II II n-uvv n-uvv ;: ;: I I 2 Typical, Typical, II II 11 11 I I = = J J v v 24. 24. "~. "~. (\I (\I (b) (b) I I I' I' -1 -1 ' ' I I I I I I I I PERIOD PERIOD --~--- Figure Figure ··' ··' \fV. \fV. I\I\/\/\,I\.''PII\1\1\I\I\/\.>. I\I\/\/\,I\.''PII\1\1\I\I\/\.>. p p i i I' I' I I (d) (d) 4H 4H p p SEC SEC ' ' I I II II M M J J mr mr II. II. INTERVAL INTERVAL ---·- il il I I I I \ \ I I TRANSITION TRANSITION ' ' ' ' I I ·1-11\/\1\.\ ·1-11\/\1\.\ •l •l !--1000 !--1000 1- \1\1 \1\1 I'" I'" f f SECOND SECOND \ \ p p ---1 ---1 -~-- i i I' I' .3 .3 (a) (a) II' II' I I II. II. rv rv i i I I -- I I 1 1 \ \ I I -· -· I I I I H H 1'\11111\l'dWu· 1'\11111\l'dWu· 0 0 > > ..J ..J (f) (f) 1- + + + + > > ~ ~ 0 0 ~ ~ w w P::~ P::~ I,p I,p \ \ : : \ \ / / \ \ I I \ \ I I \ \ Waves. Waves. /s /s \ \ ' ' ' ' //"" //"" ' ' ~/ ~/ I I , , I I Spherical Spherical \ \ \ /"' /"' , , ', ', \ \ / / of of ~IA_, ~IA_, "- ...... ', ', /'\.. /'\.. t:;;l/ t:;;l/ .,.. .,.. ~ ~ ...... _... _... / / "" "" .- \ \ \ Spreading Spreading -- ', ', to to " " -~ -~ _ _ /' /' -- '~II.~ '~II.~ Due Due __ __ lA lA I I I I I I I I I I I I I I I I I I /(---+---+----;--Y /(---+---+----;--Y t: t: / / ----- Intensity Intensity _j _j --- ~--~--- '-- _ _ ><../ ><../ /-~-...... /-~-...... Wave Wave .,_.,- ...... - I I \ \ ~- '-...... '-...... { { of of .- ...... ~ ~ / / ' ' ,....- >/ >/ '..... '..... // // / / / / ' ' I I / / / / Attenuation Attenuation \ \ I I \ \ / / Y, Y, ' ' / / / / 25. 25. / / / / ', ', I I X X , , I I \ \ \ \ \ / I I I I Figure Figure \ \ \ / I I ' ' I I l l : : ( WAVE WAVE 3 3 OF OF p3 p3 I > > > > 2 p2 p2 I ORIGIN ORIGIN > > 1 1 = = pi> pi> I A A "-' "-' w w PACKET PACKET WAVE WAVE Source. Source. AE AE ~EXPANDING ~EXPANDING an an from from Propagation Propagation Wave Wave 26. 26. • • Figure Figure SOURCE SOURCE AE AE <>' <>' <>' <>' A series of tensile and compressive tests was gives a more realistic indication of activity issuing from performed on steel and aluminum to determine the the gage section. effect of various testing parameters. The tests were Figure 32 demonstrates the Kaiser effect using fully conducted at the University of Kentucky Department annealed steel. The summing interval was increased to of Metallurgical Engineering and Materials Science using 10 seconds. The specimen was loaded until a 20,000-pound (88,960-N) capacity, universal testing discontinuous yielding took place. Then, the load was machine (Instron). Pin-type grips were required for decreased to a low value. The specimen was reloaded tensile spe~imens (see Figure 27). Threaded and and pulled to failure. The AE rate demonstrated the wedge-type grips induced excessive noise in early tests irreversibility of acoustic emission. The AE activity by deforming and scraping the specimen. A special ceased on unloading, and no emissions occurred until prestressing device, shown in Figure 28, was used to the previous maximum load was exceeded. The Kaiser preload pin-holes. The preload was higher than fracture effect was also found to exist for fully elastic and fully load of the specimens. This had two beneficial effects: plastic deformation. it work-hardened the specimen pin-holes, and it Figure 33 shows the effect of relaxation. The steel 1Silenced' noise from deformation from the pin-holes. was strained to different levels, and the strain was This minimized AE activity from the grip area during maintained for time intervals of about a minute. At a the tests. given strain, the steel yielded with time; and the The initial tensile tests were conducted using AISI effective stress decreased. AE activity took place during 1018 steel (see Figure 29). The resulting load and AE the relaxation process but ceased once the steel had fwlly rate-vs-strain curve of Figure 30 shows the difference rela,'{ed. It did not reoccur until the previous maximum between the cold-worked steel (Test 9) and the annealed load was exceeded. steel (Test 10). These tests were run at a crosshead speed Figure 34 shows the load and AE rate-vs-strain for of 0.05 in./min (1.3 mm/min). Low-noise (differential) a cornmercially pure aluminum in compression. The Dunegan Dl40 transducers were used for the tests. The Kaiser effect was also demonstrated in this test. The Dunegan Model 301 totalizer was run on the 'rate specimen was strained at a crosshead rate of 0.01 memory' counting mode with a 2-second surruning in./min. A 2-second summing interval was used for the interval and a full scale of l x 105 counts. The Dunegan AE rate. No AE activity was found in compression or Model 310 totalizer was operated on the 'summing' tension until the proportional limit was exceeded. The mode with a full-scale setting of 1 x 10 6 counts. The maximum AE rate occurred prior to the onset of gross Model 301 totalizer was run with a gain of 95 dB; the plastic flow. The AE activity for most aluminum Model 310 was run with a gain of 85 dB. The AE specimens decreased gradually with inqeasing plastic outputs of both totalizers were plotted on a two-pen, flow. The compression test was stopped when the load strip-chart recorder. reached the maximum capacity of the press. AE activity of cold-worked steel rose rapidly with Ingham, et aL, have associated the AE phenomena increased tensile load and reached a maximum value at in steel with cracking of cementite; they noted that the elastic limit. After plastic deformation began, the steels with spherodized pearlite produce less total AE activity decreased. The count rate for the annealed acoustic emission than those with a lamellar pearlite specimen increased gradually with load and reached a (38). Dunegan and Harris have associated the AE rate maximum value at the elastic limit. After plastic with the mobile dislocation density (39). The ductile deformation began and as it proceeded, the rate aluminum specimen shown in Figure 34 had of material in strain as decreased. The count rate for this steel decreased after approximately the same volume pure aluminum has discontinuous yielding, occurring thereafter only in the steel specimens. Commercially and the cause of AE activity random bursts. Whereas the cold-worked steel had the no brittle second phase, attributed to dislocation motion highest AE rate, the annealed steel had nearly three in this metal must be (40, 41, 42). If steel is considered a composite of brittle times as many total AE counts. Most of the potential cementite in a ductile ferrite matrix, the AE for AE activity in the cold-worked steel had been fracture will depend on the dissipated in the cold-forming process; therefore, less contribution of the carbide AE activity due to brittle total activity was possible. percent present; and the the contribution of Figure 31 shows the effect of preloading on the carbide fracture will greatly exceed an equivalent amount of ferrite. In most structural AE rates. The AE settings and testing·machine crosshead is small. Also, the intensity speed in Test 13 were otherwise the same. Although steels, the cementite content specimen is of the there is little basic difference in shape of the rate curve, of the AE activity in the aluminum as in the steel specimens. The the AE intensity of the preloaded specimen (Test 9) same order of magnitude 34 f---= I. 50 -~-1 TYPICAL .50 .75 1 TYP. TYPICAL .75 ----- TYP - ---- 1.50 ----- TYP. 501 +.oos DIA . -.000 . ~_____,___l \ L .56 DIA. x 60° CSK ~ BOTH SIDES TYPICAL .50 DIAMETER TYPICAL 10.00 ~- I \ - -- - r-- - i- - -- - \_.125 RADIUS TYPICAL 10o:::~ 1"o2 REF TYPICAL NOTES I. ALL DIMENSIONS IN INCHES +.006J . 1875 -.000 2. ALL TOLERANCES ± .03 IN . EXCEPT WHERE NOTED. AISI- 1018 STEEL Figure 27. Tensile Specimens for AE Tests. 35 Figure 28. Prestressing Device Use:d to Prel.oad One Pin~Hole of a Tensile Specimen. 36 I ' Figure 29. Tensile/ AE Test of a Steel Specimen. 37 i i i i \ \ Annealed Annealed and and Steel Steel Cold-Worked Cold-Worked for for Curves Curves 10 10 STRAIN STRAIN TEST TEST ~I ~I AE-versus-Strain AE-versus-Strain PERCENT PERCENT and and SteeL SteeL Load- 30. 30. ® ® Figure Figure 9 9 9 9 TEST TEST TEST TEST ~ ~ II II ,_ ,_ I I I·- I_ I_ I I I I ® ® 16 16 • • <( <( g4 g4 0 0 ~ ~ z z ~J ~J ~ ~ ' 00 00 z z .,. .,. w w 3: 3: z z X X 1214 1214 (j) (j) I I 20 20 0 0 2 2 4 4 6 6 B B 10 10 12 12 16-l 16-l 14 14 18 18 @ @ 10"-. 10"-. <( <( <..) <..) 0 0 ::J ::J <..) <..) ~ ~ ~ ~ w w (f) (f) z z iii iii 0 0 X X 0::: 0::: w w !;;( !;;( <..) <..) z z 0 0 ::l ::l ...... (f) (f) 1- <..) <..) ljol ljol 20 20 ~ ~ 00 00 ® ® 20 X 18 ® TEST 13 (NO PRELOAD) 16 (/) z 14 g (,) 5: w w (/) z TEST 12 (PRELOADED) t-' 00 z v ::l ."<;!" 0 -¢ (,) 10 10 w " 0z ~ :::> 0::: 0 z a.. 0 ~ (/) (/) ~ ::?: 6 z w :::> (,) ~ t- (/) 0 ::l 4 PERCENT STRAIN Figure 31. Effect of Preloading !'in-Holes on AE Rate Curve. 39 ' ' ! ! i i ' ' 40 40 II II 38 38 36 36 If If If If IS IS Steel Steel a a for for 16 16 Effect Effect 14 14 Kaiser Kaiser the the 12 12 STRAIN STRAIN Showing Showing 10 10 Curve Curve PERCENT PERCENT 6 6 8 AE-versus-Strain AE-versus-Strain RELOADING RELOADING REGION REGION UNLOADING UNLOADING 6. 6. Specimen. Specimen. ® ® © © ® ® and and 5 5 · 4 -- ·~! ·~! Load- Tensile Tensile 5 5 4 4 4 4 32. 32. 2 2 Figure Figure I I 0 0 © © sJ sJ 8 8 5 5 9 9 7 7 2 2 3 3 4 4 lo" lo" X X • • <( <( 0 0 '3 '3 Cl Cl 5 5 <1. <1. _l _l (!) (!) (/) (/) 0 0 -t -t -t -t w w z z s: s: f2 f2 (j) (j) :z :z 10 10 ® ® 5 5 2 2 6 6 3 3 a a 4 4 9 9 7 7 10., 10., X X <( <( 0 0 0 0 0 0 w w :J :J ~ ~ (/) (/) z z i= i= 0:: 0:: rn rn 0 0 U) U) "- w w 8 8 l;;r l;;r 5 5 [;3 [;3 1-- (/) (/) 10 10 ® ® -, -, 8 1o•-1 1o•-1 IO'J IO'J 1o•--l 1o•--l 1o'--l 1o'--l IO' IO' 10 <.9 <.9 '3 '3 ~ ~ :::;; :::;; 0 0 z z ti ti ~ ~ 0 0 w w i= i= ~ ~ rn rn 0 0 z z U) U) w w ~ ~ 0 0 (/) (/) B B (/) (/) 1-- "'" "'" 0 0 20 20 -'.'~:i~c;c---- -,-,o;;p.-r:< -,-,o;;p.-r:< ~ ~ !7_2is !7_2is ''"' ''"' "•' "•' 5 5 ·-··v<·"- HOLD HOLD 1725 1725 ----'"" ----'"" 16 16 a a for for 142.5 142.5 Effect Effect 4 4 HOLD HOLD 1414.25 1414.25 Relaxation Relaxation the the 12 12 Showing Showing 10 10 STRAIN STRAIN ., ., Curve Curve PERCENT PERCENT 3 3 HOLD HOLD Specimen. Specimen. 9.5 9.5 AE-versus-Strain AE-versus-Strain B B and and Tensile Tensile ® ® ® ® 6 6 Load- Steel Steel 525 525 2 2 33. 33. HOLD HOLD ,, ,, Figure Figure I I 4 4 I I " " I I " " I I I I I I 2 2 II II II II II II II~ II~ II II II II II II I I I I I I 0 0 l l 11 11 ® ® 2 2 5 5 3 3 7 6 "-I "-I 10 10 X X " " --' --' 6 6 8 8 z z ~ ~ ::J ::J :g :g ""4J ""4J z z 0 0 ~ ~ ~ ~ z z 00 00 :;j: :;j: f2 f2 (f) (f) - 10 10 _.,. _.,. ® ® 2 2 3 3 5 5 7 7 8 8 4- 9 9 9 6 6 10 10 X X C> 0 0 (f) (f) ::J ::J Q Q w w ~ ~ 1- (f) (f) 0 0 z z iii iii 10 10 tt tt w w C> C> 0 0 ::J ::J "' "' C> C> ILl ILl z z >- lQ lQ I 36 I 34 --r UNLO/ nc!b--o SPECIMEN I 32 ;; an 'I for ~ REGION AND Effect 20 RELOADING UNLOADING Kaiser ~..0 ~ i'itfi the 1~1 1 ~ Showing 16 Curve PERCENT Specimen. ~ STRAIN, AE-versus-Strain Compression 12 and ~ Load- Aluminum 34. Figure 1 6 8 @ @ ·---..,..----,---,--~--,---,---1 4 2 @ 3 0 8 18 6 10 16 12 2 10-· X " _14 z (/) w f2 5: z 00 ' m 0 (/) z :::l 64 0 a_ _j 0 20 @ ~ 16~ tO~ 18~ 6 12 14 2 4- 10 X '-. w (/) z 0 0 :::l 1- 0: z !;;: w 0 i7i ::;:: (/) U.l 0 ~ 0 0 F 20 N -"" to were made, and further attempts cold-work and annealed determinations difference in AE rates of the were abandoned. by mobile establish load histories 1018 steels is better explained AlSl the i.rJJtial dislocation density of dislocations. The 4 2 to 10 times greater cold-worked steel was from 10 TESTING on load FATIGUE the annealed steel. Immediately, OF RESEARCH than mobile. AT THE DMSION many of the dislocations became application, loading. steel suffered dislocations upon that acoustic The annealed AE tests by others have shown steel showed a gradual Lncrease of subcritical cracks as Therefore, the annealed emission could detect growth of mobile dislocations with 6 5 Most tests were in AE rate. The decrease as 10· in. (10" mm) I 43 ). and work-hardening small continuation of plastic flow notched fra'cture specimens. Generally, the in both performed on correlates with the AE rate behavior to increase with an increase closely of the AE rate has been found more continuous AE rate behavior 46, 47) or an increase specimens. The of K (see Equation 5) 144, 45, to the greater ease of dislocation tests following aluminum {s due fracture surface area I 48 ). Proof cubic metal. of crack motion in the face-centered have shown a correlation between sh.owed that the partial fatigue Tests at faster crosshead speeds quantity of acoustic emission loading severity and the curve would increase with increased