ML20091A898

From kanterella
Jump to navigation Jump to search
Failure Analysis of Type 304 Stainless Steel 4-Inch Recirculation Bypass Lines of Monticello BWR Power Plant
ML20091A898
Person / Time
Site: Monticello 
Issue date: 04/30/1976
From: Cheng C, Danyluk S, James Park
ARGONNE NATIONAL LABORATORY
To:
Shared Package
ML20091A893 List:
References
NUDOCS 9105220311
Download: ML20091A898 (64)
v • d • e


Text

{{#Wiki_filter:. _ _ 1 FAILURE ANALYSIS OF THE TYPE 304 STAINLESS STEEL 4-in. RECIRCULATION BYPASS LINES OF THE MONTICELLO BkV P0k'ER PLANT

  • by S. Dcmyitsk, J. Y. Park, and C. F. Cheng k

ARGONNE NATIONAL LABORATORY 9700 South Cass Avenue Argenne. Illinois 60439 Materials Science Division April 1976 i I I NOTICE: This report was prepared as an account of the work performed for the i Northern States Power Company under Ack. No. 5693. It is not to be r distributed, referenced, excerpted, quoted, or reproduced in any form without i the permission of Argonne National Laboratory and the Northern States Power Company. 9105220311 760803 PDR ADOCK 05000263 p PDR

Tabic of Contents P_ag3 List of Figures. . iii List of Tables vi

SUMMARY

1 1. Introduction 3 II. Description of 4-in. Line and In-service Environmental Conditions 3 A. Geometry 3 B. Material and Welding Procedure 4 C. Vibration Analysis, Water Chemistry, and Radioactivity 4 III. Examination of Pipe. 5 A. Defective Pipe Removal and Nondestructive Testing.. 6 B. Crack Verification 7 C. Macroscopic and Microscopic Investigation. 8 1. Macroscopic and Metallographic Examination 8 2. SEM Examination. 9 3 TEM Examination.... 10 D. Quantitative Cheuical Analysis 11 i 1. Corrosion Products 11 2. X-ray-dispersive Analysis. 12 3. Sensitization. 12 E. Wall-thickness Measurement and Stress Analysis 13 1. Circumferential 'lall-thickness Measurement 13 i 2 Stress Anal'ois of Cold Springing and Dead Weight 13 i F. ID Surface Examination 15 l IV. Discussion. 15 A. Environmental Effects. 16 B. Microstructural Effects 16 1 ~

__._._._____..m. Table of Contents (Contc. ) W C. Stress Effects. 17 Acknowledgcents 20 References. 21 APPENDIX, 22 i e e l l l I l

List of Fip,ures Fig. 1. Schematic showing the BWR water loop system. The 4-in, lines are indicated along with the gate valves and water flow direction. Fig. 2. Loop A of the 4-in. recirculation bypass line. The weld locations and bypass valve are indicated. Water flow proceeds as shown. The cut is also shown along with the hanger location. Fig. 3. Loop B of the 4-in. recirculation bypass line. The weld locations, bypass valve, and water flow are shown. The cut is shown along with the hanger location and cold springing. Fig. 4. Radiograph and ultrasonic results of welds from Loop A. The figure is an axial unfolding of the pipe with the weld shown at the center of each diagram and the clock positions numbered at the top. The radiograph indications are shown by the straight lines, and the UT indications are represented by rectangles. The view is from the pipe OD, The pipe and elbow sides of the weld are also shown. Fig. 5. Radiograph and ultrasonic results of welds from Loop B. The figure is an unfolding of the pipe with the weld shown at the center of each diagram and the clock positions numbered at the top. The radiograph indications are shown by the straight lines, and the UT indications are represented by rectangles. The pipe and elbow sides ~ of the weld are also shown. Fig. 6. (a) Schematic of Loop B with the elbow containing welds RBBJ-9 and RBBJ-10. The clock positions are shown. (b) A three-dimensional schematic showing weld RBBJ-10. The veld bead and clock positions :re shown. The cracks were located at the 71/4 to 8 3/4 clock position. One (90% throughwall) crack was observed in the RAZ, and two others were at the weld root. Fig. 7. (a) The ID surf ace of weld RBBJ-10. The weld bead (B) and elbow (A) and pipe (C) sides of the weld are indicated. The clock positions are 8 3/4 at the left and 7 1/4 at the right. (b) A higher-magnification photo of (a). The machining marks at the surface are readily seen. The three cracks are visible (arrows). Fig. 8 A cross-sectional view of weld RBBJ-10. The crack penetrated 90% throughwall after it had originated at the 1D surface (A). A weld imperfection is also visible (B). Fig. 9. An enlarged view of the crack in the HAZ. An oxalic acid etch was used to delineate the grain boundaries. The dark irregular-shape (arrows) may have contained grains and/or precipitates. Crack areas branching is also observed (A). Fig. 10. The cracks at the weld root are shown at the pipe side (A) and elbow side (B) of the weld. The cracks are shallow and transgranular and have blunted tips. The grain size dif f erences are attributed to the thermal history of the elbow, i 111

-- -- -~---- ' List of Figures (Con td. ) Fig 11. Grain structure of the elbow material. (a) Small grains (due to re-crystallization) are at the ID surface (arrow). Large grains follow the veld fusion line. (b) A magnified view of the ID surface. The notching (arrows) that ha.s occurred is seen at the grain boundaries as well as at the center of grains. Fig. 12 Schemstic diagram showing the orientation of the cracks in relation to the pipe geometry. The fracture surface to be examined with the ) SEM is shown. Fig. 13. The mating pieces of the fracture surface are shown. The regions labeled A and B were saw cut prior to separating the two pieces. Ductile tearing (arrow) appears as high-contrast areas on the photo-graph. Fig. 14 The fracture surface as seen with the SEM. Photograph (a) was taken after exposure of the surface to the NS-1 solution for 100 h. Photograph (b) is the same region af ter 219 h in the solution. " Tongue" behavior (arrow) can be seen. These two photographs show the ef fects of the NS-1 solution on removing the surf ace oxide (A). Fig. 15. A.. SEM photograph showing the fracture surface. The high-contrast regions are oxides that have become, charged (arrow). Precipitates are also seen at the grain boundaries. Fig. 16. Photograph showing slip (arrow). Fig. 17 Photograph showing slip (arrow), grain-boundary separation (A), and crack arrest (B). Fig. 18. Photograph showing transgranular fracture (arrow) at an isolated grain on the fracture surface. Fig. 19. Photograph showing intersection of the ID surf ace and fracture surface Also, pitting (arrows) is shown at the grain boundary and center of the grain. Fig. 20. (a) Crack branching, grain-boundary separation, and precipitates are shown (arrow). (b) Enlargement of (a). Dark circular areas, where precipitates had been (arrow), are seen. X-ray flourescence was performed on precipitates, bright area labelled A,and the bulk grains. Fig. 21. Cross-sectional view of weld RBBJ-10. The areas punched and used for TEM are numbered. Fig. 22. Bright-field (a) and dark-field (b) TEM of a grain boundary (arrow) of region 1 of Fig. 21. A grain boundary with precipitates (bright areas) is shown in (b). Fig. 23. Dark-field (a) and bright-field (b) TEM of a grain boundary (arrow) of region 3 of Fig. 21. Precipitates (bright areas in the grain boundary, (a)) are seen. iv

