ML19262A048

Revised Deficiency Rept 78-03 on Reactor Containment Liner Overlay Pad Fillet Weld Crack.Found 1-ft Crack Along Toe of Bottom Horizontal Fillet Weld.Cracked Area of Liner Plate Removed.Will Be Replaced Using Approved Procedures

ML19262A048
Person / Time
Site:	Beaver Valley
Issue date:	10/19/1979
From:	STONE & WEBSTER ENGINEERING CORP.
To:
Shared Package
ML19262A044	List: ML19262A043 ML19262A048
References
NUDOCS 7910260190
Download: ML19262A048 (30)

Text

I .

FINAL REPORT ON REACTOR CONTAINMENT LINER OVERLAY PAD FILLET WELD d.T.

BEAVER VALLEY POWER STATION - IMIT NO. 2 I

I COPYRIGHT 1979 1215 '.53 STONE & WEBSTER ENGINEERING CORPORATION BOSTON, MASSACHUSETTC g 7 9102 60**f O

FINAL REPORT ON REACTOR CONTAINMENT LINER OVERLAY PAD FILLET WELD AT BEAVER VALLEY POWER STATION - UNIT HO. 2 1.O _ SUM?u.tY A surface examination of the reactor containment liner plate and associated equipment, following a sandblast operation necessary for prime paint substrate, revealed a one foot long crack along the toe of a bottom horizontal fillet weld which attaches an overlay pad to the reactor containment vertical I liner plate.

7198-4",

centerline.

The overlay pad is located at elevation fourteen feet west of the north containment As a result of grinding necessary for repair, the crack depth was found to be greater than half the thickness of the liner material. Further investigation has established that the crack did not penetrate the containment boundary and that the crack is unique and restricted to this single overlay pad.

2.0 IMMEDIATE ACTION TAKEN Since the crack continued beyond the liner contractor's maximum allowed excavation depth, half the thickness of the liner material, the defect was viewed as a possible through-thickness crack, thus c potential breach of the containment vapor boundary. Grinding was stopped and a detailed evaluation iniciated. No additional overlay plates have been installed since the discovery of the crack.

Since this problem was considered a possible reportable significant deficiency pursuant to 10CFR50.55 (e) (1) , the U.S.

Nuclear Regulatory Commission (NRC) was notified orally of this situation on December 22, 1978.

I

3.0 DESCRIPTION

OF THE DEFICIEN Q I The overlay pad was fit up on December 15, 1977, tacked in place on December 19, 1977, and final welded on December 21, 1977. Visual and magnetic particle examinations were performed on January 16, 1978 and February 15, 1978, I respectively, with acceptable resuLts. 'Ite sandblast operation and resultant visual observation of the crack was performed on October 23, 1978.

The Contractor *s general procedures allow for repair of base metal linear defects which do not exceed 1/2 t (t is base metal thickness) in deptb <x 1/2 t in width. Linear defects 1215 154 1.

which exceed the above require the Contractor's Project Engineer to be notified.

Based on this information, Nonconformance and Disposition I (NSD) Report No. 9167 was issued by the site personnel.

Engineers reviewed the problems and dispositioned NSD No. 9167. Further grinding invoked by NSD No. 9167 showed The I the crack to approximate a through-liner crack, thus a potential breach of the containment vapor tight boundary.

Results of the investigation have shown that the crack was not a through crack.

A detailed description of the metallurgical examination of this problem is presented as Enclosure 1. A report by Dr.

C.M. Adams, an independent consultant retained to investigate this problem, is presented as Enclosure 2.

It is concluded that the defect is a hydrogen - related delayed crack which initiated at the toe of the fillet weld.

Propagation into the base material is explained by observed material properties.

4.0 ANALYSIS OF SAFETY IMPLICATIONS The safety of offsite personnel and the function of the containment liner to serve as a leak tight boundary are assured for the following reasons:

A. The area of liner plate containing the defect has been removed and will be replaced using controlled repair techniques.

B. The remainder of the overlay plates were reexamined I using visual and magnetic particle techniques and found to be free of defects.

C. Between initial acceptance and reexamination of all I overlay pads, hydrogen, an essential contributor to delayed material.

cracking, has naturally diffused from the D. Properties of the case material are such that it can readily perform its intended service function (Enclosure 3).

5.0 CORRECTIVE ACTION 'IO REMEDY DEFICIENCY I The cracked area of the liner plate has been removed and will be replaced using approved repair procedures. All presently installed overlay plates have received additional visual and I magnetic particle examinations to assure that welds are free of similar defects.

\5,3

)2\5 I 2-I

, j -

, e S

4 1

D I

4

, ' f '

O e

I e .

METALLURGICAL EXAMINATION OF REACTOR CONTAINMENT LIN'N OVERLAY PAD FILLET WELD CRACK AT BEAVER VALLEY POWER STATION - UNIT NO. 2 by P.A.G. Carbonaro February 21, 1979 APPROVED BY g ,s / m v R'.A. Rosenberg I

I I .

I I

I I

I STONE & WEBSTER ENGINEERING CORPORATION LOSTON MASSACHUSETTS

',51

\2\S

I I

ACKNOWLEDGEMENT The author acknowledges the efforts, 4.dvic e, and technical I support of 'tessrs . H. Aznoian, J. Dowicki, C.W. Eriksson, F. Morales, I. Sprung, P.W. Ward, and E.G. Watters during the conduct of this investigation.

I I

I I

I I

I I

I I 1215 158 I

I 3

SUMMARY

An overlay pad-to-containment liner fillet weld toe crack was subjected to a metallurgical study, which included a review of welding procedure s, environmental factors, and evaluative metallurgical tests, i.e., volumetric analysis, magnetic particle I testing, macroexamination, metallography, scanning, electron microscopy, chemical analyses, and other destructive tests.

Factors contributing to the f ailure are discussed with initiation I attributable to hydrogen delayed cracking. The report concludes that the presence of cracks is not generic, is limited to a local repairable area, and that the A537 Grade B liner material is satisfactory. Finally, special precautions are recommended to minimize the hydrogen problem during repair of affected concrete backed liner plate.

I I

I I

I I

I I

l \2\5 159

I I 4 TABLE OF CONTENTS Subiect Page

1.0 INTRODUCTION

7

2.0 CONCLUSION

S 7 3.0 RECOMMENDATIONS 7

4.0 TECHNICAL ANALYSIS

8 4.1 Review of Procedures and Environment 8 I 4.2 4.3 4.4 In Situ Visual Examination Sample Removal Nondestructive Examination 8

8 9

4.4.1 Visual Examination 9 4.4.2 Magnetic Particle Examination 9 4.4.3 Radiographic Examination 9 4.4.4 Ultrasonic Testing 10 4.5 Destructive Testing 12 4.5.1 Chemical Analysis 12 4.5.2 Macroetch Examination 13 4.5.3 Hardness 13 4.5.4 Mechanical Test 14 4.6 Scanning Electron Microscopy 15 4.7 Metallographic Examination 15 4.8 Technical Discussion 16 I

