ML20209D369
| ML20209D369 | |
| Person / Time | |
|---|---|
| Site: | Sequoyah  |
| Issue date: | 06/17/1983 |
| From: | Rich Smith, Travis R, Whitt K TENNESSEE VALLEY AUTHORITY |
| To: | |
| Shared Package | |
| ML20209B481 | List: |
| References | |
| I-83-13-NPS, NUDOCS 8702040467 | |
| Download: ML20209D369 (46) | |
Text
_ _.
~
- i..
4.a
,r
=
e ATTACHMENT D Page 1 of 44 GNS '84 03 01 150 840305X0187 &
TENNESSEE VA1. LEY AUTHOIRTY NUCLEAR SAFETY REVIEW STAFF, NSRS REP')RT NO. I-83-13-NPS
SUBJECT:
INVESTIGATION OF CONCERNS REGARDING QEB RECORDS j
l DATE OF 1
INVESTIGATION: MARCN I6 - MAY 6,1983 i
E l
1NVESTIGATOR:
l R. D. SMI y t
_ gyp--y g
DATE' INVESTIGATOR.
M
,, g -
R. W. TBAVIS /
g <, g 7 DATE' l
APPROVED BY:
K. w. WHITT jy h DATE
~
1 I
\\
MEDS, W5863 C.K
(_
G 8702040467 870129 PDR ADOCK 05000327 P
- - - - - - - - - - - - - - - - - - - - - - - ' ~ - ' ~ ~ ~
~
.'1
... i g
ATTAC10iENT D Page 2 of 44 e
TABl.E Of CONTENTS
.P,gg
- 1.
SUMMARY
n II. SCOPE '
2 III. FACTS 2
A.
Functional and Organization of QEB.
2 B.
Concerns Regarding QEB Records 3
C.
Quality Assurance Record Requirements 4
1.
Documents External to TVA.'....
5 a.
b.
5 ANSI N45.2.9..............'. J,.
6 c.
2.
Documents Interal to TVA i 6 TVA Topical Report 6
a.
b.
Interdivisional Quality Assurance. Procedures 7
D.
Implementation of Quality Assurance Record Requirements.
9
[
1.
Quality Engineering Branch 9
2.
Division of Construction Il 3.
Management and Engineering Data Systems (MEDS) 12 q
E.
Nandling Practices for Vendor-Supplied QA Records 13 1.
Quality Engineering Dranch 13 e
a.
Mechanical / Nuclear Unit 16 b.
Civil / Architectural Unit 17 c.
Weldments Unit 18 d.
Instrumentation Unit 19 e.
Electrical Unit 19 2.
MEDS 20 lsl J.
Divis' ion of Construction - Sequoyah Nuclear Plant 21 Y
4 Division of Nuclear Power - Sequoyah Nuclear ~ Plant 23 l
ll I
F.
Randomly Selected Contract File Flow j
{r..
23 1
IV. Analysis Q
26
.g A.
Availability of Complete Records at SQN 26 g
B.
Usability of Records by NUC PR.
30 C.
Ability of the TVA System to Meet j
the NRC and ANSI Requirements 31 D.
Safety Significance 32 i
e-ft s,
- r..
h ATTACHMENT D Page 3 of 44 TABLE OF CONTENTS (Continued)
{agg V. Conclusions.
33 VI. Judgement of Needs 34, ATTACIDtENTS
- - QEB Organization Chart
' - Summary of Contract Files Reviewed f
e 4
L I
i r
Ii lp
,a fl' l)i I
a t l
1 di d
l 3
i O
e Page 4 of 44'
~
ATTACleiENT D 1.
SUM 1ARY
-j q
4 As requested by the Assistant Urneral ttanager (Technical), the Nuclear i
Safety Review Staff (NSRS) investigated concerns regarding the inadequacy of QEB records. More specifically those records that are required by HRC to dociument that commodities requiring TVA source inspection for nuclear plants meet the design specifications.
The concerns were that the record system did not meet either the NRC or ANSI requirements, that control and review for accuracy and complete-ness of the records were less than adequate, and that the documents were not receiving a second and third technical review for technical accuracy.
I
'A S
QEB is responsible for assuring TVA that vendors supplying equipment h
i and materials provide those conrnodities in accordance with TVA's contract specifications through inspection of the vendor's QA program and the commodities manufactured.
QEB is responsible for collecting g
l the vendor QA documentation and providing it to the Division of Nuclear Power (NUC PR), the ultimate holder of the records.
- w In d
order to fully determine the extent of the problems relat'ing to 3
the alleged concerns, this investigation included an examination of the system for handling QEB records from generation to ultimate
)
storage. Sequoyah Nuclear Plant (SQN) was chocen as the ultimate 8
location because it was the newest facility on line and, according to the TVA Topical Report, all quality asssurance (QA) records were required to be ensite at the time of the investigation.
g 3
The investigation determined that the concerns were valid in varying Y
degrees for all of the QEB units.
It also became clear to the inves-tigators that prablems within QEB both affected and were affected by organizations other than QEB. In summary, the following problems were identified relating to QEB records.
The investig1 tion determined that, with what NSRS considers minor variations, the written procedures provide for the collection.
verification uf accuracy and completion, and transfer of vendor Q'A l
records to NUC PR in a manner which is in compliance with NRC and l
ANSI requirements. However, with inadequate implementation of those q
procedures, particularly across organizational boundaries, the com-i pleted sets of QA records were not getting to NUC PR.
Even if NUC PR 5
were to receive a complete set of records for a given product, in R
their current form the records would not be in a condition that would allow easy retrieval of information. These deficiencies place TVA In violation of 10CPR50, Appendix B, Criterion VII requirements for SQN; and unless changes are made to correct the problems, both Watts l
Bar Nuclear Plant (WBN) and Bellefonte Nuclear Plant (BLN) will be subject to the same violations.
a 1
At the time of this investigation, QEB did not know how many con-f tracts for SQN were complete, and therefore, associated QA document
)
records required to be onsite. They did not know with any degree of certainty whether or not QEB had all the vendor QA documentation 4
required to support fulfillment of the contract specifications; they 4
i
O
~
Articiaiz"r a e>5'"
e thought they did but could not verify it.
The only shipment of QA records to the Division of Construction (CONST) at SQN occurred in 1977 consisting of about 55 cubic feet. QEB estimates that it currently has over 500 cubic feet of completed contract records for all TVA nuclear plants stored in Knoxville and did not know how many cubic feet were in the regional offices awaiting shipment to Knoxville.
SQN CONST placed the 1977 shipment of QEB records in filing cabinets separate from the filing cabinets used by CONST to store vendor QA documents received with equipment. Current CONST personnel did not know they were there until NSRS enquired about them. CONST had MEDS film its vendor QA records which were received with various ve'ndor
~
shipments, but these were incomplete when compared to the reviewed QEB records. CONST provided that film to SQN NUC PR.
SQN NUC PR received the film and reportedly reviewed it for readability, llow-ever, according to a person in charge of CONST QA records, approxi-mately 80 percent of the CONST generated QA records for unit I are unreadable and the hard copy has been destroyed.
i Regarding the second and third level of technical review,.the investi-gation found that units other than the Electrical Unit were not per-forming them. Procedures require that the records be reviewed for accuracy and completeness by the unit supervisors. A third level of review is not required. NSRS believes this additional review is not g
necessarily required for all contracts but may well be beneficial on g
contracts with vendors having known or suspected problems.
The following report will show that the handling of approximately 700 million dollars of vendor QA records is inadequate and will require tens of man years of effort to collect, verify for completeness and j
accuracy, and organize and store in an easily retrievable manner
[
records for approximately 1800 completed contract purchases.
The present unacceptable condition of vendor QA records is attributed more to the lack of implementation of procedural requirements and
.3 d
intent rather than inadequacies in the ptocedures themselves. The prevention of future contract-required QA records arriving in the
[j same unacceptable condition is related to divisions and branches involved in the collection and flow of vendor QA data working together to reach a common goal beneficial to TVA and not to their i
own best interest.
II.
SCOPE i
Concerns regarding QEB records were investifated using standard inves-i tigation techniques and NORT analysis.
III. FACTS A.
Function and Organization of QEB
~
i QEB is charged with the responsibility to evaluate the quality i
control programs of vendors supplying commodities which require
[
5 source inspection (at the vendor's plant).
It is further charged I
l 1
J..
~
m ---
2-m a-
'... \\. -
h ATTACIDiENT D hPage 6 of 44 with the responsilillity of assuring that the commoditics pur-chased physically meet the contractural or purchase order speci-fications and to collect, maintain, and transmit to the user of the commodity documented proof that the commodity complies with the specifications.
Quality of the commodities is verified by TVA regional inspectorn located in eight regional offices through-out the United States, plus Knoxville, and commodities are not.
released from the vendor's plant until the inspector has docu-mented evidence that the contract specifications have been met.
The flow of those documents, bnth theoretically and in practice, are described in sections Ill.D and.E.
respectively.
Under the direction of the Branch Chief, QEB oversees the manu-facturing of commodities for TVA which require source inspec-tion.
A method of keeping track of the number of contracts and their value was established in the 1970s, and as of December 31, l
1980, QEB was inspecting 1,824 contracts valued at over two billion dollars. Since then, through tiay 10, 1983, 994 contracts f
were added and none removed.
Removal cf a contract f rom this list required the contract to be complete andsthe file tran3-y
[
mitted to the QEB Records Unit (NOTE:
In tiarch 1983 the per-g sonnel and functions of QED Records Unit were transferred to liEDS. For convenience and because the units had not been offi-cially renamed yet, the report will continue to refer to the QEB Records Unit.) or the site. Since no contract records have been sent to either location by QEB for several years, the number of active contracts and their dollar value may be greatly ir.flated.
The branch chief estimated that at least 80 percent of their work was nuclear related, and the vendors added manufacturing, g
testing, and documentation cost associated with producing a g
quality product with the required traceability added 64 percent to the normal cost of the commodity. Guaranteed quality usually requires closer attention to product processing, sometimes requires more time per unit produced, and can in some cases add to the cost of initial production. Ilowever, an ef fective QA/QC program more than pays for itself thrnugh increased reliability and performance.
