ML19276D617
| ML19276D617 | |
| Person / Time | |
|---|---|
| Issue date: | 11/07/1978 |
| From: | Hunter V, Whitsell D NRC OFFICE OF INSPECTION & ENFORCEMENT (IE REGION IV) |
| To: | |
| Shared Package | |
| ML19274C950 | List: |
| References | |
| REF-QA-99900036 99900036-78-3, NUDOCS 7901080073 | |
| Download: ML19276D617 (8) | |
Text
,.
VENDOR INSPECTION REPORT U. S. NUCLEAR REGULATORY COMMISSION OFFICE OF INSPECTION AND ENFORCEMENT REGION IV Report No.
99900036/78-03 Progran No.
44040 Company:
Combustion Engineering Incorporated 911 West Main Street Chattanooga, Tennessee 37402 Inspection Conduc'ted:
September 25-29, 1978 h[U[. [unterTContractor Inspector, Vendor
// ((
Ik h Inspectors :
V.
H Date Inspection Branch Approved by:
[vEd
- - P 74 D. E. Whitesell, Chief, ComponentsSection I, Date Vendor Inspection Branch Summary Inspection on September 25-29, 1978 (99900036/78-03)
Areas Inspected:
Implementation of 10 CFR 50, Appendix B and applicable codes and standards including control of special processes.
The inspection involved thirty (30) inspector hours on site.
Results:
In the one (1) area inspected, no apparent deviations or unresolved items were identi fied.
i 4
t 790108 e@q$)cro 13
_2_
DETAILS SECTION A.
Persons Contacted CombustionEngineering(CEj.
- J. T. Rich, General fianager
- W. A. Stone, Manager, thrclear Quality Assurance
- J. G. Rawlings, Section Manager, Quality Systems
- J. W. Reed, Section Manager, Quality Engineering
- L. A. Hoenig, Supervisor, Component Engineering Standards
- B.
J. Bates, Audit Coordinator
- L. C. Miller, fide Section Manager
- R. J. Ferester, Project Manager
- J. P. Fava, Inspection Section Manager Hartford Steam Boiler
- R. E. Fitz, Authorized Nuclear Inspector
- J. N. Garrett, Authorized Nuclear Inspector
- Indicates personnel present at the Exit Interview.
(See paragraph F.)
i B.
Special Welding Applications 1.
Objectives The objectives of this area of the inspection were to verify that special welding specifications (cladding, hard surfacing, seal welding, and weld repair without postweld heat treatment) conform with the additional requirements established by ASME Code,Section III and IX, NRC regulations, and the vendor's QA program commitments.
2.
Method of Accomplishment The objectives of this area of the inspection were accomplished by:
a.
Review of the ASME accepted QA Manual, (1) Sections 5, 8, 9,10,15, and 16, to verify the vendor had established procedures to prescribe a system for the development and qualification of special welding specifications, and for qualifying welders and/or welding machine operators.
.- b.
Review the following procedures:
Weld - SMA s.8-103-0, dated April 30, 1976.
' Radiographic - N-2.4.1.3.(F), dated September 25, 1975.
Liquid Penetrant - N-2.4.3.9.(F), dated October 28, 1975.
to verify the special requirements governing special welding application procedures and performance qualification imposed by ASME Code regarding test sample size, examination of test sample, and special essential variables are satisfied.
c.
Interviews with personnel to verify that they were knowledge-able in the procedures applicable to welding.
3.
Observations Based on records review, observation of in-process manufacturing, and discussions with CE personnel, the following is a summary of significant events that have occurred since CE identified a serious problem with their shiel e d metal arc welding process at the Chattanooga, Tennessee facility:
a.
Problem Definition l
In August 1978, a required in-process liquid penetrant examination was being performed on the nozzle to safe end i
welds of two (2) reactor pressure vessels.
Results of the examination revealed numerous unacceptable crack-like indications on the safe end to nozzle weld surface.