I E

1215 160 I

i

] 5

~

LIST OF FIGURES

=

, Fig . No. Subiect 1 Overlay Pad Location 2 Location of Test Sample

_~

3 Sample Removal i 4 View of Sample (Back Side)

5 View of Sample (Weld Side) 6 Palpable Offset Opposite Weld 7 Photoradiograph of Fillet Weld Area a

8 Test Sample Sectioned 3 9 Mocroetched Stud / Liner / Weld / Pad Junction 3

9 View of Crack Mag 12x 10 Crack Surface

] 11 Hardness Surveys 12 Plastic Deformation by Bending 13 Scanning Electron Photomicrographs

-, 14 Scanning Electron Photomicrographs 15 Photomicrographs (Metallographic) 4 J

1 j

J 1

] \2\S \bi

I 6

LIST OF TABLES Table No. Sub-lect I Chemical Analyses II Tensile Tests I

I

\2\b \62 I

i j

i

)

1.0 INTRODUCTION

Following sand blasting preparatory to painting in October

1978 a crack was visually detected at the toe of a. fillet j weld attaching a 12 in x 12 in x 1/2 in overlay pad (No. 167A10) onto the 3/8 in thick containment liner. It

, was probe ground along its entire length, and when further i grinding over a 3 in length showed penetration beyond one half thickness of the liner further grin 6ing was stopped.

_ The crack ran continuously along the entire bottom fillet length and became intermittent as it rounded the corner for about one in. An engineering investigation was initiated to determine the cause of cracking and to establish the

necessary corrective action. This taetallurgical study is part of the investigation.

2.0 CONCLUSION

S

1. The most probable cause of crack initiation was hydrogen embrittlement.

2. The cracking problem is isolated to the local area examined.

3. The A537 Grade B liner plate material is satisfactory.

4. Metallurgically required repair welds for the concrete backed liner replacement can be made successfully.

3.0 RECOMMENDATIONS

1. Repair procedures for the concrete backed liner should include the following precautions to prevent hydrogen

embrittlement

^

a. Preheating to a range of 250-350*F for at least thirty minutes af ter all evidences of moisture

have been removed prior to welding and to weld while holding materials within this temperature range. Post weld baking should be in a range of

250-350*F for a minimum of four hours; however, j longer periods may be used for added assurance.

b. Electrode holding ovens be maintained at a

! mimimum temperature of 2500F and electrode exposure to ambient temperature does not exceed

_ two hours.

I i

e

2 8

4.0 TECHNICAL ANALYSIS

4.1 Review of Procedures and Environment

_ The technical analysis included a site visitation prior to the removal of the sample for metallurigical study,

=

a review of the welding procedure used, and a review of the environment data at the time of welding of the overlay pad. Welding procedure required a preheat to 50*F minimum prior to welding. There were no known or apparent deviations from the established welding e procedures which included weld rod control. Prior to final welding, there had been periods of rain including light rain on December 20. The day of final welding, December 21, 1977, was cloudy, no rain, and ambient temperature of 300F.

4.2 In Situ Visual Examination Examination of the defect in situ disclosed a crack i along the lower portion of the overlay pad fillet weld.

! The jagged appearing crack opening was estimated at approximately .015 to .025 in at a location of minimum

probe grinding. The crack appeared to be very tight at i

the area where probe grinding had proceeded beyond the mid-thickness of the liner. No other defect, such as gross inclusions, laps, seams, etc, were observed. The

weld remaining and welds on other overlay pads in the J immediate area displayed good workmanship.

] 4.3 Sample Removal A section of metal, 6 in x 18 in containing the entire crack was removed from the liner and included approximately a 3 in portion of the pad. Figure 1 shows the location of the pad; Figure 2 shows the segment removed for metallurgical study, and Figure 3 contains a series of photographs taken during the sample removal. The removal procedure required locating the headed concrete anchor studs by ultrasonic techniques, hole-cutting around stads, and removal of

~

the sample section by cutting dry with thin grinding wheels. Views of the sample as received at Stone 6 Webster, Boston, Massachusetts, are shown in Figures 4 and 5. Figure 4 is from the back side of the liner 2

showing the two studs that were holed out, and Figure 5

~

is a view of the overlay pad weld area. Due to the geometry of the sample, which knpeded ultrasonic

_ detection of the stud, several hole cuts were required in the overlay plate to locate one of the studs n precisely. Care was exercised so as not to destroy evidence or aggravate the crack during the sample removal operation.

\l\

_ =

6 9

4.4 Nondestructive Examination 4.4.1 Visual Examination A minor distortion in the liner on the reverse side of the liner opposite the weld deposit was noteo, the depression measuring .006 in (see

, Figure 6) . This deviation is attributable to the plastic deformation caused by the welding shrinkage. At the lower left-hand corner of the pad fillet weld there was a stud attached to the

_ back of the liner, forming stud / liner / weld metal / pad / junction, all in line in the crack path. A cross sectienal slice of metal through s this junction was used for further evaluative studies which are discussed later in this s report. The overlay plate was in direct contact l with the liner plate over most of the areas cross sectioned. Various cross sections were polished and etched and showed six weld passes of uniform size and adequate penetration.

4.4.2 Magnetic Particle Examination All surfaces of the test sample were examined by magnetic particle (MP) testing prior to

, sectioning for evaluation. MP of the ground surface delineated the crack as a continuous a crack along the length of the fillet weld, as previously observed, with some apparent discontinuity as it rounded the corners. MP examination of the back side of the liner disclosed no emergence of the crack through the y steel at any location, indicating that the crack has penetrated into the liner but not completely through its thickness. No other defects were

' uncovered.

4.4.3 Radiographic Examination Several radiographs were required before the crack could be clearly delineated. A photograph depicting radiograph results is shown in

, Figure 7. Other than the crack defect under study, no other defects were observed by the radiographic examination. Radiographic examination was for exploratory purposes and not performed in conformance with any code requirements.

-l .

)2\b

a

=

10 i 4.4.4 Ultrasonic Testing m

Ultrasonic testing BME), using techiques which

• exceeded code procedural requirements, was

~

applied for crack evaluation and to determine base metal characteristics:

a. Crack Evaluation

1. Due to partial excavation of the crack along its entire length prior to a removal of the samples, meaningful UT could not be performed from the welded side of the section; therefore, all

=

~

tests were conducted from the opposite surface of the liner plate. An angle beam scan was performed using a maniature size (8 by 9 mm) 4 MHz 70 deg probe, and the ultrasonic instrument calibrated for direct n readout of depth of indication.

Results show the crack decayed at an approximate depth of .150 in from the probe contact surface, (i .e . , the back a surface of the liner plate). The response decay point was partially obscured by the initial pulse and probe interface indications, so additional tests were performed using pitch / catch, straight beam techniques to verify the angle beam readings.

~~

2. A 5 MHz, 1/4 in dia dual element probe

, (pitch and catch) was used to scan the crack. The response indicated that i

the crack, which had a general through wall direction, was also comprised of

=

steps parallel to the surface and at

various depths. The crack tip in many locations appeared to be a laminar type step. The crack generally appeared to stop approximately .120 in to .150 in from the liner plate back surface, confirming the angle beam test results and earlier M.P. test results.

i During the crack evaluation using this

] technique, a scan was also made to determine the integrity of the fillet

, weld - heat affected zone along liner j plate extending into the weld root.

Results showed no deficiencies.