List of Figur,es (Cont d. ) l Fig. 24 Bright-field (a) and dark-field (b) TEM of a grain boundary (arrow) of region 4 of Fig. 21. Fig. 25. X-ray fluorescence yield of the grains at the fracture surface. Si, Cr, Fe, and Ni peaks are observed. The relative magnitude of the peaks shows that the material is stainican steel. Fig. 26. X-ray fluorescence yield of a precipitate embedded in a grain boundary at the fracture surface. S, Cr, Fe, and Ni peaks are observed. The relative magnitude of the peaks (compare with Fig. 25) suggests that the particle is made of a chromic-rich spinel. Fig. 27. Sensitization test of an area that includes the RAZ. The samples were immersed in oxalic acid for a given length of time so that grain-boundary etching could occur. The weld fusion line is shown in (a), and each succeeding photograph [(b)-(h)] is of an area 0.1-in, farther from the weld. The heaviest sensitization is seen at a distance of 0.5 in, from the veld fusion line. The crack occurred at a distance of 0.3 in, from the veld. The horizontal lines are drawing lines produced during pipe fabrication. Fig. 28 Relative sensitization va distance from the fusion line. Sensitization is obtained by using the QUANTIMET scheme of measuring the relative area (dark area / light area) of etched grain boundary from Fig. 27 The maximum in sensitization occurs at 0.5 in., whereas the crack occurs 0.3 in, from the weld fusion line. Fig. 29. Circamferential wall thickness vs angular position. The circum-ferential wall-thickness measurement was accomplished for weld RBAJ-7 as shown in the upper right corner of the figure. The data were obtained 1/8 in. from the weld root. Fig. 30. Schematic of cold-spring measurements. Fig. 31 The ID surface of RBBJ-10. The machining marks (A) and f racture surface (B) are seen. The crack probably originated in a region where machine marks exist; however, the crack edge (arrow) in this figure is outside the machined region. Fig. 32. (a) The machining marks (A) and crack at the weld root are shown (arrow). The weld is at the right. Solidification marks (B) are also shown. (b) The crack at the weld root is shown magnified. The veld is at the right side of the figure. Precipitates at the grain boundaries are shewn (arrow). The crack follows the weld root and is both transgranular and intergranular. i l l ---e -a-- e- ,----------------,.--..,--,a ~, - n r

List of Tablen Table 1. Radioactivity Table II. Corrosion Products in the 4-in. Recirculation Bjpass Lines (Monticello, Quad Cities-III, and Dresden-II) Table III. Results of Stresses Dec to CGd Springing in Loop A Table IV. Results of Stresses Due to Cold Springing in Loop B Table V. Results of Stresses Due to Dead Weight in Loop A Table VI. Results of Stresses Due to Dead Weight in Loop B O M 6 Vi .. ~

FAILURE ANALYSIS OF THE TYPE 304 STAINLESS STEEL 4-in. RECIRCULATION BYPASS LINES OF THE MONTICELLO BWR POWER PLANT

  • by S. Danyluk, J.1. Park, and C. F. Cheng

SUMMARY

The intergranular cracking observed in the Type 304 stainless steel 4-in. recirculation bypass lines of the Northern States Power Company (NSP) Monticello power plant is similar to that observed in other boiling-water-reactor (BWR) systems. These failures are attributed to intergranular stress-corrosion cracking (ISCC). Contributing factors are: local variations in water chemistry, changes in the metallography of the heat-affected zone (HAZ) as a result of welding, and localized tensile stresses. Two types of def ects were discovered in the initial examination of the 4-in. line. A major crack, was located in the HAZ of weld RBBJ-10. It was determined to be 90% through the wall. It followed the circumference of the pipe, and is intergranular in nature. Shallow crack-like defects were also found at the veld root. These were both intergranular and transgranular. All the defects initiated at the inside diameter (ID) surface. The influence of sensitization on crack initiation and propagation is not firmly established. A sensitization test showed that the major crack was not in an area of maximum sensitization; this crack was 0.32 in. from the weld, and the maximum sensitization was 0.5 in, from the weld. Trans-mission-electron-microscope (TEM) examination showed that precipitates are It is present in the grain boundaries in the vicinity of the major crack. l not known how these precipitates affect the grain-boundary energy and/or l l influence crack propagation. l l l

  • Work supported by the Northern States Power Company.

2 Tensile stresses in the IMZ are required for ISCC to occur. These stresses are a result of residual stresses f rom ID surf ace preparation (e.g., machining) and weld shrinkage and externally applied loads such as fit-up stresses, dead weight, themal expansion, and pressure. In other BWR systems, grinding was assumed to be a major contributor to the cracking problem since it would greatly increase residual stret,ses, llowever, in the present system the ID surfaces were machined. Since the correlation between machining and residual stress has not been established, the magnitude of this stress is not known. The stress analysis of Loops A and B, including cold springing and dead weight, showed that large st resses were present at the elbows and joints. Residual-stress measurements have shown that cracks in the 4-in. line are in a region of maximum tensile stress (0.3 in, from the weld root).1 These ~ stresses 'which result from veld shrinkage, add to the externally applied loads. Metallurgical and geometric factors also affect the tensile stresses. Variation (N10%) in pipe wall thickness was measured within the IMZ. Recrystallization at the ID surface due to (a) surface deformation by mechanical working and (b) the weld heat input, provides localized areas at the grain boundaries for crack initiation. The water chemistry at the crack tip and at localized regions (pits) of the ID surface is not known. Monitoring of the reactor water and feed water did not show unusual chemical constituents. The radioactivity of the pipe was high; higher than Dresden-II and Quad Cities-II piping. It has been speculated that component radioactivity is closely tied to the corrosion properties of the surfr.e.2 Active ions are thought to diffuse to the pores of the oxide film and become incorporated into the oxide structure. Thus, monitoring of the corrosion-product activity may provide information on the corrosion rate.

3 Af ter the work on veld RBBJ-10 was completed, additional data (x-ray) became available which showed that circumferential and axial defects may be pretient in other velds of the 4-in. line. A metallographic examination of these defects was not made. I. Introduction On January 20, 1975, during in-service inspection, linear indications were detected by ultrasonic testing (UT) in two welds in Loop A and one weld in Loop B of the Monticello BWR recirculation system 4-in.-diam bypass lines. A refueling outage was in progress at the time. Radiographic examination of these three welds verified the UT results in one of these welds. The radio-graphs indicated circumferential cracking in the weld 11AZ. No obvious ex-planations that would account for the cracks were apparent, but it appeared that the cracks were similar in nature to those observed in the Millstone, Dresden-II, Quad Cities-I and II, Fukushima, and Peach Bottom-III BWRs. The sections of Loops A and B that contained the suspected cracks were removed and shipped to Argonne National Laboratory (ANL) for metallurgical examination. The present report is a summarv of the failure analysis con-ducted at ANL. II. Description of 4-in. Line and In-service Environmental Conditions A. Geometry Boiling-water reactors have two 28-in, recirculation lines. The 4-in, lines bypass a 28-in, gate valve in the recirculation lines. A schematic of the pressure vessel and recirculation system is shown in Fig. 1. The bypass lines supply reactor coolant water to warm the 2S-in. header before the re-circulation pump is started and relieve the pressure on the gate valve in

4 the 28-in line. Detailed schematics of the 4-in, bypass lines of the Monti-cello plant ate shown in Figs. 2 (l.oop A) and 3 (Loop B). Thu bypass valves shown in these figures remained open during reactor operation, contrasting with the Dresden and Quad Cities plants, which operated with the bypass valves closed. The weld locations are noted in the figures and are labeled, a.g., RBAJ-9. The water flow proceeds as indicated. The welds suapected of having defects are RBAJ-8, RBAJ-7, RBAJ-6, and RBBJ-10. The hanger positions are also labeled. B. Material and Welding _ procedure a All the piping is Schedule 80, Type 304 stainless steel. The elbows are of the same heat <.f material as are the pipes, although they have been subjected to additional heat treatments in the forming process (hot bending). The trans-mission electron micrographs suggest that the elbow was not completely solution-ized af ter forming. The exact welding procedure is not availabic but welding was conducted according to the ASME Nuclear Code, Section IX. The open-root technique of butt welding was employed using the tungsten-inert-gas and shielded-metal arc processes. The joints were prepared with a 70' minimum fit-up angle, and the ID surf aces were counterbored. The beveled edges, root face, and back side of each weld joint were cleaned prior to welding. Certified ER308 and E308-15 filler metal was used with a specified maximum interpass temperature of 350*F. Argon was used for torch shielding, and purge gas. C. Vibration Analysis. Water Chemistry, and Radioactivity A test program to monitor the pipe displacement and discharge pressure fluctuations was conducted by the Bechtel Power Corporation. Pressure trans-ducers were mounted on an instrument line outside the drywell and connected with 1-in, tubing to the recirculation system piping. Acceleration data were collected by a portable accelerometer that measured the vibration of the 4-and 28-in. lines. Data were collected as the pump speed and motor power were varied. It was found that the dynamic amplitudes were small (<1 mil) and - - ~ -