Finding problems during manufactuirng is cost effective when compared to finding them at a later date.
1 Organizationally, QEB is divided into Field Operations (con-taining the eight regional ofilees) and Central Quality Control.
Staf f sizes, including clerical, for the Field Operations was 73
[
in 1980 and is currently at 39.
l'or QED Knoxville, including the Branch Chief's Staff and Central Quality Control, it was 33
[
b in 1980 and is currently at 31.
Central Quality Control is divided ir.to five unite each responsible for a commodity manu-y
(
factured under their speciality. These units are titled: weld-g(
ments, mechanical / nuclear (including NSSS), instrumentation, civil / architectural, and electrical (see attachment 1).
B.
Concerns Regarding QEB Records NSRS was requested by the Assistant C'neral flanager (Technical) y e
to investigate concerns regarding QEB records and to prepare a p
report documenting the investigation findings..
(
1, ' '.
~
ATTACIDiENT D l' age 7 of 44
--i O
O 3
The concerns were that the records system prnbably did not meet
-g HRC or ANSI requirements; that the vendor generated QA documents that come to TVA through QEB were not being reviewed for accuracy j) or completeness or being adequately controlled.
In support of I
these concerns, a room full of documents stored in the OED Cedar Bluff office in deteriorating boxes stacked one on top of the i
other with only a contract number on the outside was described.
The concerns further identified that records turned over to the QEB Records Unit by the unit supervisors were being indexed and g
reorganized by nontechnical clerical personnel; that vendor QA
} l docurnents were not being reviewed a second and third time for technical accuracy and cornpleteness by the other unit supervisors
, (
technical engineer; and that if documentation was
-j p or the contract found missing in a contract that had been closed out for several g
years, it might be difficult, if not impossible, to duplicate that
)
document from a vendor's file because in some cases vendors have sone out of business.
N.
i j
C.
Quality Assurance Record Requirements d
i The vendor generated quality assurance records that a nuclear 3
utility is required to accumulate and store are defined by a hierarchy of documents.
The control and handling of these j g
- g records is also described in the documents. These documents and their relative position in the hierarchy'of control are listed 4
below:-
j [
10CFR50, Appendix B, Criterion VII, " Control of Purchased Q
C Material, Equipment, and Services" and Criterion XVII, j>)[,
" Quality Assurance Records."
Regulatory Guide 1.88, " Collection, Storage, and Main-l '
tenance of Nuclear Power Plant Quality Assurance Records."
ANSI N45.2.9-1974, " Requirements for Collection, Storage, and Maintenance of Quality Assurance Records for Nuclear Power Plants."
Topical Report TVA-TR75-1, " Quality Assurance Program Description for Design, Construction, and Operation.
L Interdivisional Quality Assurance Procedures '(ID-QAP)
] [
Hanual for Nuclear Power Plants.
a d f ID-QAP 17.1, " Transfer of Quality Assurance Records for Design and Construction" i[
ID-QAP 17.2, " Quality Assurance Records for Design and Construction JI 4
1 t
o i
_ t. _
__.-~nr-
- - - - ~ ~ ' ~ = ~
~
, is.
e-l
~,'~;;.
O O
age 8 yt 44 ATTACIDiENT D P
i t
ja 1.
Documents External to TVA s
i 1
a.
~
- 5
l Title 10, Part 50, fppendix B of the Code of Federal Regulations is the controlling document for all, quality i
assurance activities for a nuclear power plant!
l 10CFR50, Appendix B, Criterion VII, requires, in part, that measures be established to assure that purchased reaterial and equipment conform to the procurement documents.
Also, these measures must provide, as app ropria te, for objective evidence of quality fur-nished by the vendor.
The documenting evidence that procurement requirements are met must be available at tre plantaite prior to installation or use of the materials or equipmeint. This documenting evidence i
must be retained at the nuc2ent power plantsite and sh,11 be sufficient to. identify the specific require-l meuts, such as codes, standards, or speci,fications met by the purchased material and equip.7ent. This crite-riot. has remained virtually unchanged since its issue in 1970 Criterion XVII requires, in part, Lhot sufficient records be maintained to furnish evidence that activi-s M
ties affecting quality were ap' ro'priately performed p
and that the applicant establish requirements con-W' %
cerning record retention, such as duration, location, and assigned responsibility, b.
Regulatory Guide (RG)'1'.38 I
The regulatory guides provide methods acceptable to NRC in complying with the Commission's regulations.
i RG 1.88, revision 2, 1976, specifically gives guidance for Criterion XVII and states in Part C, " Regulatory Position," the following:
a I
e i
i The requirements and guidelines for collection, U
storage, and maintenance of nuclear power plant q
i
)p, quality assurance records'that are included in
, j ANSI N45.2.9-1974 are acceptable to,the NRC staff and provide an adequate basis for complying with the pertinent quality rssurance requirements of l
Appendix B to 10CFR Part 50.
This statem-nt is unchanged from revision 0 issued in 1974 TVA was committed to the requirements of the regulatory guide with the followint exceptions:
For Browns Ferry, OEDC will use one-hour filing cabi-nets for temporary record storage and OEDC will pro-vide a two-hour rated vault for permanent record l
A
P gs 9 cf 44 ATTACIDiENT D storage.
For Sequoynh and Watts Bar, only the excep-tion that UEDC would provide two-hour filing cabinets for temporary record stnrage was taken. For Bellefonte, Hartsville, and Phipps Bend, the exception that OEDC would provide one-hour rated filing cabinets for tem-porary storage of records was taken.
c.
ANSI N45.2.9 Since RG 1.88 endorses an ANSI standard, the standard becomes the controlling document. ANSI N45.2.9-1974 states, in part, that:
i i
This standard provides general requirements and I
I, guidelines for the collection, storage, and maintenance of quality assurance records asso-l b
ciated with the design, manufacture, construc-g tion, and operation phase activities of nuclear power plants.
1 This standard delineates technical requirements; receipt of recorde methods; storage, preservation,' and safekeeping methods; retrieval requirements; and dis-position of records.
It does not specify where the records are to be physically maintained. This stand-i ard was not intended to cover the generation of records.
Also, it does not apply to activities covered by Section III, Divisions I and 2, and Section XI of the g
AStfE Boiler and Pressure Vessel Code.
2.
Documents Internal to TVA Each organizational level within TVA had developed a docu-ment which described the handling of QA records at that level. The TVA Topical Report as the controlling TVA 1.
document and each lower level document must adhere to that
[
' control. The Topical Report may be e:: par.ded upon and explained, but all commitments made in the Topical Report
[
5 must be satisfied. Most of the TVA controlling documents o
have been revised several timen since the records being l],
reviewed were initially crented. The handling of the records, l l i
of course, must adhere only to the controls in effect at documents were written for SQN, the NRC regulations have l (l the time of their creation. Since the first procurement been tightened considerably; but the basic requirement that li QA records be maintained onsite in an easily retrievable manner remains unchanged.
p ll a.
TVA Topical Report
(
Topical Report TVA-TR75-1 presents the TVA Quality Assurance Program developed for the planning, design j
and construction, operation, and maintenance of TVA l
i 1 ". l
~
e' ATTACHMENT D P
G 9
age 10 of 44 nuclear plants.
It is thus the highest level TVA document for controlling the activities being investi-gated.
Revision 5 of this topical report dated July 1982 has been accepted by the NRC and, revision 6 submitted to NRC for approval and draf t 1 of revision 7 remain essentially unchanged in the areas investi-gated. Section 17.1A.1.2.3, "EN DES Quality Engi-neering Branch," defines the organization and respon-sibilities of the Quality Engineering Branch. Section 17.1A.1.2.3 states, in part, "The Quality Engineering Branch (QEB)... la responsbile for determining that 1
the manufacturers and suppliers of equipment and materials for the nuelcar power plants fulfill the l1 technical and quality requirementr. as defined in the l
procurement specifications." Section 17.1A.17.1 states, in part, "EN DES QEB is responsible for sup-i 1
plier QC records on those contracts in which QEB inspection is required. This responsibility is to see that such records are sent and incorporated-into the HEDS system.
." Section 17.1A.1722 states, in part, " Construction QA Procedure, ' Quality Assurance Records,' contains similar provisions for records '
generated or received at the site during the construc-tion phase." Section 17.lA.17.3 states that "the transfer of OEDC QA Records to NUC PR is established q p I
in an Interdivisional Quality Asssurance Procedure.
The transfer system nakes provisions so that OEDC QA
-y records are readily retrievable by NUC PR at all L
times."
i b.
Interdivisional Quality
...urance Procedures The 1D-QAPs allocate responsibilities to the different divisions within TVA for the various quality assurance functions.
ID-QAP 17.2, Section 1.0, " Objective and Scope," states:
This procedure establishes the guidelines and identifies the interface responsibilities 1br controlling quality assurance records within OEDC.
J Section 2.1.2.d under " Responsibilities, Division of Engineering Design," states:
Furnishing CONST with evidence that purchased items conform to the procutement requirements (attachment 2). to ID-QAP 17.2 contains a listing of the
" Records Required at the Site Prior to Installatinn or
]-
L Use."
This attachment references the "McAdoo Letter" i
which was a letter to J. D. McAdoo, Manager of Licensing Y1 h* -
- - - - - - - - - - - - - - ' - ~~~
~~ ~'
', " * ~
'1 ATTACIDiENT D Pcge 11 of 44-O O
and Reliability, Vestinghouse Electric Corporation, from C. K. Deck, Deputy Director of Regulation, U.S.
AEC (now U.S. NRC) on May 11, 1971. The letter con-tained, as attachments, " Guidance for Criterion VII Regarding Certification Systems" and the " Westinghouse Implementation of AEC Criterion VII Guidance Regarding Certification Systems." The body of the letter itself contains the AEC criticism of the Westinghouse imple-menting document. There is no description of how Westinghouse modified its program to meet the AEC requirements. Attachment 2 of ID-QAP 17.2 describes how TVA is implementing the AEC guidelines.