An in-process radiographic examination further disclosed that crack-like indications were apparent throughout each of the nozzle to safe end welds of this vessel.
CE then initiated actions to determine the extent of the indications which l
revealed that in all, a total of twenty-one (21) nozzle to safe end weld seams of vessels in various stages of fabrication exhibited unacceptable indications.
These indications were located in the inconel weld which joins the stainless steel safe end to the shielded metal arc inconel buttered 508 type nozzle.
The indications were linear in nature and were scattered throughout the weld, but with primary concentration towards the stainless steel safe end or inconel buttered nozzle.
CE immediately issued a stop-work order for all inconel welding in the shop and initiated a statistical analysis of
. material and process records in an attempt to identify any condition that would cause or contribute to the cracking problem. The documentation search and tabulation disclosed a
that:
(1) Welds made in the shop with Arcos supplied inconel electrodes manufactured prior to October 1977, did not show the unacceptable indications.
(2) Welds made in the shop with Arcos supplied inconel electrodes manufactured after October 1977, exhibited numarous unacceptable indications.
(3) Of the three (3) welding pos.i"as involved (1-G, 2-G, 3-G), only the 2-G welding position exhibited numerous unacceptable indications.
As the first two (2) digits of the Arcos lot numbering system indicates the year and month of electrode production, CE concluded that electrode lots 7K (77-October) and newer were
" suspect" with electrode lots produced prior to 7K as "non-suspect." A review by the Inspector of material test reports for both " suspect" and "non-suspect" electrodes verified that all constituents were within the limits of both the AWS and ASME Code rcquirements.
It was noted howeser, that as of October 1977, the Arcos material test reports indicated a significant increase of iron (Fe) from 1-2 percent to that of 7-9 percent (Tne ASME Code specifies a maximum of 10 percent Fe).
The inspector was advised that the increase of Fe was due to a change of core wire vendors by Arcos and that Arcos had revised their flux chemistry formula to accomodate the increased level of Fe.
b.
Weld Classification i
Classification of weld seams welded with the " suspect"
?
l electrodes are listed as follows:
(1)
Full penetration pressure boundry welds, 2-G position (horizontal) welding.
Example: nozzle to safe end welds.
This type of weld seam is liquid penetrant and radio-graphically examined.
(2) Partial penetration pressure boundry welds,1-G and 3-G (flat and vertical) weld position.
Example: CRDM housings to closure head welds and "J" prep buttering.
. This type of weld requires a liquid penetrant examination of the root pass, for each 23 inch depth of weld material deposited; and the completed weld, both before and after hydrostatic test.
(3)
Structure welds,1-G weld position.
Examples:
nozzle buttering, attachment of core stabilizer, and core stop lugs to pressure boundry base material. The nozzle buttering is subjected to both liquid penentrant and radiographic examination.
All lug welds receive a liquid penetrant examination of the root pass, for each b inch of weld material deposited, and the completed weld both before and af ter hydrostatic test.
(4)
Miscellaneous, 1-G position.
Examples: clad, back claddilig, keyway claddings, etc.
C.
CE Investigations CE conducted a formal investigation of the " suspect" inconel electrodes including those deposited.
The Metallurgical Department removed three (3) boat samples from representative welded areas for standard optical 6
microscopic evaluation, scanning electron microscopic evaluation, and chemical analysis.
Evaluation of the three (3) samples at 100X showed typical hot shortness fissuring.
Scanning electron microscopic evaluation of the fr!ctured surface showed no gross segregation of elements that would cause hot shortness fissuring.
Chemical analysis j
by direct reading spectrograph, x-ray, fluorescence, and volumetric techniques showed no unacceptable chemistry results.
To determine if the cracks in the shop welds could be reproduced, an SA-533 test plate was machined with a groove having dimensions of 3/4 U
inch dee p, 3/4 inch root, and a 30 included angle.