5 Irequencies were lower than 12.5 Hz. Turbulence in the line results in pressure fluctuations as the pump notor speed is increased. The Bechtel report concluded that the combination of low vibration amplitude and low frequency presented no hazard to the bypass-line integrity and that "it is better to operate with the bypass valve open," as had been standard practice in the Monticello plant. The water chemistry in the bypass line is not monitored. However, the feed water and reactor water are monitored for pH, conductivity, chlorides, solids, and metals. The results of the reactor-water chemistry analysis 170-200 ppb, 1 1 1 iodine showed the following chlorides <10 ppb, silicates 0.16 uCi/ml, and pH between 6.5 and 8.5. A f eed-water oxygen level of 50 ppb and a conductivity of 0.1 umho were also measured. The results are within the B'n'R operating-condition limits. The radioactivity of the pipe was measured at ANL under three different conditions: as received, after soaking in a solvent (NS-1 ) for 100 h at 250*C, and after ultrasonic cleaning. The results of the S, Y, and a-activity are shown in Table I. The solvent and ultrasonic cleaning conditions removed the loose corrosion scale from the ID surface of the pipe so that metallurgical investigation could proceed without undue precautions. The a-activity is a result of fuel-element leakage during reactor operation. III. Examination of Pipe As a result of the failure analyses performed at ANL on other Typa 304 stain-less steel piping (Dresden units I and II and Quod cities-! and II) a flexible, specific program of data acquisition that has proven useful for identification of the nature of the cracking has been developed. The program calls for: field

  • Proprietary decontamination solution of Dow Chemical Company, Nuclear Industrial Service.'

l 6 nondestructive examination (NDE), water chemistty analysis, stress measurements, corrosion-product sampling, recording of cold springing, cutting and packaging to prevent 1D surface contamination, decontamination of activity by rinsing, and laboratory investigation that involves NDE, corrosion-product analysis, and metallographic and electron microscopy. Various aspects of the analysis performed on the Monticello bypass piping are discussed below. A. Def ective pipe Removal and Nondestructive Examination The defective pipe sections were removed from the plant with a motorized hack saw. The cuts were made under " dry" conditions (no lubricant), and the pipe ends covered with polyeti 'lene sheet after cutting to prevent ID surface contamination. The inventory of pipe, included three 90' e2 bows, one 45' elbow, and one straight piece, as shown in the Appendix. Nondestructive examinations were conducted on all the pipe velds received by ANL. These examinations included single-wall radiographic, ultra-sonic (shear wave only), and dye penetrant methods. The radiographs were taken with a source-to-film distance of 43 in., with the film mounted in conventional flexible cassettes. The ultrasonic examination was conducted using a conventional hand-held scan (2.25-MHz, 1/4 x 1/4-in., 45' shear-vave transducer). The amplitude sensitivity was calibrated by means of a flat piece of 0.600-in.-thick Type 304 stainless steel with a 20% wall-thickness rectangular notch. No artificial damping or rejection threshold levels were used. The gain of the instrument was set to yield 80% of full-scale deflection when the ultrasonic signal reflects frem the 20% wall-thickness notch. During scanning the gain was increased 12 dB. A dye-penetrant exam;.s'. ion that employed Magnaflux Zyglo wrs performed using standard techniques. The results of the UT and radiographs for the nine welds investigated are shown in Figs. 4 and 5. These figures are schematics of the unfolded pipe with the weld root separating the pipe and elbow sections. The numbers at the

7 top of the figure denote the clock positions. Ultrasonic test indications are present on both the elbov and pipe sides of the veld; however, it is not possible to determine the orientation of the indications due to geometric masking of the UT signal. Radiographic indications appear on the pipe side of the weld at all the joints except RBAJ-8 and RBAJ-9. The latter two welds have radiographic indications on both sides of the joint. Dye penetrant examination of the radiographic indications located in velds RBBJ-12, RBAJ-9, RBAJ-8 and RBAJ-6 did not confirm the presence of cracking in these joints. Prior to in-depth laboratory radiography of all welds, it had been de-cided that veld RBBJ-10 would be subjected to a detailed metallographic examination. Preliminary field ultrasonic examination had indicated the presence of a significant defect in this joint. Subsequent field radiography ~ had verified the presence of a crack. Further metallographic exploration of indications located through ultra-sonic or radiographic examinations of welds other then RBBJ-10 is not planned. B. Crack Verification Upon receipt at ANL, the sections were scrubbed with a clean nylon brush and distilled water, on both outside and inside surfaces, for removal of loose radioactive contamination. Corrosion product was taken f rom the ID surf ace near weld RBAJ-8 by scraping and analyzed by x-ray powder diffraction. Weld RBBJ-10 was cut f rom the pipe using a motorized hack saw and included N1/2-in. of material on each side of the weld head. The elbow geometry is shown in Fig. 6a. Figure 6b shows a three-dimensional view of the weld and the ob-served major crack, which e (tends f rom the 71/4 to the 8 3/4 clock position.

l 8 Two weld-root defects were also observed. The ID i,urface of the pipe is shown in Fig. 7a with the weld bead and defects. The elbow side of the veld is toward the top of the figure. A higher-magnification photograph is shown in Fig. 7b. The major crack can be seen on the pipe side of the veld bead. The major crack f ollows the circumference and jogs along the counterbored machine marks. The weld-root defects are visible in Fig. 7b. Figure 8 is a cross-sectional view of the weld that shows the extent of crack penetration. It is clear from the figure that the crack has initiated at the ID surface and has penetrated at least 90% through the wall. Also visibic in the figure is an inclusion in the weld (center). C. Macroscopic and Microscopic Investigation 1. Macroscopic and Metallographic Examination A, composite cross-sectional meta 11caraphic view of the major crack is shown in Fig. 9. Sample preparation included metallographic polishing prior to etching in a 10% oxalic acid solution using a current density of 1 A/cm to delineate the grain boundaries. Several interesting features in this figure are worth noting. The crack is intergranular, and dark irregular-shape areas, which may have contained precipitates or grains, are evident. Branching of the main crack is also readily seen. A cross-sectional view of the crack-like defects at the weld root is shown in Fig. 10. These typical weld-root defects, which have blunt tips, occur on both sides of the weld. The grain size of the base material on the pipe side differs from that of the elbow nide because of the mechanical working and heat treatment received by the elbow during the forming (hot bending) process. The grain l size in the bulk of the eftew enterial is larger than at the ID surface. This difference (Fig. 11) is partially due to the forming process and the heat provided during welding. 'igure 11b, which is a magnified view of Fig, lla, l

/* 9 shows surf ace " pitting," the pref et ential attack at the gr.cic boundaries possibly due to the use of NS-1 decontamination solution. The weld was sectioned as shown in Fig. 8, and the crack was opened to reveal the fractu'e surface. Figure 1.2 shows a three-dimensional schematic with the sectioning geometry and major crack location. Figure 13 shows the two mating pieces after the major crack shown in Fig. 12 had been opened. The regions labeled A and B were car using a cutof f wheel with a silicon carbide blade. The crack front separates the high contrast region from the Care was exercised not to disturb the crack f ront intergranular cracked region. af ter saving, the sample was cooled to liquid nitrogen temperatures euch that, and the crack was opened by impact fracture. 2. SDi Examinatien The f racture surf ace of the major crack was subjected to scanning-electron-microscope (SDI) examination af ter removal of the oxide scale; the surf t :e was then gold shadowed and reexamined. The oxide was removed by exposure of the f racture surf ace to the NS-1 solution (250'C f or 100 h). The effective-ness of this solution for removing the oxide scale has been the subject of in which it appears that the interface between the oxide and 6 a recent study the base metal is preferentially attacked. The grain boundaries of the base metal are also attacked such that precipitates which may have been present are dissolved away. The effects of the decontamination solution are seen by comparing Figs. 14a and 14b. These photos show the morphology of the surface after N100 (Fig. 14a) and 219 h (Fig. 14b) in the decontamination solution. " Tongue" behavior, fine slivers of metal usually found on cleavage f acets, is apparent in these figures, it has been shown that such features result from plastic deformation (cleavage across microtwins) at the tip of the main propagating crack. The fracture-surface morphology is shown in Fig. 15,