] 1215 '.66

I I ii

b. Base Material Characteristics

1. The liner plate material of the sample was straight beam examined using a I 12 MHz 5 MM dia probe to detect possible laminar conditions and to determine whether base metal surface was consistent throughout. A number of multiple echoes from the plate were displaced on the instrument screen I forming an attenuatic curve.

scanning, any deviation normal pattern was from When the further investigated to determine whether I surface conditions contributing cause.

were Results showed the uniformity of the response curve with the exception of two localized areas, away from the crack, each approximately 1 in in diameter. A I scan was made of these areas using the dual p:c-Se at a very high sensitivity level. Resulta showed extremely small reflections within the plate at these I locations. One area was sectioned, polished, and examined under 10X-30X magnification. This material did show I some in-line inclusions.

porosity The material would be acceptable if examined by ultrasonic and/or standards invoked by ASME codes.

2. Straight beam pitch and catch examination was performed on the I remainder of the liner plate sample as a follow-up of the 12 MHz straight beam tests. Results showed occasional I small reflectors and verified earlier results.

in general I 3. Angle beam testing was performed of the liner plate sample using a 45 deg 4 MHz 8 x 9 MM probe. The sensitivity used far exceeds the three percent notch response required during ASME code examination. Results showed occasional individual reflections with no linear characteristics. The plate I. would be acceptable if examined to ASME code requirements.

In summary, the UT has shown the crack generally from .120 in to .150 in from I a2\s \61

12 the back side of the liner - again with no penetration at any point. The indications that the crack had I lamellar-type propagation associated with it was verified in subsequent metallographic study. Occasionally, line microporosity and inclusions were detected in the material. Itowever ,

from a steel soundt. ass point view, the liner material would be classified I acceptable requirements.

under existing code 4.5 Destructive Testing Following nondestructive examination, the 6 in x 18 in sample cut from the liner plate was sectioned for I destructive testing and study.

extent and type of tests taken.

Figure 8 shows the Except for one chemical analysis and hardness survey in the weld metal I area of the overlay pad, all destructive tests involved the liner material as its properties and behavior were considered of prime importance to the study.

4.5.1 Chemical Analysis Chemical analyses were made of the liner, weld deposit, and overlay pad materials, The elements carbon and sulfur were analyzed by the combustion method using a Leco automatic

.70 second carbon analyzer, a Leco indiction furnace, and a Leco automatic titration apparatus. The elements manganese, phosphorus, silicon, chromium, nickel, molybdenum, tin, and I vanadium were analyzed by wet chemical methods in accordance with ASTM E30 and Antimony, bismuth, and lead were analyzed by the E 350.

I atomic absorption method using a Jarell Ash AA Spectrophotometer and manufacturer recommended instrument settings. For nitrogen, the micro Kjeldahl Titration method was used. The results I are shown in Table 1.

A comparison is t.ade with the code-acceptable I analyses for the materials and certification covering the same materials. In the addition, elements not normally specified were mill I determined as these elements, if in sufficient quantity, may affect ductility properties of the steel.

and impact All base materials analyzed conformed with the code-required analyses, with the exception of 0.07 percent higher manganese in the weld metal deposit.

I 1215 168

i 1

4.5.2 Macroetch Examination At various locations the segmented specimen was examined after lightly etching polished areas

• 1i perpendicular to the crack. The polished surface showed the weld beads to be of uniform size and having approximately 1/32 in penetration into the base metal. The corresponding heat affected zones showed adequate but not excessive heat input. These were uniform at all sections that were polished 1.

and examined. Figure 9 shows one such area, which contained the junction of stud that had been welded to the back of the liner during construction, the liner material containing the

[ crack, the weld metal deposit remaining after i probe grinding in the field and the overlay pad material. The crack, as evidenced by the above

,1- photographs, progressed in a stepwise fashion into the liner material. The liner material 3 apparently has behaved in a ductile manner thus j limiting propogation of the crack. By cutting a

number of samples, perpendicular to the crack and fracturing these samples after embrittling the material by immersing in liquid nitrogen, it s~' ,l was possible to examine a considerable length of the crack surface. (see Figure 10) All of the samples opened in this manner displayed the same 1 woody appearance with the cracked surface being

heavily corroded, indicating long exposure to the atmosphere, although the length of time of

_1_

exposure could not be determined quantitatively.

4.5.3 Hardness

=

Hardness surveys were made using Rockwell "B" I

scale tests which encompassed the liner, weld metal, and overlay pad shown in Figure 11. In 1 addition, a microhardness survey of weld zones i at the junction of the stud / liner / weld metal and a overlay pad was made and is also shown in Figure 11.

The Rockwell "B" survey gave hardness readings 1 varying tror, $(.8 to 99.8 and averaged 96.6 at this leve3 ON wnverted ultimate strength would be about fut asi.

Vicket" ik.vgahardness surveys using a 500 oram load care taken at the stud / liner and

~

metal / overlay pad junction and showed more changes as the indentations proceeded through the weld and heat-affected zones. The liner JI 1215 169

I 14 I base metals directly above the heat-affected zone in the stud area showed lower hardness.

Around the crack itself, hardnesses were consistent, averaging about 97.6 RB (converted) .

I At the fillet weld higher hardnesses observed than in the base metal.

were 4.5.4 Mechanical Tests Mechanical tests of the liner material taken I transverse to the rolling direction, included tensile specimens, as well as selected bend specimens to ascertain the behavior of the material under extreme loading conditions.

Results from tensile testing are shown in Tahle II. Based on three tensile tests the liner material has an ultimate strength of I 97.7 kai and a yield strength of 86.1 ksi.

Elongation averaged 23 percent. Specification required an ultimate strength of 80-100 ksi, a minimum yield strength of 60 ksi and a minimum I elongation of 22 percent.

Several tests were made to observe the behavior I of the plastic

. material under conditions of severe deformation. A total of 4 bend specimens were machined and bent over a mandrel I having a diameter of 2 t (where t is equal to the specimen thickness).

Figure 12 shows these specimens after bending.

I Two specimens were taken in the long transverse direction (Figures 12a and 12d) with one specimen having the as received surface (12d)

I and the other (12a) having the as received surf ace machined off .010 in prior to bending.

This latter specimen was bent 180 deg but I displayed fine surface cracks at the tension side. Sample 12d failed prematurely after bending approximately 65 deg, with cracks bend initiating at surface defects (pock marks) on the tension side. Two other specimens (Figure 12c) in the direction of rolling with no surface preparation and the other (Figure 12b)

I bent parallel to the short transverse axis, were successfully bent 180 deg.

I I

l 1215 170

I 15 4.6 Scanning Electron Microscopy Fractura surfaces from (1) the original crack, (2) the failed bend specimen, (3) the tensile test specimen, I and (4) laboratory fractured specimens were subject to electron microscopy study.1 The original crack was heavily oxidized and was cleaned using a phosphoric I acid-and-water mixture. It was then cleaned in acetone using an ultrasonic bath, followed by preparation for electron microscopy. All other specimens were examined I in the "as-fractured" condition without surface preparation, except that required to enhance electron microscopy. Specimens used for limited the microprobe chemical determination received no I preparation. A comparision was made of the phosphoric acid cleaned surfaces against the nonphosphoric acid cleaned surfaces and, essentially, the rust-removal I operation did not show any adverse effect on the fracture surface. Figures 13 and 14 show a series of photographs obtained during this examination.