It does this by referencing the "McAdoo Letter" and adding some special requirements. Under Special Requirements it states that "for Code Material a certified Mill Test Report on Certificate of Conformance and any required radiographs are required."
The Westinghouse certification system had.three docu-ments:
1 (1) a Quality Assurance Release (2) a copy of the purchase order fl (3) a copy of the equipment specifications 2
TVA meets parts 2 and 3 by stating these items are I
i included in the procurement documents supplied to the y
construction site by the procuring organization. Part I was met by its being supplied by the vendor or by QCB.
TVA did not go into the degree of detail in describing its system as Westinghouse did, it did not address the criticism that the AEC had of the Westing-f I
r house description, nor did it have an NRC or AEC l
criticism or neceptance of its program. The Westing-0
'g house three part system was acceptable in its bare f
form, but the detailed method of implementation of the system was not acceptable to NRC and the "McAdoo y
1.etter" described what Westinghouse had to do to make p
it acceptable.
g ID-QAP 17.1 describes the method of transferring records from EN DES and CONST to NUC PR.* NUC PR is the ultimate user of the records being collected and generated and, therefore, must be in s'ponition to use those records. The responsibilities of EN DES include receiving, storing, maintaining, indexing, and trans-ferring to NUC PR supplier QA records (except receiv-ing inspection reports) required by contract specifi-CONST responsibilities include receiving, p
cations.
i storing, and maintaining supplier QA records. CONST c
also arranged for the transfer of QA records to NUC PR E
as required by the ID-QAP. These QA records, of l
course, include much more than vendor-supplied records.
q 51
s
. ATTACHMENT D Pag 2 4 0 **
ll>
ll>
NUC PR was required to accept the QA records from CONST and EN DES and maintain them.
MEDS had the responsibility of coordinating microfilming and dis-tributing EN DES and CONST QA records to NUC PR.
D.
_1mplementation of Quality Assurance Record Requirements After the responsibilities of the various organizations were defined by the ID-QAPs, each organization then had to hav.e a means to implement the requirements allocated to it.
OEDC used the Quality Assurance Program Requirements Manual for Design, Procurement, and Construction (PRM). Under this, each division issued procedures. EN DES used Engineering Procedures for this purpose and CONST used Quality Assurance Procedures and Construc-tion Procedures.
PRM 17Q PR-1, " Program Responsibilities, R1, November 15, 1982 (R0 issued June 29, 1982), assigned responsibilities.to EN DES and CONST.
EN DES responsibilities included maintaining sup-plier QA records until turned over to MEDS and ' assuring that supplier records were identifiable and. retrievable and in a system compatible with NUC PR and CONST requirements. CONST responsibilities included assurance that CONST and contractor records are identifiable and retrievable during the construction phase. CONST was also required to maintain supplier records g
until turned over to EN DES (MEDS).
B j
PRM 17 QPD-1, " Management and Engineering Data Systems (NEDS),"
RO, April 2,1979, stated that MEDS was responsible for all phases of the records management function and ensuring that all y
records program activities were in accordance with industry i
codes and standards and Federal regulations.
I It also stated that docaments which furnish evidence of quality of items and/or activities affecting quality or documents which have expected retrieval significance for any OEDC organization or for TVA, are sent to MEDS.
No mention was made about distribution of microfilm.
1.
Quality Engineering Branch Activities of QEB in the area of vendor QA records were controlled by the following division-wide engineering procedures.
EN DES-EP 5.4, " Release of QA Items from Supp13ers' Shops to Construction Site," R6, April 1,
- 1982, and the branch-wide engineering p;ocedures QEB-EP 24.11,
" Inspection - Ceneral Instructions," R0, March 4, 1983; QEB-EP 24.37, " Produce Cor.pliance Data and Quality Engi-neering Records and Dispusal," R0, January 22, 1982; and QEB-EP 24.58, "llandling uf Supplier Records," R1, September 1, 1982 (R0 issued February 9, 1979).
It should be noted that the issue dates for two of the three branch engineering procedures were in 1982 and 1983.
Prior to 1982, there was very little in the way of a formal
_g.
ATTACIDiENT D Page 13 of 44 O
e program in effect for the Knoxville office of QEB. This was verified in conversatfort with the Supervisor, Field operations and others within the QEB organization.
a.
EN DES-EP 5.43 This EP described the method for setting up a contract file at the regional offices and the Knoxville office.
It described the duties of the regional inspector and of the QEB personnel in Knoxville.
One duty of the TVA inspector was to verify that
" performance test reports are accurate and complete and that documentary evidence exists that the items conform to procurement requirements, including cali-bration of test tools and equipment." Also, the inspector " handles" supplier test reports and other data" as follows:
(1) Sends cable test reports, performance test reports, and other data that require QEB or technical engineer approval to Central QC Knoxville, as they are obtained.
(2) Requests the supplier to send data packages for structural and piping loose material and code data report forms for items which require ASME Code, Sect'.on 111, N-symbol stamping to the Jobsite with the shipment, unless otherwise specified by the contract.
(3) Handles other vessels, components, and equipment
?
items in accordance with the contract require-ments or as directed by the Chief, QES.
(4) At the end of the contract or purchase order, U
sends the data files to Central QC, Knoxville,
~g for microfilming unless otherwise specified by the contract or instructed by QEB.
-q
-g 9
b.
QEB-EP 24.1I j
This EP was issued on !! arch 4,1983, and replaces the
[
TVA Inspection Manual. The procedure is a general instruction on the practice of inplant inspection A
(source surveillance) by TVA.
In section 7.3.7,
" Final Reports," it is stated that, " Final reports are normally used to close out a contract file. The Central QC Staff reviews the file for completion after distribution of the final report."
e ATTACIDiENT D Page 14 of 44 O
O QEB-EP 2_4._37 i
c.
Under Section 1.0, " Purpose and Scope," it was stated that "this EP covers the indexing and preparation of Product Compliance Data and Quality Engineering Records produced and received by QER."
Under Section 2.0,
" Policy," it was further stated that "the QEB is responsible for the storage of Product Compliance Data and Quality Engineering Records during the manufactur-ing stage of contracts set up for inspection."'
In-
~
cluded in these responsibilities was the orderly transfer of records from QEB to ifEDS for transmittal to NUC PR.
In section 4.1, it was stated that the QC Unit Super-visor " accumulates Product Compliance Data and Quality Engineering Records by project; ensures that the data / records are complete and accurate.",The QC Supervisor was also required to turn,over-records to the QEB Records Unit.
By this EP, iiEDS microfilms the data packages and sends a copy of the microfilm to QEB and to the re-g lated plantsite.
jg i
d.
QEB-EP 24.58
)
In Part 1, " Purpose and Scope," it was stated that i
"this EP outlines the methods that QEB uses in handl-j l;
ing supplier records for TVA contracts requiring
),
i inspection by QEB. This EP applies to supplier i
records submitted to QEB to furnish proof of com-l I
pliance with the requirements of the contract."
The EP also instructed the TVA inspector to prepare a j i detailed index identifying the number of packages or l
boxes, what records are contais.ed in each package or box, and to what piece of equipment the records apply.
!h
[
The inspector was then instructed to send the records to Knoxville. The Knoxville unit supervisor was q
directed to receive, review, and check the records J{
against the original records checklist.. He was then instructed to further index the records, if necessary, I
and maintain them until turned over to,the QEB Records l
i Unit.
l
~
l d
2.
Division of Construction (CONST) 0 The CONST Quality Assurance Program for records is set forth in QAPP-17, " Quality Assurance Records," RS, June 30, Very little guidance is given in this policy state-j 1982.
g-ment concerning vendor generated QA documentation.
It does, however, state that QA records shall be identifiable I
I 1
~~~ ~
~ ~ ~ ~ ~ - '
~~'
ATTACIDiENT D e
@Tage15of44 and traceable to the item or activity, they shall be retriev-
=
able, and they shall be maintained in a systematic and,
E controlled manner.
=
At Sequoyah Nuclear Plant, SNP Construction Procedure No.
P-8, " Quality Assurance Records," R16, February 18, 1983, 4
i described the manner in which QA records are prepared, reviewed, handled, classified, temporarily stored at the construction site, and transferred to NUC PR.
1 In Procedure P-8, Scetion 6.C.4, "QA Records Originated Offsite," it was stated that "QA records received at the site which were prepared by vendors or other TVA organiza-5 4
tions shall be incorporated into the site cont ract file."
This requirement was established in revision 13 dated November 11, 1977.
l 1
d in Procedure P-8, Section 6.C, " Transferring Records to NUC PR," part 2, it was stated that the Quality Control and
(
7 Records Unit (QCRU) will arrange for MEDS to microfilm all QA records with manual folder level indexing.
Also,' the h:
microfilm was to be returned to QCRU for transfer to NUC PR.
l {
Nonmicrofilmable QA records were to.be transferred to NUC PR l
1 in hard copy. In section 6.C.4, it was stated that the 3 !
6 transfer of QEB supplier radiographs will be made by EN DES.
3.
Management and Engineering Data Systems (MEDS) i lI The eperations of MEDS was described in their procedure book.
In MP 13.01, "Backfile Documents - Processing," R0, l
March 25, 1981, it was stated that MEDS receives requests k
from user organizations to film backfile documents.
MEDS then goes through the mechanics of filming, including a setting of priority using attachment 2 (this attachment was i
not available). After filming, MEDS distributes index and I
film to organizations requesting copies.
l MP 12.07, " Processing Site-Originated Quality Assurance Ir Records - Sequoyah and Watts Bar Nuclear Plants," RO, March
]
I 4, 1982, described the mechanics for bringing the records from the construction site listed to Knoxvil'le to be filmed.
One ' responsibility of the MEDS user representative was to k
deliver the index, all illegibles, and the applicable
'gij number of copies of microfilm to the site QCRU for transfer dl to NUC PR.
i[
in MP 14.01, " Document Proccasing - Overview," R0, Narch 4, 4
I j)[k 1981, it stated that for documents processed in the MEDS online system, MEDS received documents from user organiza-tions prepared in accordance with instructions in MEDS
]
Handbook or related EPs, CEPs, and Als.
J I
12-
t g Page 16 of 44 ATTACIDiENT D E.