____ _.-.~._ ~ ~ _ -. _. _ _ _ _ _ _ l 10 i where the unremoved oxide scale (bright areas) in also visible. The high-contrast regions, indicative of oxide charging due to electron-beam bombarding, ] made high-resolution microscopy difficult. Gold shadowing of the fracture I surf ace resolved this problem by providing a thin conducting layet, which leaked accumulated charge away. A disadvantage of thic technique is that the chemistry of the original surface is changed such that quantitative chemical information is lost. The observation of slip on individual grains was common, and some examples are shown in Fige 16 and 17. Also shown in Fig. 17 is a case of branching-crack arrest at the intersection of three grain boundaries. Transgranular fracture of some grains was also observed, as shown in Fig. 18 Figure 19 shows the ID surface at which the crack initiated. Also visible in the figure are pits that formed possibly during pickling when the material was fabricated. These pits are seen both at the center of the grain and at the grain boundaries. The main crack branched during propagation. A characteristic of these branches is " clean" grain-boundary separation. A typical branched crack is shown in Fig. 20. Also seen in this figu 6 ate precipitates, which appear as bright contrast areas embedded in the grains, and dark circular areas from which these precipitates have been removed. The SEM study showed that the crack is predominantly intergranular, with some transgranular fracture also occurring. Twinning cleavage fracture and slip were identified but could not be attributed to fatigue. Pitting of the ID surface was also observed; however, a direct correlation to crack initiation has not been made. 3. TEM Examination Transmission-electron microscopy was performed on three thin disks (numbered 1, 3, and 4 in Fig. 21) punched from the ilA2 near the weld. Disks 3 and 4 were adjacent to the major crack. The grain boundaries were examined . ~. -.

11 in each case. Photographs of these grain boundaries are shown in Figs. 22-24, which display bright-and dark-field views. Care was taken to expose any precipitates that may have been present. The micrographs show discontinuous precipitates in position 1, a few traces of precipitates in position 3, and no C precipitates in position 4 The precipitates, usually identified as My3 6 carbides, are thought to contribute to ISCC. The presence of precipitates on the elbow side suggest that the solutio'nizing treatment was not complete. Continuous precipitates were not observed in these samples. It must be noted, however, that the small number of grain boundaries examined does not represent a statistically significant sample. Additional TEM studies are required to correlate the suscept-ibility of this material to ISCC with precipitates at grain boundaries. 3 Quantitative Chemical Analysis 1. Corrosion Products Tne loose and adherent corrosion products at the ID surfaces were analyzed The by the x-ray powder-diffraction technique using CuKa and FeKa radiation. adherent corrosion products were removed by light mechanical scraping before cleaning and washing. Table II lists the types of oxide scale found in the Results from Monticello plant pipe and their relative x-ray line intensity. Quad Cities-III and Dresden-II are included for comparison. In general, no significant difference in corrosion products exists among the three reactors, and no unusual products are found. Spinels of nickel ferrite and nickel chromite are tightly bound to the surf ace substrate. The he=atite (a-Fe O ) is probably a redeposition product f rom the oxygen-rich 23 In region in the core that results from radiolytic decomposition of water. the case of Quad Cities-III the hematite is probably converted If combite, which results from scrubbing with trisodium phosphate. The austenite is i likely to have resulted from metal dissolution, whereas kamacite may originate from transformed martensite at the machine-disturbed metal surface or from l debris due to galling and wear. t

12 2, X-ray-dispersive Analysis X-ray-dispersive analysis was perforned on the precipitate particles and substrate material. The x-ray intensities (CuKa radiation) are shown in Figs. 25 and 26. Chromium, iron, nickel, and silicon are present in both the precipitate and substrate; however, sulfur is observed only in the precipitate. Dif f erences in peak height are readily seen and have previously been attributed to oxide spinels. 3. S_ensitization The degree of sensitization in the llAZ was examined using ASTM A262 Practice A. This procedure is a rapid means of determining whether austenitic stainless steel is susceptible to intergranular attack. A cross section of the weldment shown in Fig. 8 was polished (3/0 final polish) and etched in a 10-wt% solution of oxalic acid for 1.5 min at a current density of 1 A/cm. The acid preferentially attacks carbides, which, in a sensitized stainless steel, have precipitated at the grain boundaries. Thus, the etched structure 4y be used to approximate the degree of sensitization. The results from the above procedure are shown in Fig. 27. The series of photographs in *he figure show the weldment (Fig. 27a) and successive 0.1-in. intervals along the axial direction of the pipe (Figs. 2 7b-h). Variations in grain-boundary etching are readily observed, with the maximum occurring %0.5 in, from the weld fusion line. Deformation lir s parallel to the direction of the pipe axis are visible in Figs. 27f-h and are the result of the pipe fabrication (ex-trusion) process. A quantitative measure of the degree of sensitization, i.e., the degree of grain-boundary etching by oxalic acid, was obtained by measuring the ratio of the area of etched grain boundaries to the total surface area, as obtained from the photographs of Fig. 27. These results, obtained with the IMANCO QUANTIMET 720, are plotted in Fig. 28. In addition to the maximum sensitization,

/ 13 f The back-NO.5 in. f rom the weld fusion line, a crack is shown at so.3 in. ground signal, due to the deformation lines, has increased the apparent value of sensitization at large distances from the weld fusion line. The true sensitization reading should be much lower and decrease to zero as the distance is increased. E. Wall-thickness Measurement and Stress Analysis 1. Circumferential Wall-thickness Measurement Wall-thickness ocasurements around the circumference were taken 1/8 in, f rom the f usion line of weld REAJ-7 The results, shown in Fig. 29, are It is apparent that plotted as wall thickness versus angular position. significant thickness changes, which could lead to stress concentrations, Weld shrinkage may may be present in the HAZ as a result of weld shrinkage. also lead to cold cracks due to tensile stresses. It may be assumed that measurements for the other welds will show similar variations in thickness. 2. Stress Analysis of Cold Springing and Dead Weight Stresses in piping may reach values close to or in excess of yield. These stresses arise from externally applied loads such as dead weight of the pipe, operating fluid pressure, thermal stresses induced at the operating fit-up stresses imposed during plant assembly, and residual temperatures, Each stresses locked into the microstructure by cold work and weld shrinkage. effect adds to the bending, axial, and hoop stresses and shear stress due to torsion. Residual stresses near the welds have not been measured. However, these measurements are being conducted as part of an ongoing program sponsored by the Electric Power Research Institute (EPRI). The results are not available to date. A stress analysis of both the A and B loops that included cold springing, dead weight, pressure, and temperature taasurements was performed. The

14 cold springing of both loops is shavn in Fig. 30, with numbers indicating straight-line pipe segments (elements). Loop A was severed at the jenetton of elements 3 and 4, and the measured separations were 1/2 in, along the pipe axis and 3/8 in, perpendicular to the axis. The cold springing of Loop B, measured at the junction of elements 4 and 5, was 1/2 in, along the pipe axis. The stresses were calculated assuming a three-dimensional beam analysis, with the flexibility of the elbows being neglected. Also, the 4-in, pipe section The analysis was assumed to be connected at both ends to immovable supports. is done by inputting the geometry and external constraints into the SAP IV Pipe Bend Element Code. The output of this code is fed into the IBM SSP subroutine SIMQ, where the unknown loads are calculated. The stresses are computed by dividing the loads by the moment areas. The results of the analysis, considering cold bending alone, are shown in Tables III and IV, which include maximum bending stress ( B), uniform axial stress ( T), and shear stress due to torsion (T). The beam-type loads due to pressure have been neglected because of their small value. Since the bypass valve on both 4-in, lines was in the open oosition, the temperature difference between the 4-and 28-in. line is assumed small; nsequently, stresses The results listed imposed by thermal expansion have also been neglected. in the tables show that (a) high stresses are present at the connection between the 4-and 28-in, lines and are distributed through a weld-o-let and (b) the stresses are distributed through the other welds, with certain welds showing higher stresses than others. For example, welds RBEJ-9, RBBJ-10, and RBBJ-ll show a relatively higher stress than the other welds in Loop B. The stresses due to dead weight have also been computed, and the results are shown in Tables V and VI. The variation in stresa for the welds can be seen. The stresses in Loop B were higher than those in Loop A. A conservative estimate of the cumulative ef fect of dead weight and cold spring l