of the proof stress 149). AE rate as accumulated during application was preceptable at crosshead rates no references rates. AE activity At the time these tests were planned, mm/min). The Kaiser effect of long-term low as 0.001 in./mln (0.025 found concerning the effect as 0.7 in./mln (2.5 could be at crosshead speeds as high environment on the mechanical persisted rolling seivice in a fatigue specimens, cut in the transverse Therefore, a series of mm/min). Steel properties of low-carbon steel. normal specimens, cut in the longitudinal using four types of direction, and axial-fatigue tests were planned, tested and compared. Both types (178-mm x rolling direction, were The steel from the 7-in. x 1 1/2-in. same yield and ultimate steel. Steel of specimens exhibited the eyebars was duplicated by National specimens showed 38-mm) C&O strengths. However, the transverse-cut Table 1). The original C&O percent less Corporation (see less elongation and about 80 acid-Bessemer steel. To five percent specification was typical for total acoustic emissions. steel, National Steel Corporation specimens were cut from duplicate the eyebar After those tests, tensile steel plate after it was The eyebar had lain phosphorized an open-hearth an eyebar used in the C&O Bridge. (178-mm X 102-mm X the bridge was rolled. A 7-in. X 4-in. X 1/2-in. a storage yard for a year since 35 and 36) and a in parallel unequal angle (see Figures were cut from its stem 13-mm) tests. demolished. Specimens were also incorporated in these It was hoped that the Kaiser C9x15 channel to the loading direction. failed by fatigue in a bridge. Both of the maximum service Both of these beams effect would give an indication and were presumed to be behavior similar to that were in riveted structures stress. However, tests revealed steel. The tensile strength that strain-aging ASTM A 7 or possibly A 36 AlSI 1018 steel. It was suspected MPa), and the of angle specimen was 54.9 ksi (378 erasing the Kaiser effect. of the ksi had occurred, of the chall!lel specimen was 55.1 possibility, four specimens were tensile strength To investigate tlris designs are shown in Figures elastic region; two were (380 MPa). The specimen stress-cycled several times in the offset was found in kept at room 37-39. Later, a 0.07-in. (1.8-mm) to prevent aging and two were A) specimens frozen alignment of the C&O (Series 55 days, the specimens were loaded the gage in temperature. After error. This was compensated for stress to see if the Kaiser effect due to a drafting above the previous high were cut in the longitudinal group showed a definite later tests. The specimens occurred. One specimen of each natural surfaces were direction of the steels. All Kaiser effect. surfaces were ground. machining was preserved, and all machined At this point, the effect of surface specimens were questioned. Four previously stressed cross-sectional area. reground to slightly reduce their removed from two The machined surface layer was The specimens were specimens by etching with nital. specimen, which had restressed 44 days later. Only one effect. Therefore, not been etched, showed the Kaiser layer was- too it was concluded that the cold-worked of the specimens. Also, small to change the AE response can occur in a short it was postulated that strain-aging load history. Tests time and that it erases the previous ( 38). No other clear by others have confirmed tbis belief 43 length) CORl'. gage STEEL mm) STEEL MPa) MPa) (203 EYEIIAR in. 0 (277 (443 DUPUCATE (8 & NATIONAL ksi ksi C 64.4 0.12% 27.0% OF 40.2 0.48% 0.021% 0.117% 0.019% STEEL COMPOSITION CO. length) DUPLICATE STEEL gage AND rnm) CHEMICAL MPa) EYEBAR MPa) (127 BENNETT ANI) CORl'ORAT!ON 0 in. & (245.4 (412 (3 C ksi ksi STEEL GRAFF, 59.8 35.6 23.0% 0.55% 0.11% 0.009% 0.055% 0.136% PROPERTIES NATIONAL MATERlAL AND strength I. strength TABLE Ultimate Elongation Yield Manganese Sulphur Carbon Silicon Phosphorous ,_ '" Figure 35. Fatigue Fracture Surface of a 7-in. (172-mm) x 4-in. (102-mm) x l/2-in. (13-mm) Angle. 45 r"c"~"''"'''""~"" ?F. ::=--:··~,.,!,._...~;:.'~~~ ' Figure 36. Matching Faces of Fractured Angie. 46 10.00" ,--.50" R TYP. 3.00 MATERIAL ~ .50" THICK SCALE • I I I I 0 l/2 I" Figure 37. Eyebar Speci.men from the C&O Bridge (Series A). 47 2.50" l 10.00 11 ~~-.50" RADIUS TYPICAL r~ 2.50" MATERIAL - .50" THICK .56" _ ...... ____,.., I '--.38" SCALE I" -I.50"--- Figure 38. National Steel (Series B) and Angle (Series D) Fatigue Specimens. 48 2.00 01 8.00 01 '<------.5011 RADIUS TYPICAL .50 II ----<-t-----1-- MATERIAL "".29 11 THICK SCALE 'I I I I 0 Y2 I" Figure 39. Channel (Series C) Fatigue Specimens. 49 .A...tter the specimens were machined, the test the poi11.ts representing failures at fatigue lives greater objectives were modified to determine the endurance than two million cycles. Figure 42 shows points for limits of the angle (Series C) and channel (Series D) fatigue lives between 96,000 and two million cycles. specimens for completely reversed loading. All 12 Figure 41 reveals that even though most of the current specimens of each group were relegated for this purpose. fatigue tests lasted six times longer than values for The 11 C&O (Series A) specimens and the National Steel fatigue lives of two million cycles, their fatigue limits (Series B) specL'TI.ens were tested at various stress ratios equalled or exceeded the minimum fatigue limits for two and fractions of fatigue lives at those stress ratios. The million cycles. Figure 42 shows that fatigue limits of fatigue tests were performed by Metcut Research most specimens having fatigue lives approximating Associates of Cincinnati1 Ohio. The tests were run on 100,000 cycles, met or exceeded the minimum fatigue a Baldwin-Lima-Hamilton, IV-20 fatigue rnachin.e. Emits of typical structural steels. Most specimens with Loading alignment was achieved by drilling pin-holes, fatigue lives approaching two million cycles had lower centered along the gage length of a specimen. Fine fatigue limits than the average values given for machine aligrunent was accomplished by mounting a 100,000-cycle lives. However, these specimens had specimen (see Figure 40) and attaching strain gages. The higher fatigue limits than 'average' steels '>jVith fatigue specimen was strained and the