Ilandling Pract ices for Vendor-Suppl,i_cd QA Records Discussions were held with members of QEB, MEDS, CONST, and 5
NUC PR to ascertain actual pract ices in the flow path of t>an 'u QA records from the vendor to the ultimate user of the commodity
-^
and ultimate holder of the offical records--NUC PR.
The flow of records to SQN was chosen because it was the newest operational
]:
nuclear plant in the TVA system and the last unit had been in i
commercial operation greater than six months, thus required by S
the TVA Topical Report to have all QA records onsite at the time of this investigation. Areas covered in the investigation 5
g process included accumulation t,f records, review for accuracy and completeness, storage, indexing of contract files, and 6
transmittal of the records. The actual review of randomly selected completed contracts is described in section III.T.
1.
Quality Engineering Branch F
i j h As described in sections Ill.C. and.D above, the overall
} j handling of vendor QA records has followed an evolutionary il process. Members of the QEB organization stated that their files were the official files during the life of an active j
r contract and until the file was transmitted to another I
group. The branch chief stated that the QEB files lose j
their offical status when the files are transmitted to MEDS j f for filming and the film has been reviewed by QEB. QEB
}
personnel stated NUC PR had told them that NUC PR did not know what they wanted or needed in the way of QA records d i and that NUC PR did not want and could not physically store i S hard copies of QEB's records. Correspondence on this prob-lem was generated in 1977. Ultimately some QEB contract g
i files were shipped in 1977 in hard copy to SQN CONST, and each file included every document QEB had at the time.
Apprently no ef fort was made to eliminate non-QA informa-tion or duplication.
gy The problem of what the required QA docttments would consist of for any given commodity also presented a problem to QEB.
Z 1
i Personnel stated that contract specifications, until very 3i f recently, were vngue regarding the required vendor QA
(( g documentation. This vagueness was described as resulting g.
in the regional inspectors having to negotiate with each 5
g vendor about what records the vendor was willing to supply, t j Therefore, documentation for like commodities varied from g
vendor to vendor and may only include a vendor certificate y j of compliance or may include as much as a data package, including certification of compliance, mill certifications, test reports, etc.
p M
There was no practice of routinely transferring QED records to CONST, and it appears that QEB records were transferred at only two time periods when they were running out c.f file space in QEB.
-. ------- -- - - - -- -- - -----~ - - ' - -- -- '~ ~
~~
~~
~
l E.
ATTACIDiENT D O
O cgs 17 of 44' P
s 8
4 In about 197:1 a van (one pernon recalled it being a 12-foot moving van) wan filled with QEB records for BFN and trann-
_I ferred to BFN.
In 1977 two shipments of hard copy records g
were. transferred to SQN CONST by QED af ter being filmed by 2
HEDS. The QEB Branch Chief said CONST could not physically accept records before then. Other QED personnel recalled that they were out of ille space and needed to make room.
They further recalled this was a hurry-up operation requir-ing overtime work, and the files were not reviewed for completeness.
One shipment n! Hechanical/ Nuclear Unit recordn to SQN 3
CONST consisted of 27 hoxen (presumed by NSRS to be 1-foot i
per box because this was the most common size) containing 377 completed contracts. These contract files ranged in
-}
size from very small, 116 contracts contained in 3 boxes, 4
to very la rge, I contract contained in S boxes. There was no documentation of transmittal produced for this shipment; however, SQN CONST had a copy of the index produced for the transfer with written-in file drawer locationn.
+-k Another shipment was made on October 19, 1977, from.the 3
design project manager to the SQN Project Manager.
This 3"
shipment consisted of 28 boxes containing 348 contracts. As with the previoun shipment, QEB stated all flies were
. filmed by MEDS before shipment. Copies of the film were g
possessed by MEDS and QEB and NUC PR Central Office, g
Indexing for these films was manual, not in the MEDS data
~
base, and included the contract number, vendor name, and a very general description of the commodity. SQN CONST had 3
no reccid of this shipment but found the records associated with these transfers after the NSRS inquired about their 54 disposition.
It was helieved by CONST that the first f
shipment was inenrporated into the CONST filen developed g
for the same contracts, but they later determined that the
)
records were not incorporated.
A similar transfer uns made to Bl.N in 1977. No record of 3
any t ransfer of QEll records to WilN could be found. There was no recorded transfer of QEll records to SQH since 1977; w
.however, both QEll and SQN CONST personnel remember that several boxes of contract filen were sent by QEB to CONST 3
in the 1979-80 time period nnd returned becabse CONST said t' hey did not know what to do with them.
-j QEB maintains a enrd file on cont racts requiring QEll source inspection issued since about 1965. On these cards such Q.
information is maintained as contract number, vendor, C
f responsible unit nupervisor, whether the contract is active or complete, and if the records file has been transferred.
m QEB also maintained a computerized data base on all active contracts f rom which completed contracts were removed af Ler the contract has been completed fbr about three months.
4
- m. g,
' ATTAC1DiENT D P
~
O O
age 18 of 44
~
There was no list other than the card file for completed contracts. QEB is present ly ih veloping such a list. As such, it would be ext remely di f ficult to identify all i
completed contracts for a given plant. Only one unit supervisor, Instrumentation, maintained such a list.
=
Cotoparing his list of contracts, which also contained disposition of record inforraation, against the list of records Pent to SQN in 1977 revealed several interesting
]
points. Of the 248 completed contracts listed by the unit
-i supervisor, only 172 or 69 percent have been sent to CONST.
Tj Of those contract files still in QEB's possession, some had n
been reopened, which was conrnon practice to avoid issuing a G
new contract, but the records for the commodities produced 5
under the contract fi rs t issuance remained as well. Within 3
the six boxes of the Instrumentation Unit's records sent in 1977, one contract--76K52-820187-4--with Standard Pipe &
y
_=
Supply Company was shown on the QEB index sent with the records. This was not a contract on the Instrumentation Unit supervisor's list nor was it on the QEB list of filmed SQN records. SQN CONST could not find that: contract with j
the other QEB recordr..
N Comparing the list of filmed contracts for SQN against the h
QEB list of transferred contract files revealed that four i$,
contracts were not filmed, contrary to QEB's belief (73C55-
] j 83535-5, 74C56-83612-8, 73C35-83571-2, 74C57-85443).
l.
Another discrepancy found was that the contents of boxes 21 j
and 28 shipped on October 18, 1977, had identical contents.
} l Of the records sent by the Mechanical / Nuclear Unit in 1977,
$l 3
l l
there are several discrepancies which appear to be typo-graphical errors between the listed contract numbers on the A
hard copy sent to SQN and the filmed list. Of-more signi-ficance, two boxes of records (boxes 1 and 2 standard com-ponents) believed to be shipped by the unit supervisor had not been filmed and could not he found by SQN CONST. An additional four contracts on the list of contracts sent 3
t were not on the list of filmed enntracts (74C56-83612-8, M l 74C54-83635-6, 74C54-83635-8, 74P63-14703). These dis-crepancies were communicated to the QEB Branch Chief for l
j his action.
y The Chief, QED, said they would like to turn 'over completed u ; j filea to MEDS for their filming and distributinn three I
months after a contract was completed. lie further stated f l(j that MEDS could not keep up with the volume of QEB records.
A completed contract to QEB meant all the compcnents required
_ )
on a contract had been shipped from the vendor's site and j
did not necessarily mean that TVA had all the paperwork.
In the past it was QEB's respunsibility to transfer their y
records to CONST and the branch chief delegated that
$[
=
respon-sibility to his unit supervisors.
In the opinion of the l l' Tield Operations Supervisor, QEB ifoes not have to maintain
- l 1 'G 3
~~
9 Y
W%
a r
c gg 3
p e
G age 19 of 44 their records in an indexed or easily retrievable manner indexing and organization for retrievability is required only when the records are transferred to HEDS.
4 With regard to radiographic (RT) film produced by the vendors during their tests, QEB stated they do not want to handle it because they don't have the storage faellities for those kind of records.
RT film until asked for or send it with the equipment.The vendor w In general, QEB personnel were unclear as to the disposition of RT film and when, how, or if TVA received possession of
~
it.
Of the contracts reviewed by NSRS, the contr'act lan-guage regarding RT film was equally vague.
stated that at the completion of the contract the filmIt specifically became the property of TVA, but did not specify how or if TVA would take possession.
Hechanical/ Nuclear Unit n.
This unit consists of a unit supervisor and an SD and an SE responsible for all NSSS and mechanical contracts. For the NSSS contracts the vendor is responsible for collecting all QA records and turning them over to TVA at some predesignated time.
With the TVA direct contracts under this unit, the vendor QA records are (1) maintained by the vendor until told to release them, or (2) collected and maintained by the regional office until the Knoxville QEB office requests they he sent in, or (3) period-ically sent by the regional office to Knoxville.
Because of the vagueness in consensus standards and contract specifications regarding required QA docu-mentation, the unit supervisor stated it was difficult to have any consistency in records. Skid-mounted equipment, for example, was described as having no governing consensus standard, and considerable negoti-ation by the regional inspector was required to estab-i lish what QA documents the vendor was willing to produce.
In some cases this could result in an in-crease in the contract cost.
The unit supervisor stated he relied heavily upon the regional inspector and his ability to verify that the commodity met contract specifications and that the documentation was complete and accurate. Hanpower within this unit was described as insufficient to adequately review the QA records as they arrive; spot checking is the best they could do.
He stated that some one-of-a-kind records are maintained by the vendors until asked for, but was unable to specify how the TVA request for these records would come about.
~--~~-~ ~~~~
r b
ATTAC1DiENT D P
~
O O age 20 of 44~
The supervisor stated that approximately 200-300 activecontractswerebeingadminiugeredbyhisunit.
Furthermore, approximately 140 feet 4
of records on completed contracts for that unit were being stored in boxes in a two-hour fire-rated room at QEB.
?
The boxes were stacked on top of each other with a contract j
number on the outside.
visor that approximately four man years would beIt was estimate required to go through his completed files, review them, and put them in order.
-' q With regard to the 1977 shipment of records to SQH, the supervisor recalls it was done very rapidly and that he personally took a van load down to SQN CONST.