15 stresses can be ma'e by simply adding the values at each veld. in We believe the absolute magnitudes of the quoted stresses are not themselves especially significant because of the simplifying assumption made in the analysis, although the relative values of stress may be significant. The stresses due to cold springing and dead weight will, of course, add to the residual stresses and may contribute to the conditiens for crack initiation. F. ID Surface Examination The ID surf ace of weld RBBJ-10 was examined for unusual surface-pre-paration marks and weld shrinkage. Figure 31 shows the ID surface near the crack. The crack edge and fracture surface are seen at the left of the figure. Machine marks are clearly seen in the center of the field of view and are recognizable because of their regularity. The surface preparation at the wald root is seen in Fig. 32. Figure 32a shows the machining marks due to counterboring (at the left) and precipitates embedded in grain boundaries. The veld-root crack is also shown (cross-sectional view in Fig. 10) and solidification marks can be seen at the right of the figure. A higher-magnification photograph of the crack is shown in Fig. 32b. The crack follows the weld root but not necessarily the grain boundaries. It is surmised that this crack is the result of weld shrinkage which led to tensile stresses at the ID surf ace and is not a stress-corrosion crack. IV. Discussion The factors that cause ISCC of stainless steels in BWR systems have been f discussed in detail by various authors.' No single cause has been found to predominate either in the crack initiation or crack propagation astects of ISCC. It seems clear, however, that ISCC is promoted by environnantal ef fects such as the chemistry of the water in contact with the steel, micro-i ^ ~ ~ ~

16 structural effects at the surface as well as the bulk of the matrix, and/or stress effects which result from residual stressts left by cold working or weld shrinkage. We proceed with a discussion of these three aspects as they pertain to the present failure analysis. A. , Environmental Effects The role that chemical impurities and temperature play in ISCC in BWR systems has not been established. However, it has been suggested that primary water lines of pressurized-water reactors (PWRs) which contain oxygen of <0.1 ppm do not show intergranular cracking. O Concentrations of oxygen in the Monticello plant feedwater were continuously monitored and showed a value of <50 ppb, although concentrations in the 4-in. lines and crevices and notches at the pipe surface may be substantially different. Large increases of the oxygen level in the plant water have also been reported.11 These occurred during plant shutdown; however, direct correintion with ISCC has not been determined to date. The transport of other impurities through the water has been reported in PWR plants, and it was suggested that these may 1 play a role in the failure-initiation stage of ISCC.12 Unusual chemical activity has not been identified in the water nor in the material. B. Microstructural Effects The relation of microstructure to ISCC in BWR systems is not clearly The grain size, degree of sensitization, and surface cold working understood. Sensiti-all contribute to some extent to ISCC in Type 304 stainless steel. zation is caused by precipitation at the grain boundaries of chromium carbides The cracks do not, however, and/or compounds that contain Mn, P, S, Si, and N. form in the regions of highest precipitation but in areas that are only lightly The 4-in. lines of Dresden, Quad Cities, and Monticello all sensitized. from the weld fusion line (a region of light developed cracks at NO.3 in.

17 s ens i ti r...lon). The role that sensiti:ation plays in crack initiation has not been 11rmly established. The TEM results presented in Figs. 22-24 show the discontinuous nature of the precipitates on the pipe side of the weld near the crack region. Areas examined farther from the weld showed grain boundaries wi thout precipitates. The SEM results revealed precipitates on the fracture surface that may fracture in a brittle manner and weaken the grain boundary, thus providing an easy path for crack propagation and material dissolution. The main crack exhibited f eatures of slip, grain-boundary separation, crack arrest, cicavage, and transgranular fracture. These effects indicate that the plastic deformation of individual grains had occurred as the crack propagated but corrosion also was taking place. Fits were observed at grain bou.daries and at the centers of grains (Fig. 19) and may be due to pickling during pipe fabrication. The association of pits with crack-initiation sites was made, but no direct evidence has been found. The ef fects of cold work and recrystallization combine to change the grain size at the ID surface of the pipe. The correlation of grain size with ISCC has not, however, bean established. C. Stress Effects Stress is necessary for the repeated rupture of the passivating oxide film and subsequent electrochemical dissolution of the metal in the de-passivated areas. The possibic sources of stress are: cold fit-up of the pipe joints prior to welding, hangar-imposed utresses at the operating temperatures, dead weight, pressure, and mechanical cold working of the pipe surface, which leads to residual stresses. Increases in the surface stress may be caused by thickness differences due to pipe weld shrinkage, surface roughness from machining during veld preparation, postweld cleanup, and pickling pits. Stress levels approaching yield are possible if all the effects that increase the stress are accounted for.

._.----. - - __..~ _ - 18 Measurements of the cold springing at room temperature were used to calculate the stresses in the weld joints. The stresses are nominally small, although values of 10 ksi (Table IV) have 'een calculated. The addition of the dead weight and cold springing stresses shows values of approximately 1 half the yield stress. Examination of Tables III-VI and Figs. 4 and 5 reveals that the welds which show the strongast UT and radiograph indications are those with the highest stress values. Weld RBBJ-10 showed a maximum in stress; welds RBAJ-6, RBAJ-7, 'ind RBAJ-S also showed a high stress level and UT and radiograph indications. Residual stresses have not been measured for the present report but are being measured under the sponsorship of EPRI. The pipe thickness of weld REAJ-7 varied by 10%; this variation would change the stress by the same percentage. The ID surfaces of the pipe were counterbored (machined). It has been 1 speculated by other authors that surface deformation leads to residual stresses and grinding is a more severe surface preparation procedure than machining. It seems clear that the stress induced in the surface regions of 4 the pipe will depend on the metal removal rate and the force used in the removal process. In the present case these variables are unknown. Additionni insight into the present problem may be gained from the residual-stress measurements. Weld shrinkage was evident (Figs. 31 and 32) f rom the crack-like defects i observed at the weld root. The nature of these defects is different from that of the intergranular crack which was examined metallographically. Weld shrinkage will induce stresses in the liAZ that will add to tite externally applied loads. 1 -r-w-- e -ee ir~-- - - - - - - - -, m e------er-m

s- - - - -

e--

--- - - + - - - - - - - - - - ' + - ' -

19 Fatigue striations due to alternating stress levels were not observed. Slip was quite evident in many of the grains (Figs. 16 and 17), and the corners of grains had undergone plastic deformation. The accelerometer testa performed by 3echtel Corporation showed that dynamic amplitudes were small (<1 mil) and of low frequency (<12 Hz). llowever, low-amplitude and low-cycle tension-tension stresses may not produco striations and these effects have not been ru3ed out. e 9 1 ~ a

20 Acknnwledgments The authors wish to acknowledge the close cooperation of the many Haterials Science Division staf f and technicians at ANL who made the com-pletion of the present study possible. Special thanks go to G. H. Dragel and J. C. Florek for metallographic work; W. A. Ellingson, D. S. Kupperman, N. P. Lapinski, A. Sather, and K. J. Reimann for the nondestructive testing; A. Purohit, H. W. Knott, and M. H. Hueller f or x-ray analysis, D. E. Busch, J. E. Sanecki, D. R. Diercks, and L. A. Neimark for help with the SEM f acilities; K. R. Ziech for materials handling; and D. K. Cavanaugh, W. H. Allen, l and H. F. Adams for manuscript preparation. W. J. Shack provided the atress analysis. C. H. Harmsen, L. Elianen, and P. Krumpos of NSP provided the criti-l cally needed background and data so that the present work could proceed.