top grip adjusted until lives of two million cycles. the gage readings were approximately equal. The Figures 43 and 44 show S-N diagrams for the specimens were tested in air at room temperatures (70 channel steel (Series C) and angle steel (Series D) ± 10 F (21 ± 5.5 C)). The specimen temperature was specimens. Both show fatigue limits of approximately limited to 200 F (93 C). The loading was applied in 15 ksi (103 MPa). a siJmsoidal manner at the rate of 1200 cpm. The total The yield and ultimate strengths of the National dynamic load error was limited to ± 3 percent of the Steel (Series B) specimens were greater than those of applied stress. the C&O (Series A) specimens. Therefore, higher fatigue The test results are shown in Table 2. The fatigue limits were expected from the National Steel specimens. limit (also the runout) was 12 x 106 cycles. On The maximum ratios of the fatigue limit to the ultimate inspection of the specimens, after machining, some strengths for the angle and channel specimens were 0.27 discontinuity was found between the fillet radii and the and 0.26, respectively. These ratios are in the lower ground faces of the gage sections. This was not range of values compiled by others ( 14). The low values corrected; it was feared that any subsequent grinding are attributable to stress risers at the gage~section fillets. would undercut the specimens at the gage root. All Tlility days after the fatigue test ended, AE tests failed specimens fractured at this location. However, the were performed on some untested specimens and the continuity of the results suggests that the stress specimens which survived the fatigue tests. Prior to each concentration factors which contributed to the failures test, the coupling efficiency between the specimen and were abnost constant for all specimens. Also, when the the transducer was measured using a Trodyne, 'Sim-Cal', specimens were inspected, they were checked for spark-gap, AE calibration device. The 'Sim~Caf flatness. A period of nine months had elapsed between duplicates an AE wave source with a signal repeatable their manufacture and use. During alignment tests at within± 20 percent. The average number of AE counts Metcut, some specimens were found to have developed recorded by the AE detector from ten 'Sim-Ca!' a slight bow. This was probably caused by long-term excitations was used in a simple ratio between the lowest relaxation of residual stresses in the specimens and led average (as a reference) and other tests to standardize to problems in testing which required completely data. The specimen pin-holes were preloaded, and the reversed loading and stresses greater than 30 ksi (207 tests were performed at a crosshead speed of 0.05 MPa). Plastic flow in these specimens led to buckling ~"!./min (1.3 mm/min). The gain was set at 95 dB with and caused three, Series D tests to be aborted. Within high-pass filtration of 0.1 MHz. The rate memory the limits of lo-ading accuracy specified and other counting mode was used with a 1 x 105 scale and a factors, fatigue life variations of 30 to 95 percent were summing interval of 10 seconds. Thirteen specimens possible (50). were tested to failure. Three were loaded to their yield Figures 41 and 42 show modified Goodman points, removed, notched, and tested to failure. The diagrams for the fatigue tests ( 14). The average lines results of these tests are shown in Table 3. indicate average fatigue limits of low-carbon structural steel for fatigue lives of 100,00 a.Dd 200,000 cycles. The shaded areas show stress variations between the average and minimum fatigue limits. Imposed on Figure 41 are 50 Figure 40. Strain Gage Alignment of Specimen. 51 TABLE 2. RESULTS OF AXIAL FATIGUE TESTS CYCLES TO SPEC. MAX!Millll STRESS MINIMUM STRESS"" CYCLE CYCLE FAILURE NO. (bi) (~!Pa) (ksi) (MPa) DESIGNATION REQUIRED (THOUSANDS) RESULTS AI 19 13l.O ·19 -131.0 NAl To failure 12,000 failure A2 29 199.9 -19 -199.9 NA2 75% of NA5 86 failure AJ 29 199.9 -29 -199,9 NAJ 50% of NA5 600 Removed A4 29 199.9 -29 -!99.9 NA4 To failure 6ll failure AS 29 199.9 ·29 -199.9 NA5 To failure 1,197 failure A6 35 241.3 0 0 NA6 To failure 204 Failure A7 35 241.3 0 0 NA7 75% of NA6 153 Removed A8 55 241.3 0 0 NA8 50% of NA6 102 Removed A9 54 372.3 0 0 NA9 To failure 96 Failure AlO so 344.7 25 172.4 NAJO To failure 12,154 Runout All 56 386.1 28 193.1 NAil To failure 526 Failure 81 29 199.9 ·29 -199.9 NB1 To failure 181 Failure B2 29 199.9 ·29 -199.9 NB2 75% of NB! 136 Removed BJ 35 241.3 0 0 NBJ To failure 12,057 Runout 84 35 241.3 0 0 NB4 75% of NBJ 3,985 Failure B5 35 241.3 0 0 NBS 50% of NBJ 6,000 Removed Cl 25 171.4 -25 -172.4 NC1 To failure !08 Failure C2 25 172.4 -25 -172.4 NC2 To failure 48 Failure CJ 30 206.8 ·30 -206.8 NCJ To failure 19 Failure C4 30 206.8 ·30 -206.8 NC4 To failure 20 Failure C5 20 l37 .9 -20 -137.9 NCS To fai!me 672 Failure C6 20 137.9 ·20 -137.9 NC6 To failure 1,627 Failure (7 !0 68.9 ·!0 -68.9 NC7 To failure 12,000 Runout C8 15 103.4 ·15 -103.4 NCB To failure 12,143 Runout C9 19 131.0 ·19 -13 J_Q NC9 To failure l ,550 Failure C!O 18 124.1 ·18 -124.1 NC!O To failure 1,762 Failure Cll 16 !10.3 -16 .JJQ,J NC11 To failure 6,998 Failure Cll J6 110.3 ·16 -110.3 NC12 To failure 4,712 Failure Dl 45 3\0.3 45 -310.3 ND! To t·ailure Te~t aborted 01 40 275.8 ·40 -275.8 ND2 To fa.ilure Te~l aborted DJ 30 206.8 -30 -206.8 NOJ To failure 64 Failure 04 30 206.8 ·30 -206.8 N04 To failure 99 Failure 05 20 137.9 -20 ·137.9 ND5 To failure 835 Failure 06 20 137.9 -20 ·137.9 ND6 To failure 1,169 Failure 07 !0 68.9 ·10 ·68.9 ND7 To failure 15,395 Runout 08 19 131.0 ·19 ·131.0 N08 To failure ],832 Failure 09 15 103.4 -15 ·103.4 ND9 To failure 12,000 Runout 010 18 124.1 ·!8 ·124.1 NOlO To t"ailure 12,000 Runout Oil 19 131.0 ·!9 ·13!.0 NOll To failure 474 Failure 012 35 241.3 ·35 ·24\.3 N0\2 To failure Test aborted denotes compression 52 60 _j __ I _L __ Y 40 Bridge of 2,000,000 I/ I than Limits " " " CYCLES Greater KSI Fatigue 20 " " " the Lives " U II 2 M II II " > ~<<:<'<'.a!' " N " " Fatigue STRESS, ...... Comparing for I 0 UNIT " " " Values, Diagram SPECIMENS 'A' 'c' 'o' :>K'<:<:.