The original cracked surface showed no striations of the type which would indicate fatigue-type failure.

The fracture exhibited a fibrous " woody" appearance, suggesting that the base material was heavily banded or contained a large number of stringer inclusions. For comparison purposes, tensile and bend test fracture surfaces were examined, and these exhibited a fibrous appearance containing shear dimples, indicating ductile failures. Inclusions were observed in all fractured surfaces and a microanalytical scan of one of the I induced fractures indicated dispersions of inclusions high in manganese content.

4.7 Metallographic Examination Metallographic examinations were conducted at several I locations to study (a) base material, (b) crack geometric appearance, and (c) selected indications from the ultrasonic test. Photo micrographs are contained in Figure 15. The liner base material displayed a I fine-tempered martensitic structure typical for the quench and tempered heat treatment given A537 Grade B steel. A segregated structure of nearly parallel bands I aligned in the direction of rolling pronounced, and inclusions both stringer-type (banding) was globular appeared at a higher frequency than normally and I encountered in average quality hot-rolled steel of this grade. Banding, a condition normal to such plate

• Massachusetts Materials Research Inc. Letter Report No. F153-22 dated February 6, 1979 g 1215 171

I 16 I materials, was aggravated somewhat by the inclusion content. Propagation followed planes of weakness caused by the banding and cut across metal laminates in a ductile fashion, creating the step-like appearance.

Selected indications from the ultrasonic testing were identified as porosity or inclusions.

4.8 Technical Discussion I The palpable offset of the liner opposite the weld indicates that weld shrinkage stress exceeded the yield strength of the material to cause deformation. The liner material displayed the necessary ductility to I absorb this stress without cracking, a behavior considered normal for the material.

I various tests and examinations showed that the crack propogated into the liner in a step like fashion. When opened, the crack displayed a " woody" appearance I suggesting a lamellar condition and possibily lamellar tearing for cause of defect. However, lamellar tearing is ruled out as the reason for crack initiation because

a. The thickness of the liner material does not make it a prime candidate for lamellar tearing, a condition normally associated with thick plate material.

b. When lamellar tearing is observed in thin plate I it occurs under conditions of severe welding restraints, a situation not present during the containment fabrication and.

Extensive ultrasonic examination revealed the character of the crack and extent of its propagation. No evidence of subsurface lamellar tearing was detected in I the liner away from the crack in the weld area.

Overall, no defects considered rejectable by ASME Code requirements were found, other than the crack.

Chemical analyses disclosed that, for all intents and purposes, all materials, overlay pad, weld metal deposit, and liner met specification requirements. The I weld deposit had exceeded the weld rod requirements for manganese by 0.07 percent an< this inconsequential increased was due to weld metal dilution during I trelding. In addition, no trace elements were found in undesirable amounts.

Hardness surveys disclosed good uniformity of the I materials witL the weld metal slightly higher than the overlay pad or liner, a normal relationship. As might l 1215 172

I 17 I be expected, variation in hardness was observed in the heat effected zones of the weld area.

hardness was uniform in the immediate crack area.

However, Tensile properties were all within code requirements, with elongation being low but within acceptable limits, and yield strength and ultimate strength well above the I minimum code requirements. Of the four bend tests taken at different orientations, the transverse tests failed, one bend specimen whose skin was removed I achieved a 180 deg bend with slight surface cracks developing during the tests; and the specimen with no surface preparation, achieved a 65 deg bend prior to cracks initiation and premature failure. Specimens I taken in successfully.

the longitudinal direction passed I Finally, metallographic examination and electron microscopy disclosed the anticipated structure for the heat treatment given, sulphide and silicon inclusion I within normal limits, and ductile behavior during failure.

Because of the loss of evidence containing material, removed during the initial repair attempt at the site, it was not possible to establish the exact cause of crack initiation. Base metal properties would not in I themselves provide a basis for crack initiation. In fact, the material displayed a level of ductility sufficient to prevent through liner penetration.

Based on the results of the metallurgical investigation:

1. The most probable cause for crack initiation was hydrogen embrittlement because:

I a. The crack was not detected two months after welding during the final MP inspection but was found some eight months later, indicating a delay failure.

b. Base metal properties do not in themselves provide a basis for crack initiation. In fact, ductile behavior was demonstrated.

c. Moisture was available during welding in I the immediate environment, i.e., moisture being trapped between the tack welded overlay pad and liner. The preheat requirement of 500F would be insufficient I to insure moisture removal prior to and during final welding.

1215 173 I

I .

18 I d. The observed crack initiation was located at the toe of the weld.

e. The chemistry of the liner makes it I susceptible to hydrogen embrittlement and

f. Stresses imposed by welding were present.

2. The cracking problem is isolated to this local area because.

a. No similar cracks i.e., hydrogen delayed cracks, or other defects were detected after magnetic particle reexamination of other pads.

b. Natural hydrogen dif fusion has removed the threat of delayed weld cracking

3. The liner plate material is satisfactory for use because:

a. The base material is of sound quality.

Nondestructive examination, such as visual I and magnetic particle, radiography, and ultrasonic inspection has disclosed undesirable plate defects no

b. Strength properties and elongation are within specified limits

c. Deformation to a degree required for crack initiation in bend testing will not be encountered in service

4. The required repair welds can be made successfully with proper precautions to prevent hydrogen embrittlement as follows:

1. Preheating to a range of 250-3500F for at least thirty minutes after all evidence of I moisture has been removed prior to welding, and to weld while holding materials within this temperature range is suggested. Post I weld baking should be in a range 250 -3500F for a minimum of four hours. A 24 hour2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> post bake will absolutely assure of hydrogen removal.

2. Electrode holding oven should be maintained at a minimum temperature of 2500F, and electrode exposure to ambient. temperature prior to use is not to exceed two hours.

I 1215 174

I I

l 33' EL 722*-4" I 14 p OVERLAY PAD NO.167A10 EL 719'- 4" 10' EL. 716*- l"

-- l'-IO" -

13'-II k~

EL 712'-3" PLATE I.D. NO. 5782 o

~

NINSIDE OF 3/8" LINER PLATE l FIGURE i l OVERLAY PAD LOCATION I

12" SQ r m I OVER P  %

SAMPLE REMOVED FOR I METALLURGICAL EX AMIN ATION

/

[ l) '

• "#C*'

_ _~__

}l s 6,,

t 2_

I = 18" I

I FIGURE 2 LOCATION OF TEST SAMPLE 171; 175 I -

g '

.}j' . ~ g* } , .; 7;f .*

~-

l4 l , _;

..- 2 :3 ??- V ..

' ' ~

M . /, c. $- . .;., g

+,js >

f. i* ' I y. ' sI 3 . .. ..

..go.

rg .t .

_, 7.

j-

,4j j;;s . .

7,

=

>].

\

E s ,. . .

lf .. l -fh ._ ;

},h:?: .,

,e f 5^ .

ll [;,' . N

-Q, , M ; . \. ., -

3.1  ;,' - , -

l LOCATING STUJ I .a -

f . . m  ? BY HOLE CUTTiMG

<.~.% 4 ;A.. .@k ,

! (LOWER LEFT)

W ,Jye} *.?:

. p. -  ; c- .

