ML19326B659

AO 50-313/74-10A:on 741025,metallurgical Analysis Revealed That Leak from Cold Leg of RCS (AO 50-313/74-10) Caused by Failure of Weld Metal Due to Fatigue Fracture.Weld Rewelded & Expansion Loop Added

ML19326B659
Person / Time
Site:	Arkansas Nuclear
Issue date:	01/22/1975
From:	Cavanaugh W, Phillips J, Rueter D ARKANSAS POWER & LIGHT CO.
To:
Shared Package
ML19326B655	List: ML19326B654 ML19326B659 ML19326B663 ML19326B665 ML19326B675
References
NUDOCS 8004170505
Download: ML19326B659 (15)

Text

hm, ,}. \/

]'

/

(\ . .'

• L

\s-(/ ,-

1. Abnormal Occurrence Report No. 50-313/74-10A

2. Report Date: 1/22/75 3. Occurrence Date:. 10/25/74

4. Facility: Arkansas Nuclear One-Unit 1

, Russellville, Arkansas

. 5. Identification of Occurrence:

See AOR 50-313/74-10 dated l1'/4/74 -

6. Conditions Prior to Occurrence:

~~

Steady-State Power X Reactor Power 1922 MWth flot Standby Net. Output 637 MWe Cold Shutdown i

Percent of Full Power 75 %

Refueling Shutdown Routine Startup

,_ Operation V Routine Shutdown Operation .

Load Changes During Routine' Power Operation

• e Other (Specify) s

7. Description of Occurrence: -

. i See AOR 50-313/74-10 O

-~x ,

- poo - q0 $ 0I 7

e f, ,

---  % g00p July 5, 1974 NSP-10, Rev. O Page 1 of 3

]

- J  ;

<j .

Abnormal Occurrence Report No. 50-313/74-10A Sheet

8. Designation of Apparent Lyse of Occurrence:

i Design

+ . Procedure l* Manufacture Unusual Service Condition Including

. Installation / Environmental Construction

! Operator

~ ~ ' ~

Other (Specify) f 'See attached report.

,g H s;

9. Analysis of Occurrence:

Fee attached report. ._

l I l 1 2 4

1

10. Corrective Act3.on: .

i See attached report.

l4

~;

i ,

s July S, 1974 NSP-ID, Rev. O Page 2 of 3 J

m

~, ) _,

. )

1

^

Abnormal Occurrence Report No.

50-313/74-10A Sheet 3

11. Failure Data:

The affected Report 50-313/74-9. weld had been reported as failing in Abnormal rrence Occu

. M

12. Reviews ar.d Approvals:

4 Reviewed and Approved by: Plant Safety Committee Yes (X) No ( )

Plant Superintendent Yes (X) No ( )

Reference:

___ JWA-824 Date:_1/29/75 Reviewed b :Y_

2!3!74 ~

r, n'.

' . ' Licensing Supervisor .

Approved'y:/f o *

,/>/ ,

// [

[ fety'Reviewf.o 'ttee t~

Date: )-V 2(-

Approved by: O Date: 2!J!7[

~

-d'-%

)fanager of Nuclear Servp[_,,

Approved by fjh2/ ,- ,e , e , / ,_

/ Director of Ppwer Production Date: 8-Y- 21 ~

Approved by: '; I t ,p

)'. ' 5

• / , Date:

Senior Vice F'cesident ,}

July 5, 1974 NSP-10, Rev. O .

lage 3 of 3 .

-)

l

, q q

~

ATTACHMENT TO AOR S0-313/74-10A A leak was detected on a 11/2 inch drain line from the cold leg of the Reactor Coolant System (RCS) located in the Reactor Building at Arkansas Nuclear One-Unit 1 on October 17, 1974. Upon investigation a crack was

discovered on one side of a socket weld and was subsequently repaired.

On October 2S,1974, another leak was detected'in the same weld located 180* from the previous leak. The failed weldment was removed and was replaced with a new assembly. Chronological fabrication details of the drain line from the inception of the project to this date are given in Appendix A. The failed weldment was taken to the Babcock 4 Wilcox

- Alliance Research Center, Alliance, Ohio, for metallurgical analysis.