No emphasis was placed upon reviewing them for accu-racy or putting the records in some logical order.
-]'
They were indexed only to the contract or file level.
A copy of the index was' supplied to the investigators.
The supervisor believed that the recordsfwere complete but could not be certain.
Records from his group were to be transferred when a contract was complete to the QEB Records Unit for filming and transmittal to CONST. He said that the Records Unit was not able to keep up with the records and the unit supervisors were told to hold on to their records.
b.
Civil / Architectural Unit This unit consists of a supervisor and two technical staff members recponsible for administration of con-(
tracts for commodities which correspond to what their a
As with the Mechanical / Nuclear Unit, j
title implies.
concensus standards and contract specification were a
described as vague in some areas regarding required M
vendor QA documentation.
This results in negotiations j]
with vendors regarding the QA documents they will supply.
-J 1
For some contracts administered by this unit, bulk quantities of structural steel, piping.
documentation and mill certifications we,etc., the QA re sent with the commodity to CONST at the site, and QEB did not receive a copy.
For other vendor-produ' red commodi-ties,.QEB received the QA documentation from the regional inspector.
If more than one copy was pro-vided by the vendor, then a copy was sent to the site with the commodity.
for SQN, however, the unit supervisor stated that only a
QEB received the QA documentation; the site did not.
This unit supervisor relies upon the regional inspec-tor for assuring the QA documentation is correct.
As., &
E
- *e * * *
'?.
w O
ATTACIDiENT D O
ge 21 of 44 Pa the contract is filed until the the contract is complete.information They were reviewed for completeness and turned over to the Records Unit when they could receive them.
The supera visor stated that they were approaching 100 percent review of all files sent to the QE8 Records Unit.
- l However, he said the Rec'rds Unit had a backlog of l
o records and none had been sent in about a year.
Currently he estimated that his unit had between 3
100-200 feet and the regional offices had aboutof completed file records in K
~ ~,
200-300 feet to be shipped to QEB in Knoxville. lie estimated three to four man years of work would be required to review and index the files he currently has before sending them to the Records Unit.
He said the Records Unit "unfor-tunately" is indexing the files they currently have and dosen't believe they. have had sufficient guidance to do it properly, lie stated that he would like to review the film ifEDS makes but that was the Record Unit's job.
This supervisor stated he sent some records, he could not remember which ones for sure, to SQN CONST in I
about 1979 or 1980 but they were returned.
-)
There was no record of transmittal found by QEB.
l
{
a Weldments Ung c.
J This unit, under the direction of a supervisor and one 4
engineer, adminis.tered contracts involving the welding
- h y
fabrication of large components. This unit was created out of both the fiechanical/ Nuclear and Civil / Architectural
=
Units and was not in existence in 1977.
With regard to Q4 records, this unit required one copy
~
j 1-of the vendor data package to be r. hipped with the convuodity and required the of fical QEB copy to be maintained by the regional office. Approximately three j
months af ter the completion of a contract, the Weldments Unit Supervisor would request the records be sent from the region to Knoxville. Reviewing his completed contracts for SQN reveeled one, unit's list of k
completed for over a year that had not been eslied in.
Prior to a record shipment, the region was to index the file.
According to the unit supervisor, additional indexing was performed in Knoxville by QEB and finally ifEDS.
lie further stated intelligence was prepared for files describing the filing system to aid retrievability.
1 I
{
~
ATTACIDiENT D ege 22 of 44 Reviewing the recently completed file on the Bellefonte containment, filling five cabinet drawings, transferred to the Records Unit; it took about three hours to j
track down the pedigree of a component for the polar crane using the combined talent of the person indexing the l'ile, the unit supervisor, and by telephone the TVA resident inspector at the vendor's site.
9 d.
Instrumentation Unit This unit, under a supervisor and one engineer, admin-
~~
8.sters contracts for instrumentation requiring source inspection. This was the only unit that maintained a working document listing all contracts handled by the unit including the disposition of the contracts' records.
Unlike the previous three units, the Instrumentation Unit contracts were described as more specific regard-ing the required QA documentation. This unit required the QA documentation to be sent to the site with the conssodity and to QEB Knoxville as it was generated.
The documentation was reviewed for completeness in Knoxville, but not technical accuracy. The regional inspector was relied upon to provide the technical revleu necessary to assure a quality product. The regional office also maintained files on the contracts and was considered the duplicate file for this unit.
3 This unit had some contracts previously completed that were reopened to purchase additional equipment.
In those cases the documentation for the previously l
i completed portion remained a part of the contract file.
p I
The unit supervisor stated that the QCB Record Unit was unable to accept any records for the last two years so they had been building up in his files. He also stated that the Records Unit clerks, for the most part, do the contract indexing.
e.
Electrical Unit This unit, under a supervisor and two ' engineers, administers contracts requiring source inspection l
under their purview. Arcording to the unit supervi-j l
sor, like the Instrumentation Unit contracts, the F
3 specifications in contracts regarding QA records are adequate.
4 1
t
)
Regarding QA records, this unit supervisor required I
I l
the regional inspector to review them for completeness and accuracy. When natisfied, the regional inupector would send eight copies to QEB Knoxville where they __ _
_ _ _ _ _. _ _ ~. - - -
J
- 0
~
Page 23 of 44 ATTACIDiENT D O
O were revirwed again for completeness and accuracy.
When satIslied, the Electrical Unit Supervisor would then send seven copies of the QA records Lo the techni-cal engineer for review, approval, and distribution, f
This nuit supervisor expressed concern regarding the transfer of his records to the Records Unit. Ile felt they were not capable of adequately handling them, lie further stated he had never sent any records to the Records Unit and was reluctant to do so but in about 3
1979, seven feet of SQN and seven feet),of VBN records were taken from him and sent to the Records Unit.
i a
QEB had recognized that a problem existed with their records and a number ni actions and program changes had begun.
The inspection load of the regional offices had decreased and those personnel were being used to review files for com-h pleteness. A recent procedural change to procurement specifications will require future vendors to supply a list of QA documents they will provide to show' compliance with the contract specifications. QEB was also developing a list of all contracts they had handled since 1965, includ-ing the disposition of the contract and location of the N
contract file. A data base for QEB records had been devel-ll oped and was in the testing and evaluation stagen.
gl 2.
ME,D'S D
A discussion was held with the MEDS Branch Chief regarding the role of HEDS in QEB records. He stated that once the records were filmed and indexed (for either computer sided or manual retrieval) by MEDS, they became the officiai documents. He unid t. hat for QEB and CONST QA records, MEDS l
performs the filming and QEB and CONST were responsible 'or distributinn of the film to NllC PR.
lie staterd that con-siderable effort had been expended in developing a data base for QA records which could be queried on the document level.
This system would include both QEB and CONST records.
When asked if HUC PR would have access to this, he stated that POWEH in Chattanoogn had a MEDS terminal and could use it if they'wnnted to, but their experience had been that I
NUC PR did not want HEDS material because it was not in a form that NUC PR enuld use, lie also expressed the position that MEDS was inr OEDC not l'0VER, and MEDS would not put
~
7 the records in a form NUC PR would like. The branch chief stated that.the new data base would be used on currently produced QA records and older QA records would be added time permitted, probably Luking several years.
as With regard to a quention on duplication of records, a task force was identified by the branch chief which was review-ing records produced by OEDC and specifying on a specific
)
hasis which groups of records were 'to be retained and which j ;
f 9 57
--d
[,.
- e J
4 ATTAC1DiENT D Pagn 24 of 44' O
e were to be discarded.
Copics of the task force reports for SQN CONST and QEB were reviewed and no effort to reduce duplication between the two nets of records could he found.
The task force that reviewed CONST records did not have a representative from NUC PR, the division that would receive the retained CONST records.
NUC PR did review the task force report and determined there were records scheduled for destruction that NUC PR wanted to retain.
NUC PR requested that they be included in the transfers and they were.
Interviews were held with the QEB MEDS representative.
The representative stated that MEDS would index completed con-tract files to the package level.
In doing so MEDS would look for any obvious errors and, if found, would send the ille back to QEB for correction.
A rough estimate of the MEDS vorkload in QEB for all' QED-inspected contracts was made by the representative as follows:
Completed contracts in QEB Records Unit i a.
3 150 ft b.
Completed contracts in QEB Materials Engineer possession 3
520 ft Open contracts in Materials Engineer c.
i possession 3
345 ft The MEDS representative assumed one cubic foot of files would fill a 2300 image roll of film, and two cubic feet of records per week could be made camera ready by the three MEDS clerical persons in the Records Unit.
Using these assumptions, 19 person years for MEDS Records Unit person-nel alone conid be expended on currently completed records.
This does not include the files held by the regional offices that are not duplicates of QEB Knoxville file material or records found in QEB files previously sent to the TVA records center which require filming.
Discussions with QEB Records Unit clerls revealed that l
they had been expected to index and organize coropleted contracts without sufficient guidance. They stated that they were just expected to know how to index.
they did the best they could.
Consequently, to document a procedure for indexing.One clerical person attempted This was the only guidance, described by them, they had to work with.
3.
Division of Construction - Se,guoyah Huelcar Flag In addition to CONST generated QA records involving CONST activities at SQN, large quantities of vendor generated and TVA-related QA documentation is accumulated.
Vendor gener-ated documentation includes that which is provided as a part of contracts requiring QED nource inspection.
CONST personnel interviewed indicated that they believed their records were complete and QEB rechrels only duplicated what. _ _
- ~ ~'~~ ~
~ ~ ~ ~ ~ ~ ' ~~
.m._o
.+
ATTACHMENT D Pags 25 of 44 O
O
~
they had.
Personnel recalled that several boxes of QEll records arrived one day (in the 1979-1980 time period), and they did not know what to do with them. Personnel recalled glancing through the recorda and saw duplication nf those records CONST had already. They indicated it would be a large effort to try to incorporate the two sets of records, and since they saw no value in doing so, chose not to.
CONST ultimately sent the records back to QEB in Knoxville.
