ML20203K133

Forwards Addl Info Re Facility Phase II Welding Project Repts,Per NRC 860610 Request

ML20203K133
Person / Time
Site:	Sequoyah
Issue date:	08/01/1986
From:	Gridley R TENNESSEE VALLEY AUTHORITY
To:	Youngblood B Office of Nuclear Reactor Regulation
References
NUDOCS 8608060178
Download: ML20203K133 (10)

Text

_

TENNESSEE VALLEY AUTHORITY CHATTANOOGA. TENNESSEE' 374o1 SN 157B Lookout Place AUG 01888 Director of Nuclear Reactor Regulation Attention: Mr. B. J. Youngblood, Project Director PWR Project Directorate No. 4 Division of Pressurized Water Reactor (PWR)

Licensing A U.S. Nuclear Regulatory Commission Washington, D.C. 20555

Dear Mr. Ycungblood:

In the Matter of the ) Docket Nos. 50-327 Tennessee Valley Authority ) 50-328 Please refer to your letter to S. A. White dated June 10, 1986 which requested additional information on the Sequoyah Nuclear Plant Phase II Welding Project Reports. Enclosed is the response to your request.

If there are any questions, please set in touch with R. H. Shell at FTS 858-2688.

Very truly yours, TENNESSEE VALLEY AUTHORITY l

R.Gridley,Direflor Nuclear Safety aAd Licensing Enclosure cc: U.S. Nuclear Regulatory Commission (Enclosure)

Region II Attention: Dr. J. Nelson Grace, Regional Administrator 101 Marietta Street, NW, Suite 2900 Atlanta, Georgia 30323 I

I l

8608060178 860801 7 Og DR ADOCK 0500 An Equal Opportunity Employer l

I

, ENCLOSURE SEQUOYAH NUCLEAR PLANT WELDING PROJECT REPORTS

1. In 2.0 APTECH ENGINEERING REPORT (Supplemental Information), Page 3, 8th line, it is stated that "In the case of the feedwater lug, no engineering evaluation was requested by the plant." Why was installing the missing welds to drawing requirements chosen as the means of resolving a missing weld problem rather than performing an engineering evaluation as had been done with a very similar problem? Demonstrate that code requirements were met without installing the missing welds.

Response: When missing welds are identified during inspections, it is usually much easier to add the welds as required by the drawing (provided there is sufficient access) than to request engineering disposition to leave "as is." This was the case for feedwater lug FDH-203. However, when sufficient access does not permit welding, engineering disposition and subsequent drawing changes are initiated. This was the case for the Safety Injection System stanchion-to-pipe weld, 1-SIH-17.

In the case of 1-SIH-17, engineering gave preliminary approval to leave "as is" since cursory calculations showed that the actual weld provided was adequate for design loads. -

Therefore, addition of the weld was not required.

Engineering will provide final calculations to demonstrate structural adequacy of the subject support when the drawing is revised and reissued. '

2. The term " separated weld" is used in 2.0 APTECH ENGINEERING REPORT (Supplemental Information), Page 3, 12th line. Define the basis for your assessment of this weld failure as being due to operating transients and not having been due to poor weld quality or cracking during fabrication.

Response: Since no cracked welds have been found during the reinspections, there is no reason to question the quality of the construction welds. Conversely, there is not a readily identifiable basis for attributing the occurence to an operating transient. TVA determined this to be an i solated case since no other cracked welds or damaged supports were found in the same area and the cause is indeterminate .

3

3. In the APTECH ENGINEERING REPORT, the Table titled, "NOI DESCRIPTIONS

- SEQUOYAH NUCLEAR PLANT UNIT 1". NOI Number SQ0201, under Disposition and Additional Comments it is stated: " . . . clean weld area per SQM-17, paint and re-examine." Explain how code requirements were met with the examination following painting.

• + +46-m 4 - e ,4 eg. # . A ea + 3. m gy ,

Response: The note under NOI SQO201 in the APTECH ENGINEERING REPORT is an editorial error. The Maintenance instruction required that the subject weld be added, cleaned, visual and PT examined, then painted. The inspection report shows that the weld passed final examination (visual and PT) on 12/9/85 and has not yet been painted.

4. In the APTECH ENGINEERING REPORT, Table 4-1 lists 5 Licensing Event Reports concerned with welds. Provide the number of LERs evaluated in this search. Here any failure analyses conducted of the welds covered by these LERs? If so, please provide them.

Response: There were 840 LERs evaluated in the search. A metallurgical failure analysis was done in conjunction with LER 80156. The failure analysis involved a vendor weld (seal water injection line to reactor coolant pump weld).