The following naterials of construction were used for the 1 1/2 inch drain pipe assembly:

Coupling ASTM-A-182, F 316 Stainless Steel per USAS B31.7, Class 1.

Jones and Laughlin,1 1/2 inch 6000# Coupling Code J6AT j.

Pipe ASTM-A-376, TP 316L Stainless Steel per USAS B31.10, Class 1. Sandvik SR60 (316) Stainless,

, S'e amless, Annealed, and White Pickled, Schedule -.

160, Cold Finished Heat No. 430997 Welding Electrodes 3/32 inch TP 316L Stainless Steel, Heat No.

711008(Sandvik)

The analysis by B4W consisted of visual observation, dye-penetrant inspec- ,.

tion, chemical analyses, metallographic examination, fractographic analysis

and a microhardness surve~y. Results were as follows: .

1. Visual 31xamination:

The fracture extended approximately 90' around the circumference

of the weld metal and was located closer to the coupling than the

pipe. A b rown discoloration (oxidized appearance) was noted on the weld metal and pipe surfaces near the crack. A white powdery substance (borate) was loosely attached to the surface beside the ,

crack. e;

2. Dye Penetrant Inspection: .

Dye penetrant inspection of the coupling pipe assembly OD surfaces indicated that the assembly was free from indications other than the leaking crack. i 1

\

• e e I sme

~ , -en-- ---- ~,---a. --- -

, - < - . - , , - , - --n,, e

r's ..

. J )

,-~q .

3. Chemical Analyses: *

a. Chemical analysis by the wet chemical method on duplicate samples indicated that the materials of construction were TP316 stainless steels in accordance with the , applicable -

specifications. The results are shown on Table 1.

b. The results of a qualitive chemical analysis on the Srown

discoloration using a scanning Auger electron energy micro-scope indicated that the discoloration contained C, 02 , Si.

Na, P, 8 or C12 and N 2+

f

c. The results of a chemical analysis of the radioactive water taken from the drain pipe are shown in Table 2.

4. Meta 11ographic Examination:

A portion of the crack was cut out for metallographic examination.

Examination of the cross-sections of the crack in the optical d microscope showed that the crack appears to have originated from the root of the weld at the coupling-pipe interface. The crack was confined. entirely to the. weld metal. It propagated trans-granularly through the austenite and ferrite phases. Some crack-branching occurred from the main crack. Evidence of deformation

}

of the weld metal microstructures was noted near the root of the weld. No evidence of grain boundary oxidation, stress corrosion ~

cracking, gross porosity or microvoids was apparent.

5. Fractographic Analysis:'

The ' fracture surfaces were examined under the scanning electron microscope (Type JSM Scanning Electron Microscope, Japan Electron 3 Optics Laboratory Company, Ltd.) to determine the mode of failure.

The results indicated that the fracture was typical of those resulting from low stress intensity fatigue in austenitic stain- -

i less steels. The crack initiation appeared to have occurred at the root of the weld. Evidence of thermal' fatigue was also J observed near the initiation point at the root of the weld. No _ _ ,

evidence of dimple rupture was observed; thus ductile overload -I fracture was ruled out.

6. Microhardness Survey: l A microhardness survey, using a Tukon hardness tester; was. con-ducted on the coupling-pipe weldment sample taken 180' from the -

crack. Hardness values ranging from 181 to 271 KHN were observed for the weldment. The microhardness values obtained on both -

sides of the crack ranged from 168 to 221 DPH, Table 3.

3 '

) .

~

_ d

V)

(

These results indicate that the weld metal failed by a fatigue fracture mechanism. The crack started at the inherent stress concentration at the weld root due to excessive eccentricity in the fit-up gap of the pipe- .

coupling assembly.

Crack initiation may have been facilitated by the thermal fatigue cracking observed in the weld root. Thermal fatigue cracking could have been produced by (leaking) water striking the hot weldment during the repair welding operation.

1 Irregularities in the fit-up between pipe and coupling suggest:that sub-stantial residual stresses may have been present in the finished assembly and contributed to the crack growth process.

Alternating mechanical or thermal loading is a necessary requirement for ,_

fatigue crack growth. The source of such loadings cannot be speculated upon at this time.