Y,. ~ : f. ., .$y ;;$ ~ .

'. ,);

j .) <*

I n[ '

6 .. >

i IE

1. .y . - .. . ,

@+s - d3, . ' , 4.

4..'-

^

P'

• 4  :

.) Q m *,,, - . .,

t , ', t_ % . ~ ~ ,

l 4 .,, e u

I h-h p@

. }. *, ff. r n - " l

h ,

a j

l

, 46 s.t. . ;*

~

4$

t.

m.,'.9L*45'- ~ . . . . _ _ _

- _ . _ _ _ . = --

1 s.

. s, .-

mv y.. m<  ; -

?-r, 3 g b ,- -

z.

QG. c,7  :'.~4 A s ., 'T ,

J d'.

I

. l+ .

,'55 .g:-lg7*-.

_. 9 4

, .

,4 .f Q.;  :

s.. c m. ,

( . [t ; i dlt - ,

. age :17 * , lJ .

lQ h 7y  ; g,: ~ # o _ -;p (*

I

< ,1 ;), u 'he- a ,c [,1

j <

r. -

HOLE CUTTING FOR STUD p.p ea y-.. ,

y,,.

DISENGAGEMENT 4 (LOWER RIGHT)

'..y/ a I z...'

1. 1 , Q,g (WM y p~ r.,,.'"n' gg..

m ( ' "

f.O

'spe- . , ce' y I ).

!;-gy g% -

f)/ - y:nl

', ". SK c

I T ^#f*

• 6-

{.. y, nb j ,,f r=.>t.q

? ?.}

.-t~

.m 4

. 1.

'. *o>f f

1; .

~;f ?

e. "4t .W o $w%.

- .v

-s g$ fj- s,-a 4

I g.p y Lj.[4 ,;ig,7

,. *qfgd a- A w..

p.' .y-0

. .~ e >

if.& >w n a N by?)f .E - .p hk y {y

-Q i.hfy D ;

Qity

.,. s f. ' .a_

)jf

,.  :)

• 1, Q-

~ '

Y [

h1 . .

85 p .p. ., -b; 4 t*'$-

Sy ;j 11., ,

[ k '/ i

.; e OVER ALL CUT OUT AREA

- E- .eo.ug' u. 't: w ,-

,2_ , _ m~ c -A. ( . . e ? ' *

~

3 - . - - , .

g f Q g" W w 4 ,

I ,v-

?_._

l_l ;?

. :s

• ~? ?

' -n

~ %_ w. y;. -

,,, j.N ^

s ,;. ,

e '*

I e

a FIGURE 3 SAMPLE REMOVAL FROM CONTAINMENT I 1215 176

I DMD~f D MIAl

= vun o n 1 m t.

/

b4/ , g , _I ,

k 4 4

• 5 7, * , , 'S '

9,

s .. .

(*7,,

y M r' -

,,g '

(,

y=s __

y .

S+ g ~

- 5

• y

+ , 4 o. -

-6

, - ,f -

8  ; - , 4 .

,y=*

e

t. g.4 ,

,,+Q

,4g 4t s

'4

+.

g ,' , ,

e., 's'. *sg?'

• s I

. s ,* , - Q.

,,y. ~/** *

. . . y 5 ^

4 -

f .- , ~ .

n. 4. .

, (i Q e t

g. "'# g$^

• t*

. y. 3 s.

ew -. )

.'34

+ , 4' ,M+'m" ..

4

' h(- % w ', +

2

-g

^

. -- . + -.,. .

. g ,

4e , ;. ..m.. 4'. -(. ,g L 'M,y~, . 4 4

.4s

'*Q' gW'.r w> . g

- o. g *

?

y.c. eay "

+

A .

a V; y **h, A .

\

s ., ,, c v9h g.N

, . . ~*

/

q # of s> # # *** g4G h I c~

g

- -_ * <A s gW , , (/ '-~

, 3 d*-

t.

+

~ll ,(*

• t ,,: .-

o ,, . _

Tl

j. , >f g ,

Y

'y ,u a ,. # -

s. l 4

.f s.

r 3

4 . yn. I' 4

g v 1*

• g, ,

4O -

,>\

, k. &. , < ' ~N

[

px.y r

w. -

• _..

/s.

+4

~ y y .

.# y -

7 '

g ,

l.

}

p . .

,j V \

p.

o:  % .?..$

t u p.

- p.

• k g 4

\.,

5

m. 1..

.s

~3

~'

Y'+' ,

l;s_

--Q;g

e. .

. +

+ .;,

V 4 N,  ; if s'-

.p.,,f  ;

'w '

M yN. - ,.,,,.> '+' .

\ MM of go

@O* #9 se e

N,e)\ % -

0' ,

I I

I I

I WELD ZONE LINER g

' t . (

l l .oos" I

I I

I I

I I

I l FIGURE 6 PALPABLE DEPRESSION OPPOSITE WELD DEPOSIT (CROSS SECTIONAL VIEW)

\2\5 \13 I

I I

I 300ROR8NAL I .

I I

I .

y I

I I

I I

I I

I I ,,e u e s z PHOTO RADIOGRAPH OF FILLET WELD I 1215 179 I

I I

I #

g 4'l' l

CHEMICAL

(( ( DC

,c t I y' ':

L1 -

H3 y,

~. /s Hi ~

s AI 42 B

,,T I '

I

)

le" I RO,y,.

I '

I TEST NONDESTRUCTIVE TEST NG MAGNETIC PARTICLt ENTIRE SAMPLE LOCATION I X- R AY ULTRASONIC CHEMICAL ENTeRE SAMPLE ENTIRE SAMPLE A (AS INDICATED), LINER ONLY AT J I

MACRO ENTIRE SAMPLE AND ALL CROSS SECTIONS WITH CRACK METALLOG R A PHIC C,L HARDNESS C TENSILE A,N BEND M, I, A FRACTURE SURFACE Al, A2, D, E, F, H1, H 2, H 3, H 4, H 5 SCANNING FLECTRON I MICROSCOPE H1, Al, TENSILE TEST I

I I FIGURE 8 TEST SAMPLE-SECTIONED FOR METALLURGICAL TESTS (LAYOUT)

I

?

P00RORBlhM c

, .1 -.--,,,,.,m..-_,,, _c,.

72 m. .,. ,

, 3..

, + .n . f. -

p

_. , , u. >

.,^ .-

( ,

e

, m

%.c. L i .. . 4 r.., ....'.,

~.,

v, . ., , %e .- .

4.  ; . , -, ,

~

- $' ' * ". et%*

• s ,

g j?

h)5%.- ' ';$y_

a .

iIL

), ' '..'

c. 4- , 4

_=

r.,b

,s .

, , . , ; .T f '362  :-

l - '

,n 'RVr;. . . .  :

a .r.c :. . :,', p. . i,O,"

S: 1. ,

, y .

I ' ' ' h. [* ,

I p _ df'. , ,

y. . 3 ~  ;;.+, .