Finally, I

we thank B. R. T. Frost, R. W. Weeks, and R. B. Poeppel for their support and i ) encouragement. i ? i 9 ) t t 4

21 References 1. C. F. Cheng, W. A. Ellingson, D. S. Kupperman, J. Y. Park, R. B. Poeppel, and K. J. Reimann, " Corrosion Studies of Nuclear Piping iu BWR Eri-vironments," Quarterly Report for the Quarter Ending June 30, 1975, Argonne National Laboratory, prepared for the Electric Power Research 2 Institute under contract #31-109-38-3138L. ] 2 D. H. Lister, "The Transport of Radioactive Corrosion Products in High Temperature Water-1. Recirculating Loop Experiments," Nucl. Sci, and Eng. 58, 239 (1975). i 3. J. R. Darnell and S. W. Giampapa, " Dynamic Test Report-Reactor Recirculation Bypass Line Test-Monticello Generating Station," Bechtel Power Corporation, i January 1975, prepared for the Northern States Power Company. 4 D. E. Harmer, F. P. Fauson, M. A. Snyder, O. U. Anders, and J. J. Holloway, " Developing a Solvent, Process, Equipment and Procedures for Decontaminating the Dresden-1 Reactor," presented at the 37th Annual Meeting of the American j Power Conference Chicago, Illinois, April 1975. 5. Craig F. Cheng, " Failure Analyses for Cracked Type 304 Stainless Steel Piping in the 4-in. Recirculation Bypass Lines of the Dresden-II and Quad Cities-II BWR Systems " Argonne National Laboratory, December 1974, prepared for the Commonwealth Edison Company. 6 C. F. Cheng, unpublished. 7 Metals Handbook, Vol. 9. "Fractography and Atlas of Fractographs," 8th Edition, American Society of Metals, Metals Park, Ohio, 1975. 8 Craig F. Cheng and Edward E. Potter, Nucl. Met. 19, 273 (August 1973). 9 Proc. Symp, on Materials Perf ormance in Operating Nuclear Systems, CONF-730801, Nucl. Met. lj[ (August 19 73). 10. Craig F. Cheng, J. Nucl. Mater. 56, 11 (1975). 11. J. Danko, presented at the meeting on BWR Pipe Cracking, O' Hare Hilton hotel, Chicago, March 4, 1975, sponso"ed by the Electric Power Research Institute. 12. T. E. Rummery, 5th Canadian Seminar on Surf aces, University of Laval, Quebec, Canada, June 1-3, 1975 4 . - - - -,.. - - ~ -. _ _,, _ -.

Table 1 Radioactivity E+Y Y a @ 1 in. (ar/h) 0 1 in. (ar/h) (DPM/cm ) 290 50 984 (1180)D As-received ID surface NS-1 solvent 38 498 (250'F, 100 h) Ultrasonically cicaned 0.3 0.1 <1 after US-1 solvent (250'F, 100 h) " Dry smear. The a energies were found for the following: 239 and/or Pu '"; t.m 1 0 and/or Pu

Cm

, possibly Cf Pu and Cm Chemical dissolution method.

l Table 11 Corrosic.n Products in the 4-in, Recirculation Bypnos Lines (Monticello, Quad Cities-111, and Dresden-II)^ Dresden-II Monticello Quad Citten-III Loop A !.oop A gpJ e As-re-As-re-Loose Scrubbed Loose Scrubbed Product ID Product ID y A S M M M M Spinel (tlife 024 + NiCr 0 24 Hematite S M A A M W (a-Fe O ) 23 W W W W S W Kan.acite (a-Fe, BCC) S V S S S S Austenite (Y-Fe, FCC) A A A A W A Ciprite (Cu 0) 2 A A M V A A Lipcombite 3(P0 )2} "}2 Te 4 S = s t r on g, M = me d itas, A = ab s en t, and " Relative x-ray intensity of oxide scales W = veak, bDeionized water scrubbed, Trisodium phosphate scrubbed. c

- _ -. ~. - _ _... -, _ _ _ _ _ _. _ _ _ _ _ _ _ _ i Table III Results of Stresses Due to Cold Springing in Loop A k' eld Identification Number" B 0 7 (RRAJ-) we1d-o-1et 8000 -2800 -2800 13 7200 -4400 12 100 -4400 11 100 10 4400 Cut 1 150 9 2400 8 2400 -150 100 7 2800 -150 100 6 2100 -100 -1000 i weld-o-let 5700 -100 -1000

  1. From Figs. 2 and 3.

1 l .,, ~., _ - _ _, _,

l s I L l Table IV 1 Results of Stresses Due to Cold Springing in 1,oop B Wold Identification Number e o t 3 I (RBBJ-) 1200 veld-o-let 12,300 1200 17 3,800 ? 4,400 700 -400 13 2,200 700 -400 12 7,900 200 1030 10 9,900 200 1030 9 10,300 700 Cut 2 600 700 J 8 600 700 7 400 700 230 veId-o-let 33,000 200 230 l t I i i I ,---m_. p .,.......,,g.,, ,.w%.,.. ,__.,4 .,m-_,,. -.m., - - -

1 3 l d i 1 t l Table V J Peaults of Stresses Due to Dead Weight in Loop A t Veld Identification Number a o t B T (RBAJ-) weld-o-let 500 400 500 13 400 400 500 12 3200 50 1500 11 3000 50 1500-10 4100 Cut 1 1400 t 9 900 8 400 -100 400 7 3500 -300 400 6 3500 -300 300 t weld-o-let 1000- -400 300 l ? I l f f t t. I l \\ l ,. =....... - -.. _.. - -,, _ _. _ -,.... -,.,.. -. -. - -, _, - ~... - - -, _ m . _,...,~,..... - .._,m .,.-.-.,--,.-,.._.-,,e.

I Table V1 Reeults of Stresses Due to Dead Weight in Loop B Weld Identification Number o C g T (RBBJ-) 1900 we ld-o-le t 3200 1900 17 600 16 3700 -80 13 2000 -80 12 1700 30 10 4300 -60 9 4200 -80 Cut 2 2600 -80 -200 8 3500 -80 -200 7 3500 200 -100 weld-o-let 3000 -400 -100

} APPDiDIX RBAJ-6 RBAJ-9 (' RBAJ-7 4" r l RBAJ-8 e 5" x ~P RBB J-10 F RBBJ-Il 5" RBBJ -9 **a n q, 4 f 6" ~ R BBJ-13 ( ' H 6 "H-RB BJ-12== a 4 o

VESSEL a RECIRCULATION 2 NOZZLE N 2 F CORE SUPPORT I STRUCWRE SAFE END n 3 CORE B A a 28" PIPE q GATE ,/ N.., 'N VA LVE S ^

g n -

X X 4"/ \\ GATE VALVE PIPE ki k'; PUMP E VALVE Fig. 1 Schematic showing the INR water loop system. 'lhe 4-in. lines are indicated along with the gate valves and water flow direction. r

. -. -. - - ~_.- - i i 1 s 1 J R B AJ -13 / RBAJ-12 RBAJ-Il j l HANGER I POSITION l CUT 1 J RB AJ -10 RBAJ-9 N BYPASS al VALVE f l / Cs n R B AJ-8 5'- 9" RBAJ-6 q d N RBAJ-7 6" \\ \\ CUT 2 i I Fig. 2. Loop A of the 4-in. recirculation bypass line. The weld locations and bypass valve are indicated. Water flow pro-ceeds as shown. The cuts are also shown along with the hangar location and cold springing. O ,-..-.--.....-y-.m._,,...,,_m.,,,,_ ,,,-,,-,,,..-,_.-._._....c. ,. -. ~.,.-.. -..._.---... ,x

l RB8J-l7 HANGER POSITION l ' l' RBBJ~l6 RBBJ~l3 up RBBJ-12 gypggg i RBBJ -II VALVE SPR NG V 12 # N 2'- R8BJ-15 " CUT 1 u CRACK i R8 8 J -lo RBBJ-8 l / RBBJ-9 46 - / l is 4 .k / a so I o 46/2 4" CUT 2 g k' R88J-7 g i i Fig, 3. Loop B of the 4-in, recirculation bypasa line. The weld locations, The cuts are shown along bypass valve, and water flow are shown, with the hangar location and cold springing. w w,m-. ,,y w --m,. ,#..w -,1y,- ,----.y- --_--,.y,---,,.y-r.--,,,,,-y--