Test plate welds ith both " suspect" and "non-suspect" inconel electrodes.
were madu s Using ASME Sections III and IX for radiography, liquid penetrant, and side bend test criteria, both " suspect" and "non-suspect" electrodes were tested in all three (3) weld positions.
Results of these tests disclosed that the "non-suspect" inconel electrodes produced an acceptable weld and that the " suspect" electrodes failed in the 2-G weld position only.
I D.
Investigative Results and Conclusions j
Based on statistical analysis and metallurgical tests, CE has concluded that:
(1)
Both " suspect" and "non-suspect" inconel electrodes supplied by Arcos meet or exceed the CE purchase order, AWS, and ASME require-ments.
- (2)
Inconel electrodes supplied by Arcos af ter October 1977, do, under certain conditions, produce numerous unacceptable cracks when used in the 2-G weld position.
(3)
The referenced cracking condition is readily detected by in-process and acceptance nondestructive examinations.
(4) While not conclusive, the available information suggests that the weld cracking phenomana is the result of an electrode chemistry i
imbalance and/or the 2-G weld position.
E.
Corrective Actions In order to produce acceptable inconel shop welds, CE has initiated the following corrective actions.
(1)
Isolated all " suspect" inconel electrodes and will re-issue only after performing weld tests in both the 1-G and 2-G positions in a restrained machined weld groove and found acceptable.
(2) All nozzle to safe end full penetration pressure boundry welds which were welded in the 2-G position with the " suspect" electrodes have been dispositioned to remove all unacceptable welds and repair with approved materials.
(3)
Notified electrode vendor of the problem who is currently develop-ing an acceptable electrode with a chemistry balance conducive to welding in the 2-G position.
(4)
Initiated an interim requirement that for any new procurement of inconel electrodes, weld tests will be performed in both the 1-G and 2-G positions.
Test welds will be radiographed and side j
bend tested to determine acceptability.
(5)
Continue to analyze all future welding to assure that initiated corrective action measures are valid and effective.
(6)
Continue tests of the " suspect" material in an attempt to determine the mechanism which caused the unacceptable indications.
(7)
Performed in-process radiography and penetrant examinations on two (2) nozzle to safe end welds utilizing electrodes approved by the above listed criteria.
No unacceptable indications were noted, which indicates corrective action effectiveness.
.- 4.
Fi ndings_
Based on a thorough review of records, discussions with technical and management personnel, and observations of production and on-going corrective actions, it was concluded that:
a.
CE has implemented an effective quality assurance program in that the program was responsible for:
i (1)
Detecting the problem.
(2)
Providing prompt isolation of affected components.
(3)
Initiated corrective actions that corrects affected components.
(4)
Initiated an on-going program to determine problem cause in order to preclude repetition.
(5)
Notified both customers and suppliers in an attempt to better assure that the problem is noc of a generic nature.
b.
While it vias determined that the CE categorized " suspect" inconel electrodes has been sold to various other manufacturers of nuclear components, it is not considered a generic problem at this time due to the following:
(1)
The CE categorized " suspect" inconel electrodes have met or exceeded all AWS, ASME, and CE purchase order requirements and, to date, have not been conclusively proven to be the primary cause of the cracking problem.
i (2)
All Class I and Class II components intended for use in nuclear facilities receive one or more ASME Code required nondestructive examinations, any one of which readily detects the referenced Weld Cracking Phenomenon.
i F.
Exit Interview 4
1.
The inspector met with management representatives denoted in paragraph A., at the conclusion of the inspection on September 28, 1978.
2.
Management representatives were advised of the following:
8-a.
No deviations or unrasolved items were identified during this inspection.
b.
The inspector has no further questions at this time relative to the we!d cracking problem described in paragraph B.,
of this report.
Itowever, at the time of our inspection of the Arcos Corporation in Philadelphia, Pennsylvania, the inspector will pursue any generic possibilities.
(Reference Report No. 99900351/78-01).
3.
Management representatives had no questions relative to the inspection.
?
s
.~
~
g