>"' Average Goodman o • 1::::. O's' "(-' MININUM 14). ( with "''Il: -20 Modified Steels Cycles I 4!. Figure I ~4~0~-L--~~~--~~--L---~--~--JL 60r---~----~--~----,----,-----r---,~---r----~--_, 20 40 en w 0::: :::.:: en en 1- en :z ::) l X ::::!: ~ ~ w ~ r 6 'A' SERIES SPECIMENS 96,000 :::; N < 2,000,000 CYCLES 0 '8' " " " II II U " " 'c' " " " II If II " " II II II •D 'o' " " " " " 60 I: I I' (f) :.:: (f)• (f) w 0:: 1- (f) 1- z I,' :::J • ! ::2: I !" :::J 20 :.. ::2: ·.i.,::·'• 1 X -20 0 + 20 +40 +60 MINIMUM UNIT STRESS, KSI I • I i.l Figure 42. Modified Goodman Diagram Comparing Fatigue Limits for Bridge Steel j i l :•1 with Average Values, for Fatigue Lives Less than 2,000,000 Cycles and '[I . '(,,: Greater than 96,000 Cycles . 11·: !,, 1···'I•. '.'.' I. 1·11ii .,,I'. '·'.I ·:! 1. I. 54 !''.. ' 1I. 6 \7 C-8 .------J0 C-7 4 \J __ C-ll C-12 Specimens. C) ----'\7-.s;; c~o (Series C-6 /Y (lOOO's) ------IOL 3 Channel ---C-94. for :;z. C-5 FAILURE TO -- Diagram S-N C'ICLES --- 43. ------IL0~ 2 v ------ n C-1 Figure -...... _ \7 \d'- ~-...._..._ C-3 (KSI) OLIO~------IL0° 10 30 40 '20 I I (MPA) 0 100 300 2ooL w ~ (j) ~ ~ (j) ifj X ~ ;:) u, u. d d \7 \7 D-7 D-7 \} \} D-9 D-9 D-10 D-10 -~- ~ ~ '"'"="'~-· '"'"="'~-· - Specimens. Specimens. D) D) """"' """"' (lOOO's) (lOOO's) (Series (Series \} \} D-8 D-8 -=·~" -=·~" Angle Angle .. .. \} \} 0"6 0"6 ~ ~ ------v· ------v· FAILURE FAILURE for for \} \} D-5 D-5 ~~"'- TO TO --- Diagram Diagram \} \} 1!1!1!1! 1!1!1!1! S-N S-N CYCLES CYCLES -....__Dell -....__Dell ~-,;7" ~-,;7" 44. 44. i11111!11 i11111!11 -...... _ -...... _ Figure Figure '-...... _ '-...... _ d d 0-4 0-4 '~, '~, UL",.RII. UL",.RII. \7 \7 0"3 0"3 -:-:'-="~~,:~~---'----~:·-.;;::- ---- 111111111111. 111111111111. ---~-----· ---~-----· ~ ~ OL------~~------~.------J~------~ OL------~~------~.------J~------~ 10 10 20 20 30 30 (KSil (KSil 40 40 I I I I I I 0 0 "· "· """~:-,~-- 100 100 (MPA) (MPA) 200 200 300.- ::;;: ::;;: <( <( - X X ::::> ::::> ::;;: ::;;: ::;;: ::;;: (f) (f) 1- w w 0::: 0::: ({) ({) ({) ({) "' "' u. u. ~~·.:...:,_·~:::___ ~~·.:...:,_·~:::___ /mJ) /mJ) 5 IU B75.0J B75.0J fqs.so fqs.so ~44JJJ ~44JJJ J8Y.19 J8Y.19 3<11.06 3<11.06 93J.n 93J.n l,l4.26 l,l4.26 "175.1J "175.1J 8J7.1•1 8J7.1•1 .15•1.'16 .15•1.'16 Jl85.11! Jl85.11! .109.1 ,6(11_9] ,6(11_9] ,16J.Bio ,16J.Bio I I I I I I I,IM.'II I,IM.'II I I ' ' l,tiJY.n l,tiJY.n (1 (1 TUUG!JNESS TUUG!JNESS 18 18 l~ l~ .. .. .. .. 8.~fJ 8.~fJ 12.38 12.38 ~~-D ~~-D )~ )~ 2fl.BS 2fl.BS 1"1.J~ 1"1.J~ 35.75 35.75 l?.UU l?.UU 58.13 58.13 4(1.25 4(1.25 2U.I3 2U.I3 30.63 30.63 Jl) Jl) 31.75 31.75 •ll.H •ll.H •12.25 •12.25 (HL-klp/onh (HL-klp/onh li li COUNTS COUNTS 57.5 57.5 555 555 n:t n:t 6(),9 6(),9 l.li.S l.li.S 322.5 322.5 1~1-J 1~1-J 290_(, 290_(, Bll.'l Bll.'l J~].~ J~].~ l4CLS l4CLS 53(,,() 53(,,() '117.3 '117.3 COI 1,913.4 1,913.4 I.UJ~ I.UJ~ ),455.(, ),455.(, ACOUSTIC ACOUSTIC EMISSION EMISSION TOTAL TOTAL l l 6 6 .. .. 1,1 1,1 '' '' _\_!) _\_!) r,_r, r,_r, h.2 h.2 9,5 9,5 ",, ",, )(,,0 )(,,0 '"' '"' 11..1 11..1 1<1.0 1<1.0 lll lll iS.5 iS.5 IJ.:i IJ.:i II II !Cfi !Cfi COIINTS COIINTS 1'1\0ilYNE 1'1\0ilYNE (AVJ'RACf) (AVJ'RACf) ) ) 4 to- 1~9.1 1~9.1 6~'11J 6~'11J OHUh OHUh ~ ~ 1_177•1 1_177•1 5.1~](, 5.1~](, 5.061'1 5.061'1 5.2251 5.2251 5 5 :!.8'1'1-l :!.8'1'1-l J.~KM J.~KM 5.1703 5.1703 S,l7,l0 S,l7,l0 j_]~J'I j_]~J'I S S IF2J IF2J J.OHIJ J.OHIJ 5_2)"14 5_2)"14 5.!15·18 5.!15·18 ~ ~ "(,:'!6 "(,:'!6 IN IN ---~:" ---~:" VOLUME VOLUME 80~~ 80~~ 8\l 78<11> 78<11> (i.,_c, (i.,_c, ~IWI ~IWI HIJI·I HIJI·I 8025 8025 ~~-,, ~~-,, -11(,5 -11(,5 [117'1 [117'1 '/2fiH '/2fiH 7'!7.1 7'!7.1 44'11 44'11 .~II~ .~II~ 7K.l:' 7K.l:' .]"IJ(, .]"IJ(, .BIJS~ .BIJS~ SPECIMENS SPECIMENS _lio _lio 02 02 ·18.7" ·18.7" 1(>1.11 1(>1.11 CoO.IIU CoO.IIU 2i>~_IH 2i>~_IH 2BUA5 2BUA5 1 16852 16852 .164.!8 .164.!8 267.18 267.18 !OU.IOI !OU.IOI cr,r,_ll cr,r,_ll Jr,4_J-I Jr,4_J-I "51l.NJ "51l.NJ - -'2H -'2H 2-II.'IH 2-II.'IH .'lHII .'lHII 251 251 IMI'al IMI'al SII!FS~ SII!FS~ 'IS 'IS ;a ;a __ __ l"l l"l ][,_.](, ][,_.](, ,,,_1_1 ,,,_1_1 HH1 HH1 <~•'''' <~•'''' JH JH _,a_ 5:!.84 5:!.84 "--'~ "--'~ 1S.IIJ 1S.IIJ [biJ [biJ 'lfd[• 'lfd[• _,,,,()~ _,,,,()~ Hf•O Hf•O ;_115 ;_115 ~0_(,8 ~0_(,8 _\~.71 _\~.71 ..--1;:.;_.:\'.ILUl ..--1;:.;_.:\'.ILUl H H \5 \5 I I I'> I'> ll ll 2 2 '' '' .~ .~ 1''·'-~' 1''·'-~' l.l·l,l)_l l.l·l,l)_l LC~.-11• LC~.-11• l'I_ ~II'" ~II'" 1'117•1 1'117•1 ~7.1, ~7.1, lK.iJH lK.iJH 411.1111 411.1111 _1%,11'1 _1%,11'1 .1\].]_jl] .1\].]_jl] .j\IJ_.j() .j\IJ_.j() -123_1)~ -123_1)~ -1 -1 ·I~ ·I~ J'•~--P J'•~--P -l-1'1,7!1 -l-1'1,7!1 I I Sl"l FATIGUE-TESTED FATIGUE-TESTED 'IH 'IH ~I) ~I) \'1.1,\ \'1.1,\ -'~XI -'~XI 5HI>'I 5HI>'I I•Mn'l I•Mn'l [k,] [k,] l),ll l),ll 11 11 ,,,_.]\ ,,,_.]\ '''''h '''''h i(o,lil i(o,lil \) \) O.'K;i O.'K;i h22-1 h22-1 l•-1.17 l•-1.17 1>1,_1) 1>1,_1) ••!I< ••!I< ,,,.,_, ,,,.,_, l'Ul~IATL l'Ul~IATL ';'II ';'II ' ' ! ! -, -, " " !I !I .. .. "' "' -~ -~ IIH IIH '"' '"' Cl Cl 'II• 'II• 12 12 2•1) 2•1) CJ.,: CJ.,: ~~ ~~ 2111 2111 C_l.ll C_l.ll CCI> CCI> 2 2 ~I ~I _1].1 _1].1 C-l" C-l" ]i'I;IU'I ]i'I;IU'I I.IJIC.I,,\]Ill"> I.IJIC.I,,\]Ill"> PREVIOUSLY PREVIOUSLY OF OF 'T1 'T1 -1 -1 2 2 -1 -1 -1 -1 -1 -1 'Jill' 'Jill' I I AIH·-1 AIH·-1 "~ "~ ,,,, ,,,, \) \) )11)2 )11)2 \() \() _1-l_h _1-l_h '"'' '"'' 21,11 21,11 ;•>-1 ;•>-1 "' "' 'Ii-I 'Ii-I bU bU '1\ '1\ f,~ f,~ hill hill ld ld I'· I'· 11'1·1\fl 11'1·1\fl 1\I·IJL 1\I·IJL TESTS TESTS 5 5 ~ ~ .\ .\ ·- ,,,, ,,,, 1112 1112 011(1 011(1 I I i>IJH i>IJH IJIJIJ IJIJIJ 'S,\'-IlSI 'S,\'-IlSI 0.~(111 0.~(111 ,,_nuu ,,_nuu <~1111·1\ <~1111·1\ I'YCU I'YCU I.! I.! 1211_\(1 1211_\(1 I I 1211011 1211011 12.1-1!1 12.1-1!1 12,(11111 12,(11111 1!1<11 1!1<11 ';I ';I «>I''"''"'" «>I''"''"'" OF OF l.riiOI l.riiOI TENSILE TENSILE -----·· -----·· I I 1.!~ 1.!~ ) ) i ~OU ~OU 2-11 2-11 CriiJ CriiJ 0 - IJ IJ (J (J -1• '" '" '" '" ,,_1[)_1 ,,_1[)_1 "' "' '"I) '"I) '"I" '"I" '" '" ,.,() ,.,() I" I" '"]!I '"]!I FROM FROM llo-C•'I llo-C•'I IU IU J\ll',o) J\ll',o) I I "~ "~ ((j_l ((j_l 11),1 11),1 C-11 C-11 2-11 2-11 Cl1111" Cl1111" lo'l lo'l 1!1 1!1 2~ 2~ [,0, [,0, _1Hc, _1Hc, 2-1] 2-1] 211!! 211!! :-II'" :-II'" LOlll LOlll k k DATA DATA '" '" I~ I~ Ill Ill IJ IJ ·"' ·"' ]'! ]'! ,., ,., ·'" ·'" l'lllll.l l'lllll.l ·I-' ·I-' ·Ill ·Ill n n '" '" '" '" '" '" lu lu I<' I<' '" '" L"ll L"ll '" '" '"II '"II ...... '" '" '"II '"II '" '" (l.,i) (l.,i) II II I~ I~ ]j ]j l'i l'i IU IU Ill'" Ill'" -'-' -'-' ,, ,, ,., ,., ;r, ;r, '" '" 1) 1) 1) 1) 2'1 2'1 3. 