5. Here there ever other than E7018 carbon / low alloy steel shleided metal arc welding electrodes on the Sequoyah site, such as E8018C3?

Demonstrate that incorrect electrodes were not used on any weldment.

Response

A. Construction Phase Yes, small quantities of E6010, E11018X and various other types of specialty maintenance electrodes were kept on site. These materials and their use were strictly controlled. Their uses were limited to such things as construction plant (temporary construction facility) maintenance and construction; maintenance of construction equipment; hard facing of construction equipment cutting edges; crane boom repair; build up for hard facing of worn construction equipment; and the fabrication of construction jigs and fixtures.

In addition to the previously described maintenance materials, small quantitles of E8018C3 and E7010Al materials were used on appropriate permanent plant features. The use of these materials was also strictly controlled in accordance with the construction Quality Assurance / Quality Control Program.

Checks and balances were reflected in construction procedures to insure the proper procurement, storage, and application of welding materials used for permanent plant construction. These included the recording and verification by QC Inspectors of filler materials by type of safety related pipewelds and a QC surveillance to spot check proper filler material application on all safety related welding. In addition QA reviews of safety related pipeweld records included electrode type as a check point.

B. Operation Phase Yes, like construction, small quantities of various types of other electrodes are maintained for specialty welding and specialized maintenance applications. These include carbon steel coated electrodes othcr than E7018 which have not been used on safety related plant features. These applications include maintenance of shop and shop equipment, fabrication of temporary jigs and fixtures, and noncritical maintenance of non-safety related balance of plant items.

These materials and their applications are strictly controlled in accordance with approved plant procedures.

Maintenance and modification procedures provide for the QC verification of proper filler material use for safety related applications. This verification provides indirect traceability to heat / lot numbers. In addition, a QA surveillance program provides additional spot checking of proper electrode usage.

6. For the Bechtel Audit, what were the total number of welders and inspectors in the populations from which the audit samples were taken?

Provide separate totals for the Office of Construction and Nuclear Operations.

Response: Populations from which the Bechtel Audit Team selected are as follows:

Organization Welders Inspectors 4

Construction approx. 3100 approx. 180 Nuclear Operations approx. 205 approx. 120

7. The TVA Reinspections checked the relative magnetism for all welds, austenitic and ferritic. What was the procedure for this inspection method? Provide justification for different levels of magnetism and their acceptance criteria, particularly " weakly magnetic". '

Response: The magnetic check for generic filler metal type (i.e.,

i ferritic or austenttic) was performed by touching a small permanent magnet to the weld deposit and noting his

judgement as to whether the deposit was strongly, weakly, or non-magnetic. The inspector also noted whether the base materials being joined were stainless or carbon steel.

Evaluation of correctness of filler metal was done by OE according to the following guidelines:

1. The correct weld metal for welds joining stainless steel to stainless steel should be weakly magnetic or non-magnetic.

2. The correct weld metal for welds joining stainless steel to carbon steel should be weakly magnetic or non-magnetic.

, _ _ , , _ _ _ _ _ , , __ - - . _ _ _ _ - . - - + - - -

. 3. The correct weld metal for welds joining carbon stQel to carbon steel should be strongly magnet,1c.

The above guidelines are as contained in P.S.3.C.ll.1 (RI).

The " weakly magnetic" category as a permissible condition for items 1 and 2 above reflects that the correct stainless steel weld metal used in these welds should appear non-magnetic or weakly magnetic depending on delta ferrite content and/or degree of base metal dilution.

8. Cracks were not listed as one of the attributes in the tables of TVA Reinspection Report. Here any cracks found during the TVA Reinspection? Also, porosity was not an attribute listed in the structural welds table. What was the rejection rate for porosity in the structural welds in the TVA Reinspection?

Response: Both cracks and porosity were attributes that were checked in the reinspection effort. No cracks were found during the reinspection. Rejectable porosity was not found on any structural welds.

9. In 4.4.1, Page 8, line 21, of the five welds which were ground, were the manufacturer's minimum wall thickness requirements encroached upon?

If so, to what extent?

Response: Only one weld (2CCF-68) of the five which were ground to reduce surface indications had its manufacturer's minimum wall thickness encroached upon. This weld is in a 4-inch schedule 40 carbon steel pipe. The measured thickness localized ground area is 0.198". This is 0.0094" less than the manufacturer's minimum wall requirement of 0.2074" but is more than twice the design minimum wall of 0.08".