No evidence was found to indicate that a material or manufacturing defi-ciency was responsible for or related to the failure'.

The detailed report,. including photographs, is available for review at the plant site. A stress analysis report containing the results of the field instrumentation program for the drain line modification should be complete and transmitted by March 14, 1975.

b

~;

e G

v e

TABLE 1 CHEMICAL COMPOSITION OF -

1-1/2 INCH DRAIN PIPE WELDMENT (TP 316 STAINLESS STEEL) .

Elements, %

C Cr Ni Mo Mn Si Np Coupling Wet Chemical-Duplicate 0.049 17. 29 12,76 2.65 --- --- 0.044

, Sample s -.

C e rt. (Code J6AT) 0.05 17.42 12. 74 2.69 1.73 0. 65 ---

A182, F316 0.08 16.00/ 10. 00/ 2.00/ 2.00 1.00 Max 18. 00 14, 00 3.00 Max Max ---

Pipa et Chemical-Duplicate 0.047 17. 5 6 13. 5 0 2.45 --- --- 0.025 Samples ,

ert. (Ht No. 430997) 0.041 17.2 13. 6 2.67 1.67 0. 70 ---

Cheek 0.042 17,4 12.8 2.62 1.68 0.74 ---

A376 0.08 16.0/ 11.0/ 2. 0/ 2.00 0. 75 ---

Max 18. 0 14. 0 3. O Max Max Woldmetal Wet Chemical-Duplicate 0.023 18. 69 13. 70 2.40 --- - . -

0.057 Sample Cert. (Ht No. 711008) 0.015 19. 5 a.3 2.20 1.77 0.35 SFA5.9 ER316L 0.03 18. 0/ 11.0/ 2. 0/ 1. 0/ 0.25/ _

Max 20,0 14. 0 3. 0 2. 5 0.60.

i i

l l

j G a

a e6e e

. ~ ~

TABLE 2 CHEMICAL COMPOSITION OF RADIOACTIVE WASTE WATER TAKEN FROM DRAIN PIPE S =

N. D' as SO3 & SO4 Cl- = 6. 2 t 0. 2 ppm F- = 0. 3'8 I . 02 ppm 109'4 t 3 ppm

~

B =

NH2 4

= < 0. I ppm pH = 6.7 t O.I s

e 9

0

• e

=.

. . m TABLE 3 MICROHARDNESS SURVEY READINGS ALONG SIDE OF CRACK Conversion to -

Rockwell B*

Bottom Side of Weld 172 DPH 83 168 "

81 19 9 90 181 85 ,

Top Side of Weld 191 "

88 (Near coupling) 207 "

91 O' 221 194 94 89 2

i

• Approximate Hardness Values 4

6 0

\

9

f m ~

\  !

V .

APPENDIX A CHRONOLOGICAL FABRICATION DETAILS OF THE PRIMARY COOLANT LINE 1 1/2 INCH DRAIN PIPE The 1 1/2 inch drain pipe was originally welded to the primary coolant line by the following sequences:

Step 1. The coupling was welded to the B4W's Inconel nozzle using Inconel 82 filler metal (Weld No. l', Fig.1).

Step 2. The pipe was welded to the coupling using TP308L filler metal. (Weld No. 2, Fig. 1).

Step 3. The pipe was welded to the 90* elbow using TP308L filler.

metal. (Weld No. 3, Fig.1)~. -

Step 4. Due to an alignment problem, Weld No. 3 in Step 3 was removed and was rewelded using TP316L filler metal.

Material certifications and welding records are available for examination.

, On October 17, 1974, after reaching 65% power, a leak was discovered in th. No. 2 weld of the drain line. APSL and field construction personnel

[N decided to repair this weld. The weld was repaired using Type 316L filler --

metal. Documentation is on file.

On October 2S,1974, .after achieving a power level of 7S%, another leak was detected in the same weld location 180* from the October 17, 1974, crack, Fig. 2.

APGL decided to replace the components. A new coupling-pipe assembly was fabricated and was reconnected to the existing nozzle and 90* elbow in the ,

system. Weld No. 2 was welded in the shop using Type 308L filler metal. -

Welds No.1 and 3 were made in the field.