,,w ,

m b1 .,. a , *.% y

"'.,4 a -

. v. .n.1 . .* c,

+.: - ,

a

.$ ., g "fg 4: ..,

k -' d- . nGuwsz.w = ---' " ' . - ? J , .u ' . . $: _. __Lj MAG.3X M AG.12 X FIGURE 9 VIEWS OF CRACK AT STUD /LINLR/ WELD / PAD JUNCTION

] _

(LOCATION L - FIG.8)

M.

4

.};l,5 .

f *iWQgi?..f, -lhi or 4 . . . . , . ,-

.M b ".

4

,.,,3 , [

',..-.,7

. v. f.4- [1 3w ,:

, ' * ~ ~ .

_= '.___**-~~-- -

. FIGURE 10 CRACK SURFACE

) .

(LOCATION D - FIG. 8) 1215 181

I WELD METAL 97.8 99.6 94.8 95.2 95.8 95.8 f f 96.5

,, , , OVERLAY PAD I 97.5-

• j' j' /

4-

• g \

,_ -96.8

% LIN ER I / / /\

97.0 95.9 97.0 99.8 95.2

\ \

95.3 I ROCKWELL "B" SURVEY (LOCATION C - FIGURE 8)

I I

R8 EQUIVALENT

• R8 92.5 OVERLAY PAD - EQUlVALENT

• I 89.5 9 .O 3 2.O*

• 97.5 97.5 WELD f

joo,o 100.5 100.0 98.0 I

5**

j CRACK O.

103.5 \s7 103.0 la r - '*2 9

I 103.5 101.5 88.0

-LINER 94.0 94.0

, , s.

83.0 Y '

i I k - STUD I

• CONVERTED FROM HV500

• ROCKWELL 'C' ( CONVERTED)

MICROHARDNESS SURVEY (LOCATION L-FIGURE 8)

FIGURE 11 1215 ',82 I

P00ROR8NAL w.e:,a.

~~'

TRAN RSE ,

(-"'

w TRANSV RSE

' ' DIRECTION PARAL EL TR NSVERSE DIRECTION (B) 0 1 2 3 4 0 NG SCALE-INCHES DIRECTION (C)

FIGURE 12 SERIES OF SPECIMENS PLASTICALLY DEFORMED BY BENDING OVER A 2-THICKNESS RADIUS MANDREL 1215 183

.,...._mo-_.--_--,-.p..o...-.

P.M..Ff.il.f.il.N..Al_.

p -

. . . . , . . . . , - ~ . . . - . . - - - . - - - . . . - - -

l g g 7 ,f% g 1 hb', ..',

s 4 + ,

v.,. e

~.,

I z*7 -

J _

. ;p +

, , . . " *. se p .

~ ,, - [

f

~

• D .j (A)

+

OXIDE ON A $-

iM y u. ..WM: ;3 SURFACE e _ -- =u . k]:-_}=_

- e v .: f' ---s -

.2 g:n, iff f

, . , , . . -~ .r,7-. ._n. , . . ~

I ,

w n;3n ,

$ y

• s

). .-

a mujg y ~n.

b

?

} w:

.m,m

.g p ., ,  ;

v...

g.gp .yu i

/u;nm y

%, ,e . ., y c .

4 .

~--<

ejgpg a,s us.2. s - m . , ,

s y sE

., - ' e a,, 1 ,- '~_ _ . ..

n, ' f.w . l-

~

• gh

[.ll'I 'M<g.[. ~ . _-)2.fhb

1. w:. .-A.y -

g ,

g h 4 .df ff 7[ [, vl NM [ E ~ .

1 ; kj I l -

y &.f: , gg p ~ ,: i

,3, ...

._ g ~

y w 1 Wl.:;~ ;w / _ .

(c) (0)

FIGURE 13 (A) FRACTURE OF CRACK SURFACE (PHOSPHORIC ACID CLEANED) MAG. 200X (B) FRACTURE OF TENSILE SPECIMEN (NO CLEANING) MAG. 200X (C) BEND TEST FRACTURE ( TENSION SIDE) MAG.1700X (D) BEND TEST FRACTURE (COMPRESSION SIDE) MAG. 2100X 1215 181

g 3r fm.. A_l ll.N._ L. .

i h

gg\;%ra gyyf_-[.qyy A m%

l ,

,. w e w: :- . g.,5 u . ; .

,~

v ,~

l

- o,-

(A)

I e . Ji3m) 9..

~

~y N li3

~ M B M 3 n M A( M h , . , .. -

g.y . , a g7 , = , ,

c y, .

. a, n.  :<

s. yn; Qts.4< ; +:- 3

+9# :., , J::

"'y l

+ + -

. , F

• p3 , ~ ;r .

3 #.* q ; .e n . > .

J*n: v . s. .. . .., c:

, . i.

.ne, :#

e

",,..;.a< .

f ,: $.[y

, , N. % g - .n I s.f ' + &n

, 9. , , , ,

( *R ** p ?. .2 g' _ x "_, *f .%. g ,'.

j. .

4

, ,e; *

( j (3)

s. . r' - ~

i ;n.. ..

r1

!, p 3 - .- '. %;S, W

,? n .

't;i;(gs;I

-if,s .,

. &% MQ.:;

n.. .::. e > w . s :: .gga A4 e w ..n e:x $

wu= f*~ xx.: -~A: ff .  ?-

b

~ p " +% pe p.9 x. n#

E n,p.G.p. u % ..

',m y;; ;q4 y w,

.i.y 4) ; ; ; g^),g.

y . y.g R kga> 3 n..s .; .. s*g g

w

'f:;,g$g,hwh.,i;&,;.g...p% qh a f'j..

l .

7 3

f ;? t .

e.

y y-r y - m y e~i 9 v s.

u a

'a, l

i

.w , y.^ggt ;4z x .

• i (c) r ..

- y<.? E I

- '; ,.;:*,y.

7 l c.. _ :

~c., 7. J , a ' Ag.f- .,

n _

e .; - ; [ -f * - ~

e: ;  ? .

e 4 s. .

y . -

1. J' . ; 2.\e4 . .; .

I O: , F '.ap  %*

1 75- . "H

.g y.. . y .-

, ,..Am-,,

I I {'

e.,

, , . . . . ~ g C. , ' Q%[l! < lsf :l f, .,. ." j ./;.

ig sw y~ . , t . . , n a i

. ij/Mi'41 x - MUj.. d *l .. . V:v . g [-

FIGURE 14 (A) FRACTURE DETAIL OF BEND TEST FRACTURE-TENSION SIDE MAG 425X (B) MANGANESE X-RAY MAPPING OF AREA SHOWN IN FIGURE 14(A)

WHITE DOTS DENOTE MANGANESE RICH AREAS MAG 425X (C) A NONMETALLIC INCLUSION (MANGANESE RICH) OBSERVED IN TENSILE TEST FRACTURE SHOWN IN FIGURE 13(B) MAG 5000X 1215 185

L l . . , .

s I .

s I

~

I ,- .

PORTION OF CRACK AND INCLUSIONS UNETCHED M AG. 50X 1.,  ;

. . _ _ _ , , ..._...;. . , p , , , ,

I

, *- . - r -. ,, * .a v . . ..

., ., *, 4 ., ,i' J,<.  :.v i

A'

?. .-; . a u

..; . .r . *'

,p, .,+A.i 4$4 .