Questions about records shipped to them in 1977 from QED produced some consternation because those currently involved with CONST QA records were not involved in the 1977 records transfer.
Documentation was produced which indicated the QEB undocumented shipment of records by the Mechanical /
}
Nuclear Unit Supervisor in 1977 had been received und i
persumed incorporated into the CONST files. This van j
arrived at by CONST from the fact that file drawer numbers were written next to each 'cox on the index supplied by QED.
j CONST was informed that based upon the NSRS review of the
'j film of CONST QA records, it did not appear that QEB records had been incorporated into the CONST files.
No record could be found of the formal October 1977 shipment from Pierce to l
Stack.
l NSRS provided CONST with a copy of the transmittal and index.
A few days later, CONST informed NSRS that QED records shipped in 1977 were found still in file cabinets in another building. Reviewing what was found against the indexes, CONST determined that two boxes were mssing.
NSRS informed QEB of this discrepancy for their action.
A large portion of the CONST QA records had been filmed (over 200 rolls of film) and the film given to NUC PR at SQN.
CONST stated they did not have a machine to review the film produced by MEDS and, therefore, did not know if the film was complete or readable.
CONST just gave the film to NUC PR.
CONST informed NSRS, based upon their use of the film at SQN NUC PR, that about 80 percent of the film for SQN unit 1 CONST records was illegible and the
{
original records had been destroyed, i
in a discussion with the CONST Haterial Inspection Grnup (NIC) Supervisor, he stated hat each piece supplied under a contract requiring source inspection is inspected against contract specifications, including the presence of required vendor and QEH documentatinn.
If the vendor documentation or QEB release for shipment form is not present, the com-modity is rejected unless QEH verbally authorized its acceptance by telephone.
The CONST QA records clerk stated that vendors sometimes put manufacturing and test data in equipment manuals sup-plied. These manuals were given, according to CONST, tu SQN NUC PH in hard copy; they were not. filmed.
1
, 1
---"'~~ ~
~ - ~ ~
~
ATTACHMENT D Pega 26 of 44 With regard to vendor radiographir (RT) film, CUNST stated they had not seen any film come onsite in years.
CONST stated they recieved about two boxes of Chicagn Bridge and Iron RT film for the containment from QEB and it was not indexed.
4.
Division of Nuclear Power - Seguoyah Nuclear Plant In discussions with SQN NUC PR personnel, they stated that they received QA records on film from CONST and document control had custody of it.
The filmed records received admittedly were not reviewed for accuracy or completeness by any of the QA or engineering groups. They were reviewed only for readability by document control. Personnel at SQN HUC PR stated they did not know what should he in the QA records they receive. They further stated that they had a difficult time finding anything in them and generally CONST personnel were requested to go through the film to find something for NUC PR. A manual index in loose-leaf form was provided with the filra. The indexing'* level,was generally on the folder level, i.e., contract number only.
The largest contract observed by NSRS was contained on 13 rolls of fil!n generally indexed to the contract number and sorne other unidentified lower-tiered number.
Film was stored by document control in two locations--the vault and in the document control file area in the adminis-tration building. The offical files were considered to be the film and any unfilmable hard copies from the records filmed and any other hard copy records were considered to be duplication. Test records not found in any other file system at QEB CONST or NUC PR were found in what was con-sidered the duplicate hard copy file at SQN.
Vendor manuals transferred by CONST to NUC PR are con-sidered by SQN NUC PR to be for information only.
The offical vendor manual was supplied by che NUC PR Central Office.
Vendor RT film r.eceived by document control was sent to the U. S. Government Record Center in East Point, GA.
- Records, generally by system or component, were maintained by docu-ment control specifying what RT film had been sent to storage. The list was quite small and included only main coolant components.
Construction produced RT film was stored in the NUC PR and CONST vaults.
F.
Randomly Selected Contract File F, low Methodology Af ter studying the operationn of QEU, it was decided that since NUC Pk was the uit imate user of vendor generated QA records, a - _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _ _
ATTACHMENT D Page 27 of 44 O
e verification of records actually under the control of NUC PR would be performed.
In order to do this, SQN was selected as the NUC PR facility on which to perform the record check.
First, contract file records were selected at the QEB office in Knoxville for review. The contracts were selected to be repre-sentative of each section in QEB. Fifteen contracts were selected in' total. After their review in Knoxville, the files for the same contracts were reviewed in SQN NUC PR and were then reviewed in SQN CONST. A comparison of each contract file was made to ensure that the information required to be generated and trans-mitted by the vendor was actually going through TVA's prescribed routing to NUC PR.
In general, contract files were found in varying degrees of completeness with no general organizational style and with portions of files in multiple locations. More complete files were found in the possession of QEB Knoxville, but assin in multiple locations. Specifically within QEli com-pleted files were found in hard copy in the possesion of the unit supervisora, in a two-hour itre-rated room, in the QEB Records Unit files, in the Chicago Regional Of fice files held by the QEB Records Unit and segregated from the Knoxville QEB, files, and in microfilm in the QEB Records Unit. Portions of
{
the same completed contract were found both in hard copy and in g
microfilm, which had been made in 1977, in the QEB Records Unit, k
f At SQN CONST the hard copy files previously tranferred by QEB were found stored in hard copy in a location different than where CONST maintained their vendor QA records. All but about l
six file cabinets of CONST vendor QA records had been micro-f I
filmed and given to NUC PR.
n an NUC PR had CONST film on file and had hard copy files for vendor contracts which NUC PR believed to be duplicate records.
}
NSRS determined, however, that NUC PR's hard copy file, in fact, had information not found in any oliser file.
Most, but not all, of the files in QEB's possession appeared complete. Microfilmed filen at NUC PR and, therefore, CONST were inconiplete. Had the QEll f iles in CONST's possession been incorporated into the CONST files, NUC PR would have what would I
appear to be complete files for most of the contrnets reviewed.
This would exclude, of course, those cont racts wher.e QED only shipped portions of the completed contract file.
None of the files in the varinus locations followed a set style of organization. Some were found in reverse chronological order, some by items produced, some by category of information (cor-p, respondence, tests, etc.), some with a combination of the above styles, and some in no logical order.
Each file contained what was considered to be superfluous infor-mation, s.uch as nultiple copics of contrpets and attachments and
[
multiple topies of correspondence and irrelevant correspondence.
l.
'e Pogs 28 of 44 ATTACHKENT D e
Hierofilmed files contained the same multiplicities and also included machine copies of text pat terns, envelopes, and blank pages. No information was found in any contract IIIe reviewed that the file had ever been reviewed for completeness and no index was found adequate enough to. allow someone looking for information to find it without going through the complete file.
In fact, the retrieval of specific information from one large
' file that had been indexed to a greater extent than most required three hours of coordinated ef fort by the Knoxville QEB indexer, the unit supervisor, and the site inspector via telephone.
The following is a listing of the contracts reviewed. Addi-tional information can be found in the review summary for each con t ra ct in attachment 2.
NSSS/Hechanical Section t
1.
Atmospherie Relief Valves '92697 2.
Auxiliary Control Air Dryers - 83630-1 3.
Auxiliary Control Air Dryers, Dewpoint Ala rra,- 83630-2 4.
Vertical Turbine Pumping Units - 92609 5.
Steam Generator Safety Valves - 92696 l
Civil Structural 6.
Ice Condenser Scals - 82064
- 7. ' Ice Condenser liinge Blocks - 823844 8.
Aluminum and Stainless Steel Honeycombed Cushions - 820345 9.
Missile Doors for Air Conditioning Enclosure - 87262 10.
Reactor Supports - 75018 Instrumentation 11.
Spare Diesel Generator and Exciter, Voltage Regulator -
825204 12.
Level Switches - 83530-1 Electrical
- 13. 6900-Volt Switchgear and Transf ormers - 54495 14 480-Volt Switchgear and Transformer - 54523 Weldments
- 15. Accumulators, Pumps, CVCS - 826301 Also, the Bellefonte containment contract (weldments) was reviewed in the Knoxville QEB office, specifically for retrievability of J
data.
'j 1
o l
L
. 3
- - - - - - - - - - - - - ~ - - - - - - - - - -
~
~.
1r,
- *!. 's
't -
ATTiCIDiENTD Page 28 of 44 Q
I Hierofilmed files contained the same multiplicities and also l
included machine copies of test 3,atterns, envelopes, and blank pages. No information was found in any contract file reviewed that the file had ever been reviewed for completeness and no index was found adequate enough to allow someone looking for information to find it without going thruugh the cornplete file.
In fact, the retrieval of specific information f rom one large
' file that had been indexed to a greater extent than most required three hours of coordinated effort by the Knoxville QEB indexer, the unit supervisor, and the site inspector via telephone.
g The following is a listing of the contracts reviewed. Addi-tional information can be found in the review summary for each
. contract in attachment 2.
NSSS/Hechanical Section l
1.
Atmospheric Relief Valves '92697 2.
Auxiliary Control Air Dryers - 83630-1 3.
Auxiliary Control Air Dryers, Dewpoint Alarm,- 83630-2 4
Vertical Turbine Pumping Units - 92609 5.
Steam Generator Safety Valves - 92696 l
Civil Structural 6.
Ice Condenser Scals - 82064 7.
Ice Condenser llinge Blocks - $23844 8.
Aluminum and Stainless Steel lioneycombed Cushions - 820345 9.
Hissile Doors for Air Conditioning Enclosure - 87262 10.
Reactor Supports - 75018 Instrurnentation 11.
Spare Diesel Generator and Exciter, Voltage Regulator -
825204 12.
Level Switches - 83530-1 Electrical 13.
6900-Volt Switchgear and Transf ormers - 54495 14, 480-Volt Switchgear and Transformer - 54523 Weldments
- 15. Accumulators, Pumps, CVCS - 826301 i
Also, the Bellefonte containment contract (weldments) was reviewed in the Knoxville QEB office, specifically for retrievability of data.
_m
._ - e F
ATTACllMENT D page 29 of 44 O
e IV. ANALYSIS IJuring the investigation of QER rernrda, it became apparent that the investigation must include to some extent activities of groups involved with the records both before and after QEB.