Weld No.1, the nozzio-coupling joint, was welded using Type 82 filler ~2 metal and preheating 'the nozzle to 300*F. Weld No. 3, the pipe-elbow joint, was then welded using 3C8L filler metal at a preheat of 300*F.

Liquid Penetrant (PT) inspection was made on the root pass and the final pass. "

Ultrasonic Inspection (UT) was performed on all four welds on this drain -

line as well as on two other lines with the same configuration.

l l Only Joint No.1 of the rewelded drain line indicated a problem. Field repair procedures were invoked and rewelding was begun. A crack was found in the root' pass running about two-thirds.of the way around the pipe.

l

! ('ss__-)

l

~

~

S ge np

. -g -s Following the clack, it is believed that it extended into the Inconel nozzle about 1/8 inch, where it stopped. This defect was removed and re-welding the coupling was accomplished using a preheat of 200*F, after which .

I.- it was UT inspected. No. indications were found. (Certifications of this repair are available for examination.)

1 4

Strain gages were tack-welded to the pipe below Wel'd No. 2. Accelero-meters were mechanically fastened to the couplings. The , locations of both are shown in Figure 3.

The primary loop was then filled and brought up to temperature. The pumps

+

, were cycled and measurements recorded from the accelerometers and strain gages to try to find the cause of the induced stresses. The system was then cooled and drained. All couplings on the three drain lines were re-inspected by UT'to .see if there were any changes. None were noted. _

, It was decided to leave Welds 1 and 2 intact but .to add an expansion loop shown in Figure 4. Weld No. 3 was ground out and a new 90* ell was used.

All of the welds in the expansion line were made in the shop and only two welds, No. 3 and No. 34, were made in the field. All couplings on this line were UT inspected and then released. ,

i h

k t

I i

l Il 1

%)

l i

.(

L 6

l

^

J.

.5 -

-d O

2

-r, er we w mu=es) ,

,i e.y

' e 1. p.

b)

W O.  :

(dit3L.D U1cLD .

i w uera a w ato n .> u,w.= = =

t  %

4 4

((

s

\> .

t

'A g ___ .\p u x u / fZa.:.. 1 Wg$W m*It p g u p p o. I f SIG

' '___ .+

. wet.,t

i. pa. #

"'( Wtsp Ho. -s*

-* w a . o u o. 4 '

g + ,'s. ,3 1

6 .

{ j ^81G --

/' .-

~

p wea,ts av mast .

. JEj2SPECT VE -

• 6

% 9

\

l l we

• MSN - e e v.

J b 00 L '

i W .

h%

,\

~ .w l

.. t

. 7 t

. << s . ..

. .,g,,;, , -

s ..

)

, . y, .

\ J'  !

. , < p (

.. 'A I

r l

f1 + , ' ' ' .  : . l.x?)

~

.~ ' -

)

.-.. c

h. y r 4 ,, .. . ' . - . . ed i <". ,1.

1.1 -, .:

f _. , .. ,

'. i ) U-

' , .< + / tv --

. b)

'\,,.

.A og I. .v

,.'., wI ,'Il.,;_ $- />.p -4 C t

,:  % .k .hed .

. 1 Fig. 2 - Photograph s' howing leaks _

~

~

in Weld.No. 2 m

O eh.ea , .e.e ---um* * * *

^

h(0/g#._

1. 47/jgy wsuo so. i  :

I h

O / {f.M20MiTAN U<

=0 wo. 2 _

Ac:cELepop/EreN. I NA 6# "

h]'

mew vo., uh m ouera g 9 ,

WELO MO.4 .

l I .-

Y e l 9

1

. . lGUE2E 5 .

, l j

1 GT12AlfJ GAGES ALJD ACCELEROME 7EN LaATioMs .

O 1

~ - . ,

! l l l l

.'~

.mosmo

, ee 7

O

) ^

MM Dhd Gl59-LEE:C2At4&2 --. -_ _

)

$hffg g R glY N FI

. . m auUc ceue=

Ng 9 4 N. 3C e#

iN<( N y d44 e-d/

c f *Yli  :-!

9 , r q y

uuc eudeesa t#

I

' 4 I '

sinacAms wa.o wa W

3e e .

. 5.X DAN.S..QM . .# D. -

4 e

oe u