ML19282A035

Deficiency Rept:Defects in Toe of Ring Beam Attachment Weld. Cracks Caused Hydrogen Cracking,Characterized by Intergranular Fracture of Course Grained Haz.Corrected by Grinding & New Welding Procedures

ML19282A035
Person / Time
Site:	Fermi
Issue date:	04/19/1979
From:	DETROIT EDISON CO.
To:
Shared Package
ML19282A033	List: ML19282A032 ML19282A035
References
EF2-46-133, NUDOCS 7905090355
Download: ML19282A035 (3)

Text

- . .

kk DEFICIENCY REPORT

'

TORUS SHELL INDICATIONS AT THE RING BEAM ATTACHMENT WELD ENLARGEMENT INTRODUCTION The primary containment for the Enrico Fermi Atcmic Power Plant, Unit 2 was designed, erected, leak tested and N-stamped in accordance with the ASPE Boiler and Pressure Vessel Code during the early 1970's. Subsequently, new torus hydrodynamic loads were identified. The new loads were related to pool response due to postulated LOCA and pool response due to main steam relief valve operation. Detroit Edison has committed to a torus modifica-tion program that would provide design, analysis and implementation of modi-fications so that the modified torus structure will withstand these new loads and comply with the acceptable design margins. A detailed description of the torus modification program and the required modifications was submitted to NRC via Amendment No. 12 of the FSAR, Section 3.8.2 and Reference 17.

The installation of the designed modifications is being performed by Reactor Controls Inc. As part of the modification program, the existing 5/16" fillet weld between the ring girder at each mitered joint and the suppres-sion chamber shell is being increased to 3/4" fillets as shown in Figure 1.

In the process of inspecting these welds, defects in the toe of the welds were discovered. The typical locations of these defects are illustrated in Detail A of Figure 1.

Description of the Defects The indications were discovered to be relevant defects through magnetic particle examination and liquid penetrant examination. The de fects pre-dominantly are oriented parallel to the welds in the area shown in Figure 1.

Presently, ultrasonic examination is being conducted to determine the presence of subsurface indications. Documentation providing dimensional description and location of the defects is available at the site.

Failure Analvsis Ten (10) metal samples were removed from selected locations of the torus ring beam welds to determine the cause of the cracking. Metallurgical and fractographic methods were used to examine these samples. Hardness tra-verses and chemical analyses were also performed. The surface cracks were found to emanate from the toe of the outermost veld pass of the 3/4" weld.

The subsurface cracks observed in the samples emanated from the underbead 7905090355

.

Failure Analysis (contd) welds and were determined to have initiated while applying the 3/4" weld.

The cracks then ran through the heat affected zone (HAZ) downward into the shell base material. Once the cracks were in the base material of the shell, they propagated along and across the banded microstructure in a stepwise fashion.

The base material chemical analysis and hardness were typical for SA516, Grade

70. Hardness in the HAZ of most of the samples examined exceed 400 Knoop.

A complete metallurgical report is on file at the site.

Cause of Cracking The cracks were found to be a result of hydrogen cracking and were charac-terized by intergranular fracture of coarse grained HAZ. The RAZ of welds which exhibited cracking were excessively hard (Knoop 400). The high hard-ness HAZ's resulted from rapid cooling and lead to the subsequent formation of martensite, which is susceptible to hydrogen cracking.

Corrective Action After the defects were discovered, a Repair Program Document was written and submitted to the State of Michigan for the repair, as required by the ASME Code. The corrective action consisted of physically removing, by grind-ing of all relevant surface and subsurface indications. The base metal and 3/4" fillet welds were restored by utilizing new welding procedures that speci-fied electrode baking (6000 to 800 F), preheat (300 to 4000F), specific minicum energy input, and post-weld baking temperature and duration.

The weld repaired area will be magnetic particle examined and ultrasonic test-ed after a minimum of at least 72 hours8.333333e-4 days <br />0.02 hours <br />1.190476e-4 weeks <br />2.7396e-5 months <br /> after the post bake duration is com-pleted and the weld metal returned to ambient temperature. An integrated leak rate test of the containment vessel in accordance with Appendix J of 10CFR50 will be performed prior to start-up.

Design Acceptance Criteria / Code The primary containment was originally designed, erected, leak tested and N-stamped in conformance with the 1968 Edition (through summer 1969 Addenda) of section III of the ASME Boiler and Pressure Vessel Code.

For the design and analysis conducted for the containment modification pro-gram, the design rules and stress limits of the 1977 Edition (through summer 1977 Addenda) of the ASME section III code are used. In utilizing the ASME Code, the suppression chamber and vent system are classified as class MC components and therefore, the rules of subsection NE are applied.

Safety Imolication of the Deficiencv A preliminary fracture mechanics analysis indicated that the nature of cracks existing in the toe of the weld would most probably not be tolerated in view of the cyclic nature of some of the hydrodynamic loads. The indications were de-tected and repaired as part of the normal site inspection program. The repaired structure will perform its safety function as designed.

P

.

. FIGURE 2

.

.

. l

[

( \

~

\

g

'

-MITRE JOINT i

,-, E- < J ,,,1c a E,cc , _

SEE DETAIL A

.

6 RING GIRDER

~t~

~%_ ~

/ \

\ \

t

.u- x '

.

}

attNroacED RCI h' ELD TYPICAL DEFECT EXISTING REGION

-

CBI hTLD

,$- ,

< jfl --- >

' DETAIL A ,

.

T F OR A4@1 eena m - - - - - ,

_--.n_"~~~~ ',

t 1, -

. .

~

k.