It was believed that the ultimate disposition of vendor-supplied quality records and their usability to TVA needed to be addressed in total before a proper evaluation and corrective action, il necessary, could be specified.
Along these lines the analysis has been broken down into four sub-sections addressing:
(1) the availability of complete records at g
SQN, (2) the usability of records received by NUC PR, (3) the ability of TVA's record system to satisfy NHC reiguirements, and (4) the safety significance of QA records.
Although this report centers almost entirely upon the flow of records to and at SQN, there is no indication that these findings would not also apply to other TVA nuclear facilities.
A.
Availabili ty of Complete !! ecori s j,iLS_gN l
Reviewing NRC regulations contained in 10CFR50, Appendix B, and TVA's Topical Report implementing those regulations, it becomes very clear that all QA documentntion confirming that materials and equipment at SQN met purchane specifications should now be onsite.
The regulations and TVA's implementation specifically state onsite--not the central oliite, not the vendor's plant, but onsite.
From discussions with QEB personnel and in reviewing their records, it was established that SQN did nut have all the records required.
The only shipment of QEB reenrda was made in 1977 and QEB still had several hundred cuhle feet of records for SQN and other nuclear plants.in Knoxville and an undetermined amount in the regional offices. For the most part, QEli did not know how many completed
(
contracts it had for SQN. QE8 pointed out that the sites did not I
want the records in hard copy because they did not have room for them. Both QEB and NUC PR stated that N'IC FR did not know what records they wanted.
Each organization--QEB, CONST, HEI)S--appe, red very protcetive of their own records and ni.:ds.
QEli had their records in a form they wanted and believed their rernrds were the official TVA vendor QA records. With regarel tn form, there were five separate units within QEB and recurds were maintained in five different styles and levels of completenenn. CONST, likewise, believed their records were more complete than QED's and were in a form usable to them. CONST viewed QED records as duplicaten and unnecessary. Therciore, QEll records were either stored in separate file cabinets or returned to QEB and not incorporated into CONST files as their procedure required. NEDS was unwilling to put records in a form NUC PR could use because they were OEDC's recurds management gruup, not POWEH's, They, therefore, arranged to film records as nupplied and index the film on a file level
-2(-
e
f, ATTACHMENT D Page JU o e
e l
l depending on how the file was arranged by the group having it filmed. NUC PR on the receiving end of these records was taking whatever anyone wanted to give them and hoped no one would ask for the records again. Not knowing what NUC PR wanted, both QEH and CONST took the conservative approach and filmed everything
[
i they had in a " contract file, including multiple copies of the contract and attactueents. This was magnified by itEDS filming machine copied test patterns and blank pages because they were in the file. The results of this mass filming was a considerable amount of duplication between QED and CONST.
t CONST had no capability of reviewing film produced by t1EDS and just forwarded the film to NUC PR.
CONST stated that through their use of the film in NUC PR possession, they estimate that 80 percent of the CONST generried QA records are unreadable and the hard copy has been destroyed.
Since QEB did not review the files for completeness prior to shipment in 1977, no one in QEB knows for sure.whether or not the records were complete. Errors were found in the 1977 docu-mentation by NSRS which showed that QED did not review what they thought they were sending SQN CONST either at the time of ship-ment or af ter the reported hectic time of assembling and trans-mitting those records.
It appeared from discussions with QEB, CONST MEDS, and SQN NUC PR that each was unaware of the others required input or the needs of the records recipient.
It appears that the general attitude was that thh QA records are required to be stored onsite; therefore, send the records and don't worry about
{
their usable value to the recipient.
l Considering TVA's assignment of responsibilities, both EN DES and CONST are responsible for providing documentation of the satisf actory design and construction of a nuclear power plant to the user (NUC PR).
Included in this documentation are drawings, maintenance manuals, quality records, etc.
If the sources of quality records associated with a piece of equipment were identified they would fall into three general areas. Those associated with quality in manufacturing, installa-tion, and maintenance. Two thirds of these records are to be supplied by EN DES and CONST. With those QA records and the QA records produced during maintenance, NUC PR is rerjuired and should be able to retrieve records showing the pedigree of a piece of equipment in a reasonable length of time.
i During the investigation QED stated that NUC PR would not apecify what they wanted for QA records.
It can be effectively argued that NUC.PR should not have to specify what is needed.
Dased upon the regulatory rquirements, the onsite required QA docu-mentation is to show that a piece of equipment meets the con-tract specifications. EN DES prepares the specifications, inspects the manufacturing process, and collects the documenta-tion proving to TVA's satisfaction that the specifications have
- -~
- - - ~-
6
~
9 c'
been met.
It follows then, that EN DES should he in the best
=
position to decide what records.are needed to prove the contract specifications have been met, not HUC PR.
A similar argument could be developed for CONST.
The monetary value placed upon these vendor-produced QA documents!
is large. The QEB Branch Chief placed an additional 64 percent
-j on the cost of the manufseturing process to produce a commodity acceptable for a nuclear power plant. After the TVA regional a
inspector assures himself that a commodity does, in fact, meet i
_-g the contract specifications and releases that commodity for p
shipment to a site, the only archival proof TVA will have that shows proper quality will he the QA documentation on that com-
{
modity. Therefore, the Justification that the additional 64 percent was well spent, preventing the possible shutdown as a
)]
unit or the replacement of equipment, rests upon the documenta-tion QEB collects and its usability by the ultimate record holder,
.f Considering the two billion dollars in contracts QEB was inspect-E ing in 1980, the value of the QA documents collected by QEB was f
about 780 stillion dollars, most of which was nuclear related.
Most QEB pe.sonnel did not know whether or not the required documentation was available and retrievable. They all believed flE It was, but most did not have a method of verifying that belief.
J[
In the speeltic case of RT film, QEB did not know nor.did the reviewed contracts show where that film was located. l'or the most part, it was believed to be at the vendor's facility.
l 1
i Returning to why the records were not in the possession of SqN HUC PR, the investigatorn determined that each responsible organi-
-? ;
i ration had an explanation. QEB pointed out that tiEDS could not T i handle the quanity of records QED had.
CONST pointed out they h
had no room for the QED hard copy records and believed them to be f
only duplicates of what they already had.
MEDS pointed out that p ;
filming old records was a low prior!'y and they would not put i
them in a form usable to the site. There certainly was validity In each position, but it was evident also, that each was concerned more with their own problems than with what was required by TVA.
?
Each was aware of the problem of getting records to the sites, but 1 1 y there was no identified effort toward solving the problem and providing the site with OEDC QA records as required and committed by TVA.
4
.I
!.Ili Another aspect of thin whole process involves manpower and i
{ l whether or not it was adequate to perform the assigned tasks.
t Sufficient time was not spent in this area to provide a complete i
y evaluation; however, some observations are worth noting. Both j
the Instrumentation and Electrical Units administer contracts J
l that are relatively small in size. The degree to which these j
(
contract files were organized and their relative appearance of p
completeness indicated that their manpower was adequate. On the y
=g g other hand, contracts administered by the Mechanical /Nucicir and Civil / Architectural Units were in some c'ases very massive in gh nize and complexity. These two units had the added problem of
.ns.
_. ~,_. ~. _.. - - - _ - - - - -
'I------
- ~ ' - ~-'~ -'-
~~
~
ATTACIGiENT D e
e less specific specifications. Hoth factors required additional work over what the Instrumentation and Electrical Units would be expected to experience. Yet the number of personnel in each of the units were basically the same.
Both the Herhanical and Civil Units expressed difficulty keeping up with contracts regarding whether or not the paperwork was completed.
It is not known if this was by personal preference, not related to workload, or out of necessity.
In either case, no information was offered which indicated that either unit felt it was under staffed.
Similarly, no mechanism was described as being used,
L by the branch chief that evaluated the unit supervisors' work
=
with regard to how well the paperwork was handled. Such an evaluation might be appropriate since paperwork is
(
a very tmportant part of QEB's final product. The Weldments Unit did not maintain running files for their contracts; the regional office did. Periodically, about three months after the contract was complete, files were called in from the regional offices.
The one file it had on Bellefonte containment was very large and appeared complete, but was extremely difficult for someone
(
L_
not entirely familiar with it to use.
The inadequate indexing and descriptive intelligence on the file information may or, may not be related to staff size.
Although problems with QEB's records appeared well known by all within QEB, little effort appeared to be expended at identifying and solving the problem of providing the sites complete and usable QA records.
Efforts were started and described during this investigation to put QEB records in order, but none were described to improve the flow of records to NUC PR or their usability to them. Even if QEB's records were in order, and if CONST followed its demonstrated practices, the QEB records would cither be stored in separate filing cabinets or sent back to QEB. With NUC PR receiving only CONST records, the records were incomplete and could not. be used to prove that contract specifi-cations for equipment had been met.
Furthermore, the need for CONST personnel to find information for NUC PR in microfilmed records indicates that the records are no; usable to NUC PR.
Ansuming the regional inspectors were performing their function, and there was no reason to suspect otherwise, at the time equip-ment was shipped by the vendor, QA documentation existed.
Therefore, between the records stored by NUC PR SQN, CONST SQN, and QEB records at SQN, Knoxville and the regional offices and records passed by the vendor or subveniors, the QA records should exist. The data at the vendor's facility becomes i
important in those cases where TVA does not have all the required QA data.
Of the contracts reviewed 4 out of 16 appeared incom-plete, lacking test data and therefore fell into that category (87226, 83630-2, 824204, 83530-1).
Efforts were started by QED during this investigation to develop an index of all contracts requiring QEB. source inspection since 1965 and including, among other things, records disposition...,-__..n__a-___w__--_.n-_-l---------
m ATTACIDiENT D Page r.:
O.
9
,\\
Efforts were made to develop the program for a data base for QEB
{
records. It was described as being capable of retrieving infor-amtion on a document level and would include such things as heat numbers and serial numbers. This system is not operational yet, and only has some test data in it.
Supposedly, it will also have CONST QA records e'ntered into it.
This type of system was needed badly; however, it is not yet the solution to the problem and if it is not handled properly, will add to the problem. The
{
program was developed without input from the QEB unit supervi-lli sors and without NUC PR input. Granted the program was devel-t i
oped under a task force with NUC PR representation, but that g
representative stated he had no input.