10. In 4.4.1, Page 10, line 1, the rough condition of two welds found during the reinspection is discussed. Provide information that justifies the statement, "The indepth investigation of the welder and inspector qualification revealed no indications of inadequacy of the welder or inspector capabilities." What was done to demonstrate that this level of workmanship by this welder and/or judgement by this inspector were not repeated elsewhere at Sequoyah?

Response: After proper removal of paint, both welds were inspectable by the penetrant method. The inspectors' certification flies were reviewed and both inspectors in question were found to have at least two years experience at penetrant testing when the inspections were made. The welder was initially certified in May 1975 and had welded in nuclear applications off and on since that time. TVA determined that no further investigation of the inspectors' or 1

welders' work was necessary.

' 11. In 4.4.1, Page 11, in the table titled " PIPING WELOS", the rejection rate when expressed in terms of the percentage of welds rejected is 56%

(184/333). Even allowing for some rejected welds counted more than once because of more than one rejectable attribute, the rejection rate is very high. a) What is the root cause of this high rejection rate of originally inspected and accepted welds? b) Is there any basis for concluding that there is a connection between the employee concerns expressing doubt about inspectors capabilities or that harassment and intimidation of inspectors occurred? c) With respect to question a),

address in particular the attribute underfill, which has very specific code requirements. d) The arc strike / weld spatter rejection rate was 31%. What is the root cause for this high rejection rate? e) What were the original inspection criteria for these weld attributes? f)

What were the reinspection criteria for these attributes? g) What is the justification for elimination of inspecting arc strikes for cracks in G-29C?

Response: The reinspection rejection rate on a per weld basis to inspection requirements is 24% (80/333). The 184 arc strikes and weld spatter indications were reportable but not rejectable. Base metal outside the weld area was not required to be examined by the construction code. The procedure used fcr the reinspection required base metal indications outsloa the weld to be reported.

Any reinspection effort will typically have a rejection rate of 5-10 percent, h*0 Wever, a reinspection such as this can have a rejection rate approaching 20-25 percent because of the circumstances under which the reinspection was made.

a. What is the root cause of this high rejection rate of originally inspected and accepted welds?

Response: The root cause of the high discrepancy rate involves both psychological factors and a changing inspection philosophy in recent years. Inspectors performing this reinspection anticipated "second-guessing" of their judgements by others. Because there is judgement involved in weld inspection close calls will inevitably become rejects under such conditions. It is unrealistic to expect the results of a reinspection performed under the degree of scrutiny involved here to yleid results comparable to those performed in the 1970-80 era. This does not imply inadequate inspection during construction. It does reflect a change in weld inspection philosophy and mathodology over the past 15 years and most particularly in the past 2-3 years. The significant change involves less reliance on the inspector's eyes and judgement of the weld as a whole, and more on quantitative measurement of '

every attribute on every increment of weld.

. .. ~.

-z

' To a lesser degree, the current discrepancy rate is a result of changes in acceptance criteria (see "d" below).

b. Is there any basis for concluding that there is a connection between the employee concerns expressing doubt about inspectors capabilities or that harassment and intimidation of inspectors occurred?

Response: The program was working properly and inspectors were performing properly. He have no evidence that would support the concerns about inspector capability and inspector harassment or intimidation.

i

c. With respect to question a), address in particular the attribute underfill, which has very specific code requirements.

^

Response: Seven of the 11 welds rejected for underfill involve sockolet branch connection fittings to pipe runs. These fittings are proprietary products designed to provide i

integral reinforcement of the branch opening. Because of the configuration of the fittings themselves and the geometry of the connection as a whole, the correct weld size and configuration is not obvious. This is particularly so in the cases where there is little difference in the size of the run pipe and branch connection.

The remaining four instances of underfill involved welds joining members of unequal thickness (pipe to valve or fitting). Here the reported underfill was with respect to the edge of the thicker member. However, the weld thickness was greater than the minimum pipe wall thickness. (Refer to Note 6 of Appendix 4.4.)

We agree that the code requirements are explicit with regard to underfill as applied to typical piping girth butt welds. Underfill in such welds has not historically i

been a problem and was not in this reinspection.

d. The arc strike / weld spatter rejection rate was 31%. What is the i root cause for this high rejection rate?

Response: TVA procedures in use during the construction of Sequoyah Nuclear Plant prior to March 21, 1979 did not require the reporting of arc strikes unless a crack was present. The procedures used during the reinspection did require

, reporting of arc strikes. The data simply reflects the procedure requirements in the two different time frames.

l e

i l

__ _ ... _ . _ _ - - . - .- :- . _==  :=w

. Weld spatter has been prohibited by TVA inspection criteria since 1970. Neither the construction era nor current piping codes (ASME Section III and 831.1) address the condition. Although lumped with arc strikes as a discrepant condition, it was reported on only three piping welds.

e. What were the original inspection criteria for these weld attributes?