3. .. .. --·- -;,,,l> 8 ""Doma~ool ""Doma~ool Ill( Ill( ILl ILl II.; II.; ])Ill" ])Ill" r~· r~· 06 06 1!12 1!12 liS liS "' "' ,, ,, 1\IC' 1\IC' "' "' o\.1 o\.1 '" '" " " ". ". NUMUI·I< NUMUI·I< TABLE TABLE .'ii'I:C"JM!S .'ii'I:C"JM!S "' "' ._, ._, The tests revealed that the speci..mens had a brittle manner once the fatigue crack had reached the strain~aged. The AE rate curv·es resembled those of critical size. The unnotched tensile specimens had a earlier tests, and the Kaiser effect was not evident. There typical fibrous cup-and-cone appearance. Two of the was no discernible effect on tbe mechanical properties notched spechnens (A7 and C8) also had fibrous, of the specimens subjected to low cyclic stresses. A C&O fracture surfaces. However, the angle specimen (D 10) specimen (A 11), fatigued in the range from 56 ksi (386 showed a coarse, granular, fracture surface. MPa) to 28 ksi (193 MPa) for 12 million cycles, was A scanning electron microscope was used to not substantially different in ductility and toughness observe the fracture topographies at higher levels of from specimens whlch had not been fatigued. magnification. Figure 45 exhibits a region of It became apparent that no gross changes in predominently ductile fracture. Figure 46 shows a region mechanical behavior resulted from cyclic loading below of predominently brittle fracture. Figure 47 reveals a the yield region, and this indicates that local plasticity fatigue fracture with crack growth striations which associated with cracks was the main problem of fatigue, indicate a crack growth rate of 8 x 10-5 in. (2 even in older bridges. One C&O spechnen (A3) produced rnicrons)/cycle. This type of surface was the exception an unusually high AE count; most of the activity for the fatigue fractures. Most of the fatigue surfaces occurred in the discontinuous yield region. However, the had an irregular appearance as shown in Figure 48. mechanical properties measured during this test did not Observation of National Steel specimen B4 confirmed markedly differ from values of other specimens. This that it failed in a brittle manner (see Figure 49). The event represents some presently undetermined variance fracture surface of the notched, angle specimen (DlO) in properties not measurable by tensile testing. was very similar to that of the National Steel specimen Three specimens (A7, C8, and DlO) were notched (see Figure 50). with a jeweler1s hacksaw after the initial loailing. After The lack of detectable fatigue striations in most notching, the specimens were tensile tested to failure. of the fatigue surfaces is due to the high inclusiOn During the tests, the notches were observed to became content of these steels. Phosphorizing of the National bluntedJ and areas of localized plastic deformation Steel spechnen probably contributed to its brittle mode appeared on the faces of the specimens adjacent to the of fracture. The National Steel specimens had higher notches. Each specimen produced an abbreviated yield stresses and ultimate strengths than the other stress~strain curve similar to those of unnotched steels. This, combined with the small area of fatigue specimens. Compared to the smooth specimens, the crack growth, indicates that the crack nucleation process notched spechnens showed higher yield and ulthnate was predominant during the fatigue life of the specimen. strengths but less toughness and elongation. The total Due to low toughness, very little crack growth was AE count was greater for urmotched specimens. The .A.E accommodated prior to fracture. The topography of the activity began hnmediately with the load application. angle spechnen (DlO) was similar to the National Steel It reached a maximum value during discontinuous specimen primarily because of the constraining effect of its circumferential notch. It should yielding and decreased with the onset of fully plastic be noted that the other specimens were notched on only two flow. A slight increase in AE rate occurred prior to opposite faces. failure. The increase in strength and loss of ductility are results of the notch effect. The high initial rate of As a result of these tests, it became evident that AE activity was caused by the rapid onset of localized neither laboratory fatigue tests nor AE tests on coupons plastic flow. The increase of AE rate prior to fracture or samples from bridge members would provide insight was probably due to rapid growth of the plastic zone. into the fatigue behavior of an existing bridge. Some The fracture surfaces of the fatigued specimens investitmtion was made of the use of AE flaw~locating (excluding those cycled into compression) were visually equipment for field use. AE systems have been inspected. The C&O spechnens had a smooth, dull interfaced with computers to locate defect~ in pressure appearance in the zone of fatigue crack growth. A region vessels for the past 12 years. Much field experience has with a dull, fibrous appearance, typical of ductile been accumulated from oil storage tanks, rocket motor fracture, was also present in each of these specimens. cases, and nuclear reactors (51, 52, 53). Several These specimens failed in a ductile manner due to loss techniques have been developed for AE discrimination of section caUsed by fatigue crack growih. A National in a high noise environment (54, 55). However, only Steel specimen (B4) showed a small region with a dull, one significant trial of AE flaw-locating equipment has smooth appearance at the comer of the specimen. The been made on a small bridge (56). rest of the fracture surface had a coarse, granular appearance. This indicated that the spechnen failed in 58 Figure 45. "]"J.ctBe ?2:'2:.c ;u;e ·:f Angle Steel (D7). 59 1,::, I II ,, i I,i!ij,, I Figure 46. Region of Predominently Brittle Fracture of Angle Steel (DlO). 60 Figure 47. Fatigue Crack Striations in C&O Eyebar Steel (A6). 61 I i Figure 48. Fatigue Fracture without Apparent Striations, C&O Eyebar Steel (A6). Figure 49. Transition from Fatigue Crack to Brittle Fracture in National Steel Specimen (B4). 63 Figure SO. Transition from Ductile to Brittle Fracture for Notched Angle Specimen (DJO). 