O I 2 3 4 5 6 7 8 9 10 ll 12 13 a PIPE _._ l ___. WELE __ _ RBAJ-9 =. _ _ ELBOW a a a a a a aoooo o o a a a PIPE RB AJ-8 _~ Z l__ _ _y/GD) ___ _._C ~~ ~ ~ ~ ~ ~ ELBOW a aca a a PIPE aa a a cm a -~ ~ ~~ RBAJ-6 [Z_ _T C3CC ~ ELBOW PIPE aaca a aaaaaa a a RBAJ-7 ~_~ __ Z Z Z - __WELE Z __~. - ~~ ~~~ ~~ ELBOW Indications located by ultrasonic examination. a - Indications located by radiographic examination. Fig. 4 Radiograph and ultrasonic results of welds from I,oop A. The figure is an axial unfolding of the pipe with the weld shown at the center of each diagram and the clock positions numbered at the top. The view is from the pipe OD. The pipe and elbow sides of the weld are also shown. )

e 0 1 2 3 4 5 6 7 8 9 10 11 12 13 PIPE l ca _._cs .sa ca. __ __ sa __. .__ __. sa_ ss R BBJ -13 W EL. D_.__ __ _.__ _ EL80W i, i PIPE c_, a T~~~~~-~~~~~~~~_~~.__~__ R88J-12 __ w"c,__c,-.W~F.H) ~~ w._ __ __ __ _.._ f ELBOW PlPE c=, a RBBJ-II ZZ___RQ ~Z- _ Z ~ Z _~ m aa a ~~ ELBOW PIPE a e=1 ~~~ ~_ ~ ~~ RB BJ-10 WRLD - - ~ ~ ELBOW P1PE a c2 a co a ca ~ ~ ~ - ~ R88J-9 ~jETJL. ~~ ~~ ~ Z '~~ ~~ ~ c=r-ELBOW indications located by ultrasonic examination. cm Indications located by radiographic examination. The figuro l Radiograph and ultrasonic results of welds from Loop B. l Fig. 5. 1s an unfolding of the pipa with the weld shown at the conter of each l diagrato and the clock positions numbered at the top. The pipe and I c1 bow sides of the weld are also shown. v-.ee- -t c e e-ee -w w w-e e y--.w-ww,,,.--,ww-v,-- -ur,%w-.,,-wrwe wre ,+m1--m.,--.----,-c.,,.__w-,-,,w -.-- ---. -+ ,,y r p a,- 9wr-.*,-g.,-.--.9- ,.-cy w,--w ---e..,-,

4 PIPE CLOCK POSITION / 6 9 12 j ELBOW l R B B J -lO e I RB8J-9 (0) i i CLOCK POSITION / l 12 DEFECTS AT WELD ROOT N ELBOW 9 3 SIDE 8/4,' l 3 I 7/4 6 WELD BEAD ^ 90 % THROUGHWALL CR ACK Fig. 6. (a) Schematic of Loop B with the elbow containing volds RBBJ-9 and RBBJ-10, The clock positions are shown. (b) A three-dimensional schematic showing weld RBBJ-10. The weld bead and clock positions are shown. The defects were located at the 71/4 to 8 3/4 clock position. One (90% througinaall) crack was observed in the HAZ, and two defects were detected at the weld root. =c---+----,---r.., ,,-,,,._.mye,._,,--,-,-.--.,wy,-- 7,,,..--,.,-,,,,,..,,,,,y.,.,,,,,,,,,r ,y--yy,,.w-,-,,-,._ ww-,_,-,.-.y,w-, ,,,-,+w-,,.,,sw-,

l l l L l ) l i i l i I l l 10 mm A 1 i ) i I ) l B i l l 1 I \\\\ ...p i C I i i l l l (o) l l 1 1 l 4 ^ 10 mm Y."r ,, ne. fc ~ * ., s. r'., ;1.. ' g g,- . g o. e 5 ' '~. h l o. .a \\ ,1 _ g i y* f 9.' . p.a s-).g, . J ), *' N ~,L : *- y s c l "' ' }f

s

.s ' ~ o ~.. -. + l ycyted$ 7 . 4 , He6. g@...% r l

3.. -
  • +

/ - K, 2: l l M.,w m m - d;., o l ~.., ( b) l l Fig. 7 (a) The ID surf ace of weld RBM-10. The weld bead (B) and cibow (A) l and pipe (C) sides of the weld are indicated. The clock positions l are 8 3/4 at the lef t and 7 1/4 at the right. (b) A higher-magnifi-l cation photo of (a). The machining marks at the surface are readily i The major crack and two surf ace defects are visible (arrows). seen. l l ( 'I

e-----m-e------ev,.c3e-e---r,w--r,mcy----w-,e,w.,,gw,-wm,,n w w- - w - w v-

ELBOW iv. PIPE SIDE ? SIDE IOmm Fig. 8 A cross-sectional view of weld RBBJ-10 The crack (arrow) penetrated 90% throughwall after it had originated at the ID sur-f ace (A). A weld im-perfection is also visible (B). ~

- _ - -- - - - - ~ Fig. 9. An enlarged viev l I of the crack in the HAZ. An oxalic acid etch was used to delineate the grain boundaries. g3,,' J The dark irregular-3;7, shape areas (arrow) A may ii ve contained t. ~ 5 grains and/or pre- [;1*' 'I ic p,{*4 branching is also cipitates. Crack - '"' ".j d observed. ...,, kj'D ., ? _;'y r r% .(*4

e -

3. :;.;et g~ ;,;- ? ..:t : 4.m (I f,p

  • f, $.c y

.u. *, i*;e ' U, *(., r& r.:p _ <. :

  • f :,4

~ , r.

p.

l _,. Ne' b 1 : V.

!];,. A.

..m. : o su:, 9 4 < "y e ., -2 9 i i I [ ,.s ,, a. ..;A> . - - ( q. aq'Q,%] v - g, Imm epp 9. q, 4 ev y

t 6 I f r-A J*: +'

p. :.

.u,,.s. k-f 7 - g-j- r r g ~x=,j ;;a.. v. . y m p c7 y f r s~h;- e i 3 p h.,.c b,> i 'i. ~ l ,rw .s . rp. ../ 1, ' ' L l O.i mm, pf4# . O.1 mm, i B j l A f 1 I Fig. 10. The defects at the weld root are shown at the pipe side (A) and elbow side (B) of the weld. These crack-like defects are shallow (0.004") and transgranular and have blunted tips. This type of defect in this location has not contributed to the (BWR) piping i I failures investigated at ANL. The grain size differences are attributed to the thermal j history of the elbow. i I I i i 1 ) l

t t l l J l ,ej ) 9....ii. X .."h ['. ,W Q. ~ .m t.. y ,y. 0

9..

3. ju .s, f ; c - a

5. ^.. Q, m,,.

\\.iy, Q Q+ % *1 t ., m.. p.. ^ =. ~ .7 ' ' M- .,f,f..,o . f. .r*,,.s.. n. (.i (Y,,{u.h,Yj][]5 {y ?x. 4 )

+

s u_..,.... is t y.. ? 3. ..-e., i, 4 +)s 4y s d.(~w' + a. .y .? s . ~ v w r. %. (. N, f u. vQ

- h !{'v.,

v ( O.2mm s ~ ., N  %.'.i,/ n a }*m1.. !,gtu v. h., Y WELD-FUSION (a) LINE .,, $ w 4..-

,e Oi d/

y. .m .c. \\

  • ywi.
. [I.f;ff.. ' (W

,.,3.5 gj, 4 h y g L S ,.'.'Q s,. 1;ut.l? l )W.VY( es

  1. u

) ., y.. u , t 7- $W

  • I c, y

?. .'*- P; 'Qi[, t (

  • c y,: d i.p h a:.

~,. rjph Yf,T.Y b 4 h1 9-

  1. ..e,7 7

(b) yq 1,, ;,,, e .i Af~ A .A g 9 l.j] O. I mm w .I p,< e a Grain structure of the elbow material. (a) Small grains (due to i recryst.allization) are at the ID surface (arrow). Large grains Fig, 11. (b) A magnified view of the ID sur-follow the weld fusion line. The notching (arrow) that has occurred is seen at the face. the center of grains. grain boundaries as well as at -. - ~

- de l ID SURFACE DEFECTS ALONG WELD ROOT WELDMENT p FRACTURE ELBOW SIDE SURFACE MACHINING MARKS - l' / wq ~gj eiee Side MAJOR CRACK l l s l i Fig. 12. Schematic diagram showing the orientation of the surface defects The fracture and major crack in relation to the pipe geometry. surf ace to be cy.amined with the SDi is shown.

l I i' e l ,v e,,v \\ a j l 9 l

1,*

g J. , 4...yi. C', $r [ M y t. .4 5mm 1 .i i l i l i 1 t I L N% [~ 5mm u Fig. 13. The nating pieces of the 1 fracture surface are I sh wn. The regions la-beled A and B were saw cut prior to separating the two pieces. Ductile tearing (arrow) appears as high-contrast areaa on the photograph. J _________,_,___....__,__._,_,-______,,..-.._,,...._m..___.,_

1 1 .I j i i i I -g C g .e l ' <{$' 1 '.;': j r.',ca.f.htI;.MI'E.J.. h '9C I f g%gg,.n.. .t . s.. ,M' Aiy ye 4 A-o.01 mm gy .o 1 (a) 1 i I ls. s j,1 f.. Q I j l* a g i k? ( b) .1. . z.. ;

  • w..