.$.,;* z o 4 I

. .. - . , , .. +  :. .

..  ?

e

.g . , < .

-e ,. . - - . 1

^

. , , ,. r ...

j :*: g t4 h

c .

w lf I ?: '+

g t

' , ,-4.

o 5

- s s.: ..

" y; -u

• y I

.- x .

.4 .

.,  !* x , ,.

, <j w, * ..a x >

s,

, ~;

8 g

, 1 * , '

, . ; . ]

.y.s ..wsr:,

g .

g,- . ; j w - - _. r a - __ _ - - :1 = . .~_ :

I NITAL ETCH CARBIDES IN A TEMPERED MARTENSITIC MATRIX MAG. 500X I FIGURE 15 I PHOTOMICROGRAPHS (LOCATION C - FIG.8) 1215 380 e

I

I I

I LINER OVERLAY PAD WELD ROD ASTM MILL' 1* 2 3

ASTM MILL SEW ASME II MILL SEU C .24 MAX. .18 .17 .19 24 MAX. .20 .12 M AX. .05 .10 Mn .6 fel.40 I.29 .65-1.40 1.23 40-1.25 1.07 1.32 I

1.33 1.28 1.15 Si 13.55 .34 .34 .33 .13 .55 .36 .32 .80 MAX. .40 .58 S .040 MAX. .022 .020 .021 .040 M AX .021 .020 030 MAX. .012 .019 P .035 MAX .008 .024 .023 .035 M A X. .008 .021 .030 MAX. 011 <.005 Ni .25 M AX. * .21 .19 .25 M AX. .17 .14 .80-1.10 1.06 1.00 Cr .25 MAX. * .19 .20 .2 *., M A X. .13 .10 .15 MAX. .01 .08 hMo .08 M AX. * .04 .06 .08 M AX. .05 .03 .35 M AX. .10 .10 I u.!

]V dCu

.3 5 M AX.

.0 01

.12

.005

.17 35 MAX.

.30

.001

.07

.05 M AX.

.02 03

.015

.07 Pb - - <.0 01 <.005 - - <.001 - - <.001 Sn - -

< 005

. <01

<.005 - -

<005 Sb - -

<.005 <.005 - -

<005

<005 Bi - -

<.00 5 <.005 - -

<.00 5 - - <.00 5 02 - -

.0027 .0023 - -

.0022 - -

.0u0 N2 - -

.014 .023 - -

.019 - - .012 I

LEGEND:

1 LADLE AN ALYSIS I 2 SAMPLES FROM SECTION A ( SEE FIGURE 8) 3 SAMPLES FROM SECTION J (SEE FIGURE 8)

• LACKING FROM MILL CERTIFICATION REPORT

• ACTUAL WELO DEPOSIT E

I TABLE I CHEMICAL AN ALYSIS I (PER CENT) 1215 187 I

I I

I I SEW ASTM MILL LOCATION A*

N*

STRENGTH KSI)

D POl 60 MIN. 77.5,74.5 86.0,86.3 85.9 K )

I ELONG ATION (%) 22 MIN. 23.0,30.0 24,24 ~22 I

I LECEND:

• ROUND .250 DI A. TENSILE SPECIMENS WERE USED IN ACCORDANCE WITH ASTM-A-370 I

I I

i l TABLE II TENSILE TESTS I OF LINER STEEL ( A537B) 1215 183 I

3; 4 g. .,

-__M,,_ . , . , , , . -.m..",.,,,,,s.y,,_____m..._3..__ _ . . _ . _ . _ m.;__, _ , . _ . s. E .

3 a

3 1

i N

g.

P S,

g .:

m

>4 I $ '

w 4 k u,. _ _

~

I I

I I Enclosure 2 I REACTOR CONTAIt01ENT LINER I OVERLAY PAD FILLET WELD CRACK AT BEAVER VALLEY POWER STATION - UNIT NO. 2 I

I I

I Prepared By h/{

Dr. C. M. Adams I

I I

I I

1215 190 E

1 I The following paragraphs sunmarize'my conclusions in connection with the causes and implications of the subject failure:

1. The location and orientation of the crack leave little doubt that the fracture initiated at the toe of the fillet weld joining the pad to the liner plate, and that the crack was hydrogen-induced .

This conclusion would be supportable even in the absence of evidence that crack initiation took I place several weeks after welding.

reinforces this finding.

The delay

2. This failure is considered unique, in no way symptomatic of some generic defectiveness, and no problem is foreseen with the other pad welds or with implementing a completely secure repair.

DISCUSSION Some two (cold) months after welding, the pad welds passed visual and magnetic particle inspection. The crack either had not initiated or was small enough to escape detection; surely, it had I not become of such dimensions as were later observed during preparation for painting. Studs were welded to the outside of the liner within a few days before or after execution of the pad poured roughly six months after I fillet weld.

welding.

Concrete was The fracture itself is distinguished in several ways:

I a. Orientation suggests initiation as a typical toe crack, as does the original report of the crack.

The exact point of initiation has been removed by grinding.

b. The transverse profi3S of the crack indicated substantial displacement, with the maximum crack opening, in excess of 0.020 in, prevailing nearest the toe of the weld. The crack opening decreases with distance into the plate, away from the toe of the weld.

I c. The fracture surfaces, as well as metallographic cross sections, revealed some lamellar tearing as a propensity mechanism for crack for propogation. However, it is most unlikely this crack initiated as a lamellar tear, especially in light of the aforementioned displacement, the geometry of which forces the conclusion the crack I propagated from the toe inward. Moreover, lamellar tearing is very rarely encountered near single fillet lap welds in p10te less than one inch thick.

Finally, exhaustive ultrasonic testing of the plate

',9F 1215

material beneath the fillet weld revealed no indications of isolated lamellar tears. All tearing was closely associated with and functioned to extend initial toe cracking. Weld areas adjacent to each end of the crack had no lamellar tears.

d. The fracture surface indicated transverse bend ductility was low.

e. There was no fractrographic evidence of fatigue.

f. The fracture was not of the brittle cleavage type e1. associated with fast propagation below the

, transition temperature.

The mechanical properties of the liner plate, at least in the region of the crack, were marginal, and differed from the certified mill test report, indicating these marginal properties I were localized (i.e., did not prevail throughout the plate) .

the region of the crack, the ultimate tensile and yield strengths In

,j were about 10 ksi higher than the mill test report values (the U.T.S. close to the ASTM maximum, 100 xsi, for A537B) , and the 1 elongation right at the minimum, 22 percent, all values determined by testing in the long transverse direction. The

.g -

transverse bend tests failed in the marginal region, but passed j in the mill test.

The steel is quite clean, exhibiting normal microstructure.

Local property variations are considered due to microsegregation 1 of alloying elements, principally manganese, together with some variation in mill heat treatment, in that some portions of the plate were rendered harder and stronger than would have been 1 preferred, although within specified ranges. In view of the generally high metallurgical quality of the material studied, it is considered most unlikely local variations would include regions which fall outside specifications in any mechanical property except bend ductility.

The slightly low local transverse ductility is not considered to 1 have played a significant part in crack initiation. The relatively high local strength did increase the susceptibility to

_ hydrogen embrittlement. Once initiated as a hydrogen-induced

=

crack, subsequent crack growth was doubtless exacerbated by the limited transverse ductility.

