ML19323J164

From kanterella
Revision as of 18:51, 18 February 2020 by StriderTol (talk | contribs) (StriderTol Bot change)
(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to navigation Jump to search
Final Significant Deficiency 79-05 Re Weld Discrepancies Found During Insp of Reactor Vessel Structural Support. Cause Unstated.Weld Excavations in skirt-to-base Flange Weld to Be weld-repaired & Magnetic Particle Inspected by 801130
ML19323J164
Person / Time
Site: Beaver Valley
Issue date: 06/16/1980
From:
DUQUESNE LIGHT CO.
To:
Shared Package
ML19323J160 List:
References
NUDOCS 8006190321
Download: ML19323J164 (4)


Text

.

FINAL REPORT O,13 WELD DISCREPANCIES IN REACTOR VESSEL STRJCTURAL SUPPORT (NEUTRON SHIELD TANK)

AT BEAVER VALLEY POWER STATION - UNIT NO. 2 1.0

SUMMARY

OF DEFICIENCY A visual inspection of the reactor vessel structural support (neutron shield tank) at the jobsite storage area revealed some areas of weld overlap at the toe of two fillet welds, indicating a potential lack of fusion condition. Subsequent magnetic particle inspection of ran-domly selected areas of weld overlap, excavated to a maximum depth of 1/16 in. below base metal surface, produced a linear indication along one toe of the inner fillet weld of the support skirt-to-base blange weld.

2.0 IPl!EDIATE ACTION TAKEN Nonconformance and Disposition Report NO. 1033 was written on June i 29, 19'79, summarizing the results of the visual inspection at the jcbsite. Disposition aof the N&E Report proposed limited grinding of areas containing weld oterlap to investigate the possibility of a lack of fusion condition. After removal of the weld overlap con-dition, linear indicatiens were found 'to exist in the groundout areas.

Further grinding was performed on November 29, 1979, beyond the original 1/16 in below base metal surface, on four of the weld areas containing linear indications until the indications were removed.

The rendits of this grinding effort were evaluated and notification of the potentially reportable deficiency was made to NRC Region 1 on December 7,1979, under 10CFR50.55(e) (I) (iii) .

3.C DESCRIPTION OF DEFICIENCY A visual inspection of the neutron shield tank revealed some areas of weld overlap at the top of a reinforcing fillet weld on a groove tee weld, indicating a potential lack of fusion condition. The tee veld joins the 11/2 in. thick cylindrical support skirt to a 4 in. thick base flange. Grinding showed indiceons to be continuous over the entire exposed length of the weld, the indications appeared to be confined to the weld / heat affected zone, with varying depth of approximately 1/16 Lin. to 1/8 in, with a few indications requiring excavations.up to 1/2 in.

, g006190 M b

h

?

  • j' . -

A boat sample was removed from the weld at a location adjacent to the deepest excavation required to remove the defect (1/2 inch).

The sample was examined metallographically in order to characterize the defect. 'Two cross-sectional microspecimens were prepared, approximately 1 inch apart along the length of the sample, cloce to the excavation.. ;The defect observed in each specimen was similar and characterized as follows:

!, l'. At the surface of the weld, the defect appears to be located at the fusion line approximately 1/8 in deep between the toe of the

> weld and the base metal, suggesting lack of fusion at this location. Below 1/8 in., the defect traverses the base metal heat affected zone. (HAZ) , somewhat parallel to the fusion line contour and terminates in the base metal near the HAZ 1/4 to 3/8 in, below the base metal surface. The defect depth does not j exceed 3/8 in. in either sample.

2. No defects are evident in the wold beads /HAZ below the final weld part which indicates that subsurface beads are sound and that only the final pass was made using improper welding techniques resulting in lack of fusion. This condition is not unlikely, given the welding process used (submerged arc welding) in conjunction with the joint configuration and the apparent use of

): the edge of the groove in the skirt of a reference plane in I placing the first pass of the reinforcing fillet. This

! technique is considered diffic21t to achieve successfully and l ,

could result in overlap with attending lack of fusion.

4.0 ANALYSIS OF- SAFE ~"I IMPLICATIONS The neutron shield tank provides the primary support for the reactor pressure vessel, and its structural integrity is required for safe I shutdown of the' reactor. The condition of the skirt-to-base flange weld, if not corrected, would reduce the intended margin of safety for i that particular weld detail if subjected to loading from a design basis accident. The linear indications that were initially identi-

-fied. and reported to the NRC were interpreted to be potentially extensive, requiring a major repair effort to restore the weld in an acceptable . condition. However, when the weld discrepancies were removed, it became evident that the extent and nature of the weld discrepancies were relatively minor.

Certain inaccessible portions of the defective weld, which occur

! .bchind 13/4 in. thick reinforcing gusset plates located every 9 in.

along the length of the weld, were exempted from further examination andfrepair for the:following reasons:

W J

g.

O 1.~ A detailed analysis of the design load carrying capacity of the

. reinforcing gusset to base flange configuration produced design margins on the order of 8.0, which corresponds to stress levels equal to less than 20 percent of the allowable yield stress. ,

Since the portions of the skirt-to-base flange weld that are under the gusset plates are not within the primary load path, the

-stresses in,the local section of the weld, are even lower than those described above. In fact, the entire section of rein-forcement fillet weld under the gusset plates could be elim-inated without affecting the integrity of the skirt to base flange joint.

2. A f atigue-type failure will not occur since cylical loads during normal plant operation are negligible and therefore will have no effect on any weld which might remain in the inaccessible sections of weld under gusset plates.
3. A brittle fracture cannot occur unless a defect is exposed to a high stress level and the material has a relatively low toughness. As shown in Item 1 above, the maximum stress (during a faulted condition) is very low, so this condition alone can preclude brittle fracture. In addition to the low stress, fracture toughness of the material involved is very good. The impact requirements were specified to be 15 f t-lbs (average) at

) -30C F whereas, the actuals for the material (SA516, GR70 normalized) taken from certified mill test reports were on the order of 46-50 f t-lbs at -300 F. Also, since the lowest service temperature is specified to be 60 to 1000 F, the weld joint in question, operating at more than 900 F above the mil ductility transition temperature, will not be exposed at any brittle fracture condition regardless of defect size. Finally, the base ring weld is not exposed to a high radiation level and therefore will not be degraded by radiation exposure.

5.0^ CORRECTIVE ACTION 5.1 Shirt-to-base Flange Welds:

The weld overlap conditions in the neutron shield tank skirt-to-base flange weld have beencompletely removed and magnetic particle inspected except for limited areas behind gussets. All weld excavations in the skirt-to-base flange weld will be weld repaired and magnetic particle inspected in accordance with the

-original specification requirements.

L - - . - - - - - - - - _ _ _ _ _ _

O 5.2 Other Welds:

After'a thorough examination of all other welds on the neutron

' shield tank (exclusive of the skirt-to-base flange weld), most of the weld discrepancies noted proved to be minor or cosmetic in -

nature. Of the weld discrepancies that were classified as other than cosmetic, the majority resulted in weld excavation depths on the order of 1/8 in, below base metal surface with a few isolated areas with maximum depth excavations of 1/2 in.

Rework of these limited weld excavations (representing less than 1/2 percent of the total weld length of the structure, exclusive of the skirt-to-base flange weld) will be performed even though structural integrity could be established through rigorous analytical procedures.

All proposed weld repair on the reactor vessel structural support (neutron shield tank) will be completed by November 30, 1980.

)

.