64 most bridge testing interval can be determined. Since further explore the listening and probing which To have been short-life, brittle fractures a large, welded plate girder, failures capabilities of equipment, more frequent tests would be a web length cannot be predicted, approximately 50 feet (16.4 m) long, with 3/4 inch (19 required for new bridges. 27 inches (690 mm) and a thickness of dose, of 1 1 critical bridge members warrant AE device and the Sim-Cal Only mrn) was tested using the evaluation. A large number of 1 produced a repeatable, periodic, nondestructive calibrator. The Sim-Caf of a bridge can be eliminated from of lower magnitude than an structural members simulated, acoustic pulse by applying the following Using a gain inspection and monitoring AE burst in a notched, tensile specimen. 0.1 to 0.3 MHz considerations: of 97 dB, with band-pass filtration of bridge is one whose the 1. a critical member of a Dunegan S-140B transducer, or and a single-ended, failure would lead to the sudden collapse could readily be detected from 45 feet 'Sim-Cal' signal impairment of a bridge; A decrease in AE level permanent (14.8 m) by the AE device. designed members are those between the 2. the most critically occurred with increasing distance the largest loading range and However, the test subjected to transducer and the AE calibrator. those with a tensile blunting and highest stresses, especially showed that AE energy caused by crack over long loading component; and crack growth can be detected by AE systems flaw-locating areas of a structural member are distances. Recent improvements in AE 3. most critical complex of geometric discontinuity systems have enabled their use on geometrically points of electronics cover plates, etc.) or points of structures and reduced the amount (connections, defects over long distances. greatest strain on loading. required to locate tests and Good judgment will eliminate unnecessary be tested per increase the number of bridges which ca..n PROPOSED year. ACOUSTIC EMISSION TESTING zone CONCLUSIONS As a crack propagates, a plastic damage and generates acoustic emission. preceeds the crack tip evaluation is the the plastic zone will grow, 1. Comprehensive, nondestructive As the crack size increases, failures which is not only method capable of preventi..11g catastrophic producing more AE activity. A flaw emission. growing, will not continue to produce acoustic of bridges. emission can be used to locate growing in a bridge member to grow and to generate 2. Acoustic For a crack member. bridge must be stressed to a level cracks in a large structural acoustic emission, the to be necessary greater than the 3. Acoustic emission testing appears less than the yield stress but perhaps of have to be make comprehensive, nondestructive evaluation maximum service stress. A bridge might to member at a practical. proof-loaded, one span at a time or one large bridges time to achieve the desired conditions. by The frequency of testing can be determined a statistical various methods. Payne (57} developed with both model for determining risk factors involved 1 in a fatigue 'safe-life' and 'fail-safe aircraft structures 1 1 to bridges environment. Safe-life structures correspond members and 'fail-safe' structures having no redundant 1 Payne S correspond to bridges having back-up members. to determine model correlated risk and testing frequency developed a an acceptable safety level. Rolfe (22} also in which method of determining inspection frequency 1 1 not be detected it was assumed that a flaw size C would stress and in an initial inspection. Considering working where K(' a plain-strain situation (to be conservative) can be determined = Klc• the critical size of a defect be used for from Equation 6. Equations 6 and 7 can program to the applicable steel; employing a computer Krc• using determine the interval required to achieve traffic studies: the load frequencies determined from 65 18. Osgood, C. C.; Fatigue Desigu; Wiley, 1970; pp. REFERENCES 49-53. 19. Grosskreutz, J. C.; ,Fatigue lvfechanisms in the 1. Parker, E. R.; Brittle Behavior of Engineering Sub-Creep Range; Metal Fatigue Structures; Wiley, 1957; pp. 259-261. Damage-Mechanism, Detection, Avoidance and M., Editor; Brittle Fracture in Steel 2. Boyd, G. Repair; STP 495; American Society for Testing and Butterworths, 1970; pp. 19-21. Structures; Materials, 1971; pp. 5-60. Report· Collapse of the US 35 3. Highway Accident 20. Feltner, C. E. and Beardmore, P.; Strengthening 15, Bn'dge, Point Pleasant, West Virginia, December Mechanisms in Fatigue; Achievement of High Transportation Safety Board, 1967; National Fatigue Resistance in Metals and Alloys; STP 467; Report No. NTSB-HAR-71-1. American Society for Testing and Materials, ·1970; 4. Dieter, G. E.; Mechanical MetallUigy; McGraw-Hill, pp. 77-112. 1961; pp. 190-220, 298. 21. Tetelman, A. S. and McEvily, J.; Fracture of 5. U. S. Steel Corp.; The Making, Shaping, and Structural Materials; Wiley, 1967; p 382. Treating of Steel; Herb rick & Held, 1971; pp. 22. Laird, C.; The Influence of Metallurgical Structure 1089-1091. on jl;fechanisms of Fatigue Crack Propagatzon; 6. Liebowitz, H.; Effects of Alloying on Fracture Fatigue Crack Propagation; STP 415; American Characteristics; Fracture: An Advanced Treatise; Society for Testing and Materials, 1967; pp. Academic Press, 1969; Vol 4, pp. 2-77, 84-133. 131-168. 7. Bruner, R. J.; Fatigue Analysis of Central Bridge 23. Paris P. C. and Erdogen, F.; A Critical Analysis Eyebars Deduced from Stmin Gage Data and of Crack Propagation Laws; Basic Journal of Probability Analysis; Division of Research, Engineering, American Society for Testing and Kentucky Department of Highways, 1973. Materials, Vol 185, 1963, pp. 528, 183. 8. Uhlig, H. H.; Corrosion and Corrosion Control; 24. Rolfe, S. T.; Use of Fracture Mechanics in Design; Wiley, 1971; pp. 120-14'2. International Metallurgical Reviews, ASM, Metals 9. Biggs, W. D.; Fracture: Physical Metallurgy; Calm, Park, Ohio, September 1974; pp. 183-199. . J. W., Editor; North Holland/American Elsevier, 25. Achter, M. R.; Effect of Environment on Fatzgue 1970; pp. 1222-1223. Cracks; Fatigue Crack Propagation, STP 415; 10. Conlangelo, V. J. and Heiser, F. A.; Analysis of American Society for Testing and Materials, 1967; Metallurgical Failures; Wiley, 197 4; pp. 11 7 · 119. pp. 181-204. 