Although fixing the blame positively on hydrogen as an initiating cause may seem unduly speculative, since the locus of crack initiation was destroyed by grinding, field experience witt A537B supports this conclusion. Much of the support is indirect: even I with low bend ductility, this material is so tough and plastic i

that a classic toe crack almost defies any explanation other than hydrogen. And, where toe cracking has been a problem in other

\2\b 192

3 instances, with steels like A537B, instituting programs which positively eliminate hydrogen, notably a post-weld bake, has I constituted an absolute fix. Thus, although hydrogen difficult positively to detect or measure, the circumstantial evidence is compelling. The delay in cracking, if real, would, is in and of itself, leave no doubt. Absent fatigue loading, there is no probable way to initiate a delayed toe crack, except by hydrogen induction.

In the context of the stresses and strains expected in service, the mechanical properties of this liner plate are entirely satisfactory. Only if severe plastic bending or other I deformation were contemplated would the limited short transverse ductility be grounds for con e',rn . By now any hydrogen has diffused from the steel, and a post wald bake will ensure this also for the repair; therefore, tha hazard of toe cracking hac been eliminated. By the same argument, the other pad welds and other liner plates can be given a clean bill of health.

The conclusion is this pad weld was unique in that three adverse circumstances prevailed simultaneously:

a. Hydrogen in the form of moisture entered the arc atmosphere, either from the electrode coating, which would permit an inference of improper electrode handling, for which there is nc evidence, or from the interface between the pad plate and che liner, which seems more likely, since substantial preheat was not required, the ambient temperature I was quite low, and rain had occurred the day before welding.

b. Location of the studs on the outside of the liner offered somewhat more than normal restraint to loads in the case of this particular weld.

c. Transverse ductility in the innediate region of the weld was low.

REPAIK OF CONTAINMENT There is no reason to doubt that repair welds against the concrete can te successfully and permanently executed, provided that all steps are taken to prevent hydrogen embrittlemqnt of the heat-affected zones of the repair welds.

Specifically:

1. Preheat in the range 2500F to 3500F should be sustained for at least 30 minutes prior to and during welding.

g 1215 193

3 4

l

2. Postweld baking at 2500F to 3500F for 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> will absolutely ensure removal of any hydrogen.

5 3. Electrodes should be transferred directly from their hermetically sealed containers to a holding

. oven at 2500F, and should suffer no more than two hours total exposure to the ambient atmosphere between removal from the oven and welding.

e Dr. C. M. Adams m

N M

=

b

=

J m

5 i

2

',4h

\2\b 2

y k.

=

i-w.

e . v g-  %>

• E

-. . ... . _ . . - . _ .. . _ , __ - . = _ _ . ; .

p ,,

e

- p , y D

J.i I

. L d

_

• D

i. t e

l'

$. 9 r *

.,N 4

e k

h a

1. +

, g u.

E E y

v. s M

~.h

~%

l

' ' j

.e 8

e 4.

4

• i+:

Il t i ~

l r b.

I 4

S yb i

)t-1 -  ;

e .

- , .. g

, = = + -

.. i

Enclosure 3 REACTOR CONTAINMENT LINER OVERLAY PAD FILLET WELD A'1 BJ.3VER VALLEY POWER STATION - UNIT NO. 2 LINER PLATE SERVICEABILITY REVIEW Prepared By [A.ll/.22

,P.W. Ward I

I 1215 4

,96

I Liner Plate Serviceability Review I The metallurgical investigation of the overlay pad fillet weld crack has found the local liner base material to exhibit a I transverse ductility lower than mill certified. The material is satisfactory for use as required by the ASME II, Part A, i.e., minimum elongation of 22% in 2 in.

SA-537 The CL-2 requirements; following will show that this local area of the liner base material is acceptable for meeting the intended service for the life of the m wer plant.

Stresses are calculated in the containment liner in a very conservative manner. A liner finite element model was developed by representing the composite reinforcing steel and liner steel I as an equivalent orthotropic shell, neglecting any strength contribution by the concrete. This model was subjected to the combined axisymmetric loadings of deadweight, DBA pressure, and in order to establish the membrane and bending I DBA temperature stresses in the liner. The total seismic shear force in the reinforced concrete containment wall (neglecting the strength of the concrete) was then assumed to be applied to the liner and I rebar in order to establish i conservative estimate of liner shear stress. The shear stress was combined with the liner finite element model membrane and bending stresses to determine the maximum stress intensity range. This stress range was compared to and found to be less than the established allowables.

g Stone & Webster Engineering Corporation calculation 12241-SM-EA-3 m " Stress Analysis of Containment Liner" demonstrates the adequacy of the containment liner design for the pressure, temperature, and seismic loads associated with the conditions specified in the I Beaver Valley Power Statiem Analysis Report. The conservatism Ifnit of the No. 2 Preliminary Safety above analysis is also and strains recorded by strain gage shown by the stresses I readings during the Reinforced Concrete Containment structural acceptance tests of Beaver Valley Power Station - Unit No. 1.

The geometry and test pressures for Unit No. 1 and Unit No. 2 of the Beaver Valley Power Station are the same. A Comparison of I the stresses resulting from metallurgical tests, analysis and the BVPS No. 1 structural acceptance test is shown in Table 1.

The current industry code applicable to the design of containment liners is ASME Section III, Division 2. This code recognizes that liner forces are displacement limited and provides liner allowables in units of in./in. of strain strain. to compare I stress results with current code In limits, the order membrane bending stresses calculated by elastic theory have been converted and to strain (by dividing stresses by the Modulus of Elasticity) and I are listed in Table 2. Since these strains are mostly membrane strain, they are conservatively compared in the table to the lower code allowable for membrane strain of 5 x 10-3 in/in. (The code allowable for membrane plus bending is 14 x 10-3 in/in.)

I I m5"U

I The above discussion and the tables clearly show that the local liner material is adequate for the intended service.

I I

I I

I I

I ..

I I 1215 193 I

I Table 1 Stress Comoarisons Stresses ASME III Allowable Max Intensity I Case Principal Cire (psi) Long (psi)

Intensity (psi)

Stress (psi)

Analysis Emergency -18000 -32000 32000 60000 Severe Operational -8000 -11500 11500 20000 Normal Operation -18000 -23000 23000 60000 Structural 25000 13000 25000 54000 Acceptance Test Metallurgical Test Yield (psi) Tensile (psi)

Results 86100 97700 BVt1 Structural Maximum Minimum Acceptance (psi) (psi)

Test Results 12000 8000 Table 2 Strain comparisons Strains ASME III Allowable Case Cire (in/in)* Long (in/in) * (in/in)

I Analysis Emergency .00065 .00115 .005**

Severe Operational .00029 .00054 .005'*

Normal Operation .00065 .00082 .0 0 2 * *

• Structural .00090 .00047 .0 0 3 * * *

• I Acceptance Test I Metallurigical Test Yield

• Tensile Results .0031 .220 BV41 Structural Maximum

• Minimum *

, Acceptance .0004 .00027 Test Results

• E = 27.9 x 106 psi

^

• • Factored Compression

• • • Service Table CC-2720-1

• • • • Factored Tension