Response: Please refer to item "d" for response.

f. What were the reinspection criteria for thcse attributes?

Response: Both arc strikes and weld spatter were treated as discrepant conditions during the reinspection.

g. What is the justification for elimination of inspecting arc strikes for cracks in G-29C?

Response Cracks have been and are presently prohibited in welds and adjacent base material in TVA inspection procedures. This prohibition includes cracks in arc strikes or anywhere else within the zone of inspection.

12. In 4.4.1, Page 11 and 4.2.1, page 13, in the tables titled " PIPING WELDS" and " STRUCTURAL WELDS" respectively, expressing weld rejection rates based upon the attribute inches is misleading. There was only a finite number of welds inspected, and a qualified craftsman should be capable of making welds which meet all of the attributes in all of the inches submitted to inspection. For these tables, please rearrange the data as follows: -

Response

PIPE WELDS NO. OF WELDS N0. OF WELDS WITH NO. OF WELDS TYPE OF WELD REINSPECTED REJECTED REPORTABLE INDICATIONS BY CODE BY CODE Socket Welds Office of Const. 204 78 0 Nuclear Ops. 34 6 0 Butt Welds Office of Const. 68 46 0 Nuclear Ops. 22 6 0 Attachment to Pipe Wall Office of Const. 5 3 0 Nuclear Ops. 0 0 0 Total Welds Office of Const. 277 127 0 Nuclear Ops. 56 12 0

_ _ a _:

STRUCTURAL WELDS

• NO. OF HELD NO. OF WELDS JOINTS NOT NO. OF WELDS WITH REPORTABLE MEETING TYPE OF WELD REINSPECTED INDICATIONS DESIGN REQUIREMENTS Fillet Welds Office of Const. 1080 160 0 Nuclear Ops. 148 21 0 Butt Welds Office of Const. 50 4 0 Nuclear Ops. 0 0 0 Other (specify) - Flare Office of Const. 92 24 0 Nuclear Ops. 24 2 0

• Weld joints were evaluated not individual weld segments.

13. In the TVA Reinspection Report, a comparison is made between original inspection results and the reinspection results for piping welds. If such a comparison can be made in a quantitative manner for structural welds, please present the data.

Response: The original inspection was made on an item basis rather than individual weld, consequently, we do not believe it possible to make a meaningful weld-by-weld comparison between the reinspection results and the original inspection results for structural welds.

14. Referring to the Legend for Table 4.2, in the Final Resolution column, define the meaning of the letter codes in parentheses.

Response: The letter codes located within the parenthesis in the legendofthefinalRekolutionofTable4.2denotevarious s

design sections within-the Divisico of Nuclear Engineering that had lead responsibility of the resolutions addressed by the code of Al through A10.

NEB CSM - Nuclear Engineering Branca - Code Standards &

Materials Section CEB M2 - Civil Engineering Branch Mechanical Analysis Section #2 SQEP M3 - Sequoyah Engineering Project Mechanical Design Section #3 SQEP C3 - Sequoyah Engineering Project Civil Design Section

1. 3 I

., , 15. There are some employee concerns about various structures not

• being in accordance with the as-Suilt drawings. Did the TVA reinspection address this issue? If so, report the deviations from the as-built drawings found. Report the deviations in configuration as to type of deviation, the rate of a, type of deviation compared to the.

number in the reinspection population, and if such deviations resulted in not meeting code requirements.

Response: No. This reinspection program was not intended to address deviations in configuration from as-built drawings. This subject is being addressed by TVA's employee concerns program.

16. Table 4.3 shows that a total of 50 structures were reinspected in the TVA reinspection program. However, Table 4.4 shows only 31 structures as having been reinspected. Explain the discrepancy.

Response: Table 4.3 is correct for number of structures. Table 4.4 shows number of items or what was defined in Phase I as a package. An item nay contain only one structure or a number of structures.

To correct the Table 4.3, the title should read " NUMBER OF REINSPECTED STRUCTURES".

There are 31 packages (items) shown in Table 4.4.

Two packages (No. 10 and No. 30) are not reported in' Table --

4.4. Item #10 was not reinspected and Item #30 is reported in the Mechanical Reinspection (Table 4.2).

The remaining packages breakdown to the following number of structures.

Items 2 thru 9) 12 )

14 thru 16) All contain one structure i

18 )

20 thru 21) ,

23 thru 29) 31 )

item 1- 2 structures 11 - 2 structures '

13 - 3 structures 17 - 2 structures 19 - 14 structures 1

- . . .