11. Szczepanski, M.; The Brittleness of Steels; Wiley, 26. Paris, P. C., et a!.; Extensive Study of Low Fatigue p 5, 24. 196"3; Crack Growth Rates in A jJJ and A jQ8 Steels; 12. May, M. J.; Applications of Fracture Mechanics; Proceedin•s on the 1971 National Symposium on Fracture Toughness; Iron and Steel Institute, pp. = . Fracture Mechanics, Part !; STP 513; Amencan 89-104. Society for Testing and Materials, 1972; pp. T.; Mechanics · 2; An 13. Barnby, J. Fracture 141-176. introduction to fracture mechanics; Nondestructive 27. Burr, W. H.; The River Spans of the Cincinnati and !PC Press, Guilford, UK, December 1971; Testing; Covington Elevated Railway, Transfe-r and Bridge pp. 385-390. Company; Transactions, American Society of Civil 14. Bamby, J. T.; Fracture j}'fechanics 4 3, Toughness Engineers, August 1890; pp. 65-95. and critical defect size; Nondestructive Testing; IPC 28. Welded Highway Bridges, An Analysis of Inspection Press, Guilford, UK, February 1972; pp. 32-37. Factors; Special Report 114, Highway Research 15. Carter, C. S., et al.; Stress Corrosion Susceptibility Board, 1970; pp. 52-66. of Highway Bridge Construction Steels · Phase I; 29. Roberts, R. and Irwin, G. R.; Fracture Toughness FHWA Report No. FHWA-RD-73-3, 1972. of Bridge Steels· Phase II Report; Federal Highway 16. Munse, W. H.; Fatigue of Welded Steel Structures; Administration, 1974; Report No. Welding Research Council, 1964; pp. 44-59, 74, FHWA-RD-74-59. 136. 30. Whittaker, V. N.; A Review of Nondestructive 17. Drew, P.; Inspection of Steel Bridges for Fatigue Measurement of Flaw Size; Nondestructive Testing; Damage; Journal of Structural Division, American IPC Press, Guilford, UK, April 1972; pp. 92-100. Society of Civil Engineers, August 1971; pp. 2107-2118. 66 31. Nondestructive Methods of Fatigue Crack 44. Willertz, L. E. and Hunter, D. 0.; An Detection in Steel Bridge Members; Research Acoustical-Emission-Dampening Fatigue Study of Results Digest, January 1975. 17-4 PH Steel; Westinghouse Research 32. Rummel, W. D., et al.; The Detection of Fatigue Laboratories, Scientific Paper No. Cracks by Nondestructive Test lVfethods,· Materials 72-1D9-CRAKD-Pl, December 1972. Evaluation, ASNT, Columbus, 0, October 1974; 45. Morton, T. M. et al.; Effects of Loading Variables pp. 205-212. on Fatigue-Crack Growth; Experimental Mechanics, 33. Schofield, B. H.; Research on the Sources and SESA, Westport, Conn, May 1974; pp. 208-213. Characteristics of Acoustic Emission; Acoustic 46. Hartbower, C. E., et al.; Acoustic Emission from Emission, STP 505; Americfu.l Society for Testing Low Cycle High-Stress-Intensity Fatigue; and Materials, 1972; pp. 11-19. Engineering Fracture Mechanics, Pergamon Press, 34. Brindley, B. J. et al.,; Acoustic Emission- 3, The NY, May 1973; pp. 765-789. Use of Ring-Down Counting; Nondestructive 47. Morton, T. M., et al.; Acoustic Emissions of Fatigue Testing, !PC Press, Guilford, UK, December 1973; Crack Growth; Engineering Fracture Mechanics, pp. 299-306. Pergamon Press, NY, May 1973; pp. 691-697. 35. Bell, R. L.; Acoustic Emission Transducer 48. Mirable, M.; Acoustic Emission Energy and Calibration Transient Pulse Method; Mechanisms of Plastic Deformation and Fracture; Dunegan/Endevco Corporation, 1973; Report No. Nondestructive Testing, !PC Press, Guilford, UK, DE-73-3. April 1975; pp. 77-85. 36. Havens, J. H. and Deen, R. C.; Bridges: Synthesis 49. Harris, D. 0., et al.;Prediction of Fracture Lifetime of Load Histories and Analysis of Fatigue; Division by Combined Fracture Mechanics and Acoustic of Research, Kentucky Department of Highways, Emission Techniques; Dunegan Research January 1972. Corporation, Report No. DRC 105. 50. Yokobori, T.; Strength, Fracture and Fatigue of 37. Lynch, R. L.; Analysis of Traffic Loads on Bridges; Materials; Noordhoff/Groninger/Netherland, 1965; Report If, Characteristics of Traffic on Ohio River pp. 20. Bridges · 1968; Division of Research, Kentucky 51. Tatro, C. A.; Design Criteria for Acoustic Emission Department of Highways, March 1969. Experimentation; Acoustic Emis~ion, STP 505; 38. Ingham, T., et al.; Acoustic Emission American Society for Testing and Materials, 1972; Characterization of Steels, Part 1: Acoustic pp. 84-99. Emission _Measurements from Tensile Tests; 52. Cross, N. 0., et al.; Acoustic Emtssion Testi'ng of International Journal of Pressure Vessels and Pressure Vessels for Petroleum Refineries and Piping, Applied Science Publishers, UK, February Chemical Plants; Acoustic Emission, STP 505; 1974; pp. 31-49. American Society for Testing and Materials, 1972; 39. Dunegan, H. L. and Harris, D. 0.; Acoustic pp. 270-296. Emission - A New Nondestructive Testing Tool; 53. Harris, D. 0. and Dunegan, H. L.; Acoustic Ultrasonics, !PC Press, Guilford, UK, July 1969; Emission-S, Application of Acoustic Emission to pp. 160-166. Industrial Problems; Nondestructive Testing, !PC 40. Frederick, J. R. and Felback, D. K.; Dislocation Press, Guilford, UK, June 1974; pp. 137-144. 111otion as a Source for Acoustic Emission; Acoustic 54. Liptai, R. G., et al.; Acoustic Emission Techniques Emission, STP 505, American Society for Testing in Materials Research,· International Journal of and Materials, 1972; pp. 129-139. Nondestructive Testing, May 1971; pp. 215-275. 41. Gillis, P. P.; Dislocation Motions and Acoustic Emission Monitoring Emissions; Acoustic Emission, STP 505, A.t-nerican 55. Nakamura, Y.; Acoustic Detection of Cracks in Complex Society for Testi_n.g and Materials, 1972; pp. 20-29. System for Evaluation, ASNT, Columbus, 42. James, D. R. and Carpenter, S. H.; Relationship Structures; Materials 1971; pp. 8-12. Between Acoustic Emission and Dislocation 0, January A. A. and Smith, B.; Stress Wave-Emission Kinetics in Crystalline Solids; Journal of Applied 56. Pollock, of a Military Bridge; Nondestructive Science; AlP, Menasha, Wis, November 1971; pp. Monitoring Testing, !PC Press, Guilford, UK, December 1972; 4685-4697. pp. 348-353. 43. Smith, S. and Morton, T. M.; Acoustic-Emission A. 0.; A Reliability Approach to the Detection-Technique for High-Cycle-Fatigue 57. Payne, Fatigue of Structures; Probabilistic Aspects of Testing; Experimental Mechanics; SESA, Westport, Fatigue, STP 511; America..Tl Society for Testing Conn, May 1973; pp. 193-198. and Materials, 1972; pp. 106-155. 67