3 ) , '3 .Q? ' h

  • 4) f e

1 j, 0.01 mm '%...N.I.- sq r 1; The f racture surf ace as seen with the SDi. Photograph (a) was Fig. 14. taken af ter exposure of the surf ace to the 11S-1 solution for 100 h. Photograph (b) is the same region af ter 219 h in the solution. " Tongue" behavior (arrow) can be seen. These two l photographs show the ef f ects of the 115-1 solution on re-moving the surf ace oxide (A). l

i i i l 1 i 4 l i l l 0.1 m m 1 l Fig. 15. An SD4 photograph showing the l fracture surf ace. The high-contrast regions are oxides l that have become charged j J (arrow). Precipitates are ) l also seen at the 7 rain boundaries. h ( 1.'VQQ k t ..a k., .t I ',3,{h', I 4 i I$ ' i 0.05 mm f 4 .r Fig. 16. Photograph showing slip (arrow).

0.05 mm l-Fig. 17. Photograph showing slip (arrow). grain-boundary separation (A), and crack arrest (B). 0.05 mm i Fig. 18. Photograph showing transgranular fracture (arrow) at an isolated grain on the fracture surface.

I ( i i (a) l i i i D' I i l I i Y '( l / a l \\. .\\ * : l ~ w,-. j h,, ac MPP f j 0.01 mm {,i, i W'., _A ] _a l (b) I l Fig. 20. (a) Crack branching, grain-boundary separation, and precipitates are shown (arrow). (1) Enlargernent of (a). Dark circular areas, i vhere precipitates har been (arrow), are seen. X-ray fluorescence j was performed on precipitates bright area labeIIed A and the bulk j I grains. i r e

%;1

gl;%
4

- ~. 4 !, 3 L ,a 10 mm Fig. 21. Cross-sectional view of weld RBlu-10. The area', punched and used for T'M are nun.be red. a

s 4 [' M 9 1 y . ud x5 s Fig. 22. Bright-field (a) and dark-field (b) 'I Di o f a g r a i n b o un da ry of reg.on 1 of Fir. 21 A grain boundat y with pre- . I p i t a t.. <, (bright areas) is de:vn in (b),

a i I i l l ) 4 4 f-j i i d g i n~ t s. t e i I i (b) l Fig. 23. Dark-field (a) and bright-field (b) TDI of a grain boundary i (arrow) of region 3 cf Fig. 21. Precipitates (b righ t areas in the grain boundary, (a)) are seen. 4

~. _ i p- \\ l 'I ( l_ l I \\ 'A, 1 g ,- n,a. ';,.. y a e 6'. ' $.y l i I i 1 a I l l l 1 i l i (0) J l ~ \\ I \\ i ~ s s /

j. '

) l 9 I ~ El Fig. 24. Bright-field (a) and dark-field (b) TDi of a grain boundary ) (arrow) of region 4 of Fig. 21.

l X i x I l 2 1 -{ j m z N W -i l _-h Cr Cr Fe Feg Ni a g a Si X-RAY ENERGY (kev) the fracture surface. Si, X-ray fluorescence yield of the grains atThe relative magnitude et the Fig. 25. Cr, Fe, and Ni peaks are obsersed. peaks shows that the material is stainless steel. b.

l I I I I I X l I x _.z i -4 l t m l l Z i us d L l I l 1 ~~~~ -< 1 Feg Ni Ni l Cra Crg Fe Sa Sg a a g 4 I Fig. 26. X-ray fluorescence yield of a precipitate embedded in a grain i boundary at the fracture surface. S. Cr, Fe, and Ni peaks are 'f j observed. The relative magnitude of the peaks (compare with Fig. 25) suggests that the particle is made of a chromium-i rich spinel. i i f I i i .. i

L

  1. M E l E E E III O.O 0,1 0.2 0.3 0.4 0.5 0.6 0.7 FR0 Q

FUSION UNE (INCHES) l Fig. 27. Sensitization test of an area that includes the HAZ. The samples were innersed in oxalic acid for a given length of time so that grain-boundary etching could occur. The veld fusion line is shown in (a), and each succeeding photograph [(b)-(h)] is [ of an area 0.1-in. farther from the weld. The heaviest sensitization is seen at a I distance of 0.4-0.5 in, from the veld. The horizontal lines are drawing lines pro-i i duced during pipe f abrication. r I t i

e l 2.0 I o m 2 i v >x -m z r o h ip i CD -1 d O g C m z 1'0 MAXIMUM GRAIN-O CRACK l LOCATION BOUNDARY ATTACK $r<p j _< C r' i m u -4 r -i o,1 n x i l 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 t I DISTANCE FROM FUSION LINE (in.) l j i i Sensiti-Relative sensitization vs distance from the fusion line. [ I i Fig. 28 ration is obtained by using the QUANTIMET scheme of measuring the l (dark area / light area) of etched grain boundary from relative area l Fig. 27. The maximum in sensitization occurs at 0.5 in., whereas l the crack occurs 0.3 in. from the veld fusion line. I 4

It ~k8 m l1 j

  1. lt j a-a A'O E

O.340 s E ~ a g N g a-A-A'A-A a O'320 f A Ag j a 3 3 _o 7z O.300 l m A V) U) C O.280 z 10 0 200 300 ~ O AN GULA R POSITION (DEG) The circu:2-Circumferential wall thickness vs angular position. ferential vall-thickness measurement was accomplished for veld Fig. 29. Tne RBAJ-7 as shown in the upper right corner of the figure. } data were obtained 1/8 in. fron. the weld root.

O N 'p5 '2 COLD SPRING 1 8, 7 4 / LOOP A @k @ i "/COLD SPRING @ /j2 LOO P B Fig. 30. Schematic of cold-spring measurements.

.i l l 1 i l rh "iff;r *. 7 'f I' ,)/ l \\ .r (s. .9 s..u.$ . 5,..,. s, . ' sg w, f 1 l t ..c..- i l r,' ' ' t

  • 't

' g I. ll.. VJ I. ~; 4 1 WELD re - f ~. l>.f.y . /., l. ,L.' DIRECTION 's j: - g4 ; 's-( -+- I..* 'j .p < .'.~ l .? ?, t *4 ' 'I '; t 1 ' r ' l :;j', .. s. ~ ' J.ht , '7l 0.5mm j f . f.. g l r Fi g. 31. The ID surface of RBBJ-10. The machining marks (A) and fracture surface (B) are seen. The crack probably originated in a region i where machine marks exist; how-ever, the crack edge (arrow) in this figure is outside the machined region. -l g I i e~w m -. - ~ -. ,nn-----,en.,, - -, - - - --n..v,--a

',x.,0,5 mm, '.j. ,/ y. ,3.... .~ ',. { ? p s.. ' ll II. t. b' f ' ',.': t,. fi ? ? r .h. ' [ " $ / 1, f

y. p j&-

f,.',:ra t,x, fhf.'jy n '

Q k A, j,

,( r idliME w(lL. g .g. h(j [s -, f., - h'hE ' (a) r'.' } kI' 'i (($(,(, 6- < r1 M.,.'!?K.'&, ,.y ,t,' .s d* 4;, 't l@if $1, o,'. Y. a 'fg *; * .): ), t-y h: h4 (,. ss . ( b' 7 [;;;((f)(:/g s O. I mm jgi 3.:.eaAet. a (b) Fig. 32. (a) The machining marks (A) and defect at the weld root are shown (arrow). The weld is at t.he right. Solidification marks (B) are { also shown. (b) The defect at the weld root is shown magnified. The weld is at the right side of the figure. Precipitates at the 4 grain boundaries are shown (arrow). The def ect follows the weld root and is both transgranular and intergranular. _}}

Retrieved from "https://kanterella.com/w/index.php?title=ML20091A898&oldid=4338041"