It appears that at g
present it will be just a QEB tool. With regard to NUC PH
'q access to the data base, the only NEDS terminal in POWER is in j j the Central Office. The sites do not have a terminal. Further-f more, NUC PR, according to tiEDS, does not want MEDS film; there-fore, this system, unless accepted by NUC PR and in a form they 9
can use, will not satisfy the 10CFR50, Appendix B, Criterion VII i
requirement.
In addition, the use of this system will be for l
current data and old data will be added only as' time permits.
If one considers that most of the equipment for a nuclear power 9
plant has already been purchased and delivered, the data base M
will have limited vclue to NUC PR.
l f
Recent procedure changes for QEB require vendors to supply a list of documents that will provide proof that the commodity i
satisfies the contract specifications'. This should help elimi-nate the problem of negotiating with the vendor and certainly provide a checklist for the materials engineers to check for h
contract documentation completeness.
jl I
The changes in the program and actions described above, while l
necessary and well intended, only address symptoms. They do not i
l address the overall TVA problem of collecting and collating in a
'I O retrievable manner information from several divisions and trans-mitting that information to NUC PR.
In order for the QA records system to work effectively, the overall problem must be addressed and solved. This will require a cooperative effort from TVA top l
management in POWER, OEDC, and OQA.
t i
j ll l
B.
Usability of Records by NUC PR H
i l
I m
NUC PR is the ultimate recipient of all QA records generated during the procurement and construction phases of the nuclear
~
i power plant. As such, they must be able to use the records once 8
l they are turned over to them. To ensure their usability, the 1
records must be retrievable; and to enhance retrievability, the j
records should be edited down to only the essential records and indexed in a manner that is understandable.
ANSI N45.2.9-1974 states that the storage system for QA records
]
will provide for the accurate retrieval'of information without undue delay.
The standard also requires an indexing system i
.g.
r
(-
ATTACIDiENT D Page 34 of 44 e
\\
i l
that Indiraten where the records are to be stured.
ID-QAl' 17.1 requires that QEU provide document-level indexing for engineered i
equipment and tolder-level indexing where nppropriate.
l Retrievability without undue delay is difficult to define.
Some InlC PH employees claim that the information they have been requested to find in the past has been found in*a few hours at I
most.
Others were skeptical that certain typen of data, espe-cially mill certifications, could be located after CONST left I
the site.
g The contract files at NUC PR SQN are indexed to what is called the fulder level. At times the folder is the entire file. Many files were poorly organized. Also, there are many documents in the files that are not required. Examples were film of enve-lopes and machine copied test patterns. Still, for small contracts, the retrieval of information, if in fact the information is present, would not he difficult. But for inrge contrnets, the low level of indexing, the poor organization, and lack of editing would present a real problem in retrievability.
I Still, a fair level of near-term retrievability could be expecte<'
4 long as the people responsible for the nrginal collection of as the records are available for consultation, either in EN DES or CONST. The overall lack of uniformity in the entire area of vendor generated QA records makes the long term usability of these records suspect.
The Bellefonte containment contract which was reviewed in Knox-f ville la n gnod exemple of the complications that might arise.
j This was a very large file--several drawers full. Since it was on a later plant, the file was in better order and better indexed that those at SQN. The piece of data that was requested was locateel by QED but only af ter consulting with the inspector at the regional office who set up the file. As time passes and the i
recordn are transferred, the availability of finding data for a contract such as this will probably be g.eaoly reduced. With the decreasing QEB workload and normal attrition, some of the j
regional inspectorn may be going to other Johs. This problem won encountered in the NSRS review of SQN CONST during thin f'
investigation. Pernonnel responsible for enllecting and storing records were no longer onsite.
e The overall usabillty of the QA recoris fninni at SQN is highly suspect for the reasons stated above.
s C.
A,bility l the TVA Sy_ stem to Hect the NRC and ANSI Heyuirements i
l Since 'RG 1.88 endorses ANSI N45.2.9-1974, this section will nnly l
address the ANSI standard. TVA implements the standard through the Topical Report, the ID-QAPs, and lower level procedures.
Generally, each level of control adequately proceduralizes the requirements of the higher tier controlling document for the I a
- - - -. - - ~. - -
,l
- R..'.*
- ATTACHMENT D Page 35 of 44 0
0 I
I speelfic area for which the controlled organization has juils-
' lict ion. TVA's procedures, with some interpretation and it imple-l, i
mented, should meet the requirement of the ANSI utnndard.
I In actual practice they are not completely implemented and a prob-tem la encountered when different divisions interface. Not all required QA records were given to NilC pH by EN DES.
Even though every organization believed they were doing their Job, the intent i
of getting all QA records in the contaol of the ultimate user was k
not being met.
There was no overall check of the enntrolling g
function. Managers may have reviewed the work perform.d by per-g nonnel under them, but no one manager reviewed the overall TVA QA accords control.
A D.
S dejy Significance I
i A logical concern after reviewing the previous portions of this report would be the safety significance of a poor set of QA f
records.
In reviewing the QEll records, the paper trails of
~
quality for specific pieces of equipment, and'in discussions b L h
with personnel involved in the paper Liail, it is difficult to state unequivocally that there is or is not a safety problem.
g Tn place some perspective on this question, it is best to briefly j
review the process by the vendor and the line organization for annuring a quality product.
The first level of quality is supplied by the vendor in its QA program. That program is reviewed by QED to assure that the vendor can produce a commodity with the quality desired by TVA.
Therefore, before a piece of equipment is offered to TVA, it has gone through the vendor'. QA program and the vendor believes it p.
meets TVA's specifications.
I Next, the TVA regional inspretor physically inspects the com-mndity 4..nl the documentation or data attesting to i t e. rpra l i ty.
1 1herclore, at the time of shipment, both the vendor nnd the TVA j
QEll regio..al inspcetor are sat infied tha6 the equipment meets j
the contract specifjrations.
A third review is perfor meil ley OlN'iT fil0 when t he commnili t y arrives at the site.
1here the til(;.isnures itself that all the equipment is accounted for, that it meets the contrart specifi-p cations, that QED has authorized its shipment, atul tila t any vendor data to be sent with the cepeip.. tent is acronnted for.
l The equipment, therefore, has three levels of inHpertion by the time ll Is accepted at a plantsite i
The QEll unit supervisors in Knoxville are required to further inkpcet the records to assure they are accurate and complete.
l Exnmining the inspection prorcss it becomes apparent that the m
l QLil regional inspector is the key to assuring that the vendor meets the contract specifications. With the exception ul the
.n.
i 1
' gCys,% N9 % W P"H*(it
.T.
"',' : $' ^'S^E*TP d, T.
c'. : t. t *
- J. * *.*j ! ^.' ' ! **" f.'lji. i,. p&, -
P
'4
9 g age 36 of 44
. ATTACIMENT D
.a ;, *....q..
... w.
W n.-
vendor, and in some cases the HIG, the regional inspector is the only one in the QEB organization that can physically inspect the commodity against the vendor QA documentation and the TVA specifications. All other QEB cvaluations for a convuodity will involve a backup check of what was previously done by the vendor I
or the regional inspector by making a comparison between QA docu-mentation and contract specifications. Such a backup check could detect such deficiencies as using the wrong materials in construc-tion, a deficiency which should have been detected in previous inspections. The same would also hold true for a technical g
,r engineer's review of QA data.
In the case of nondestructive tests performed by the vendor, the regional inspector is the only one in TVA that can physically observe those tests and the only one in TVA who, in most cases, ever reviews RT film. All other TVA reviews will be of the paperwork certifying that a test..
i c
was performed and that the vendor and regional inspector recorded the pass / fail test results. Therefore, all TVA reviewing, in the
- L,' :
QEB chain of paperwork, beyond the regional inspectors will only involve a recommendation of documentation previous,1y examined by 5'
the regional inspector.
=
for example, the 480-volt shutdown boards for SQN went through the normal review process--vendor QA, regional inspector, and HIG. The paperwork went through a secondary and tertiary review i
I by the QEB Electrical Unit and technical engineer. The shutdown boards were accepted, but during installation, TVA determined that the welds in the shutdown board,had every defect possible.
No amount of paperwork review could find that kind of defect.
7 The probability appears very steell of finding a defect in a 4
comenodity through a paperwork review. At best the review might have established that the initial reviews were not made or that certain portions of inspections were incomplete.
Quality assurance records can be of significant safety importance in a number of situat. ions. When acueric safety problems are identified with a given vendor, either due to material problems, process problems, or function performance problems, it is extremely foportant for the utility to determine if the identified problem exists at the operating plant. This requires both the existence of the records and the retrievability of the records.
In the I
absence of adequate documents, it could be necessary to either shut down and replace or require some degree of yerification.
Records are also important in evaluating causes of failure of a component or system. Exarsination of QA records may provide key information to determine the cause of failure or to identify corrective action.
In their present form, QEB generated QA records may not be sufficient for the above purposes.
V.
CONCLUSIONS
)
A.
Vendor-supplied QA records obtained by QEB for SQN have not been i
completely turned over to NUC PR and t,here does not appear to be any plans to do so in the near future. This places TVA in 1
a.s
.f=
7,,,
es h
violation of 10CTRSO, Appendix B, Criterion Vil, and Topical
' Report requirereents. Although this report examined the flow of records to and at SQN, there is no basis to believe that this
' conclusion would not also apply to other TVA nuclear facilities.
B.
There is no overall OEDC cffort to provide NUC PR a complete QA file integrating both CONST and QED records in a usable form.
This is in violation with TVA QAP PRM 17QPR-1.
C.
The level of effort being expended to assure that more than g
700 raillion dollars worth of QA records are properly assembled and accounted for is less than adequate.
l D.
The retrieval of information from vendor QA records at SQN by NUC PR is considered to be inadequate due to the poor organiza-tion, lack of completeness, low level of indexing, and lack of editing. Retrieval depends upon use of individuals located
/y away from the site that are not readily controlled by HUC PR.
D.
There is no coordinated TVA effort to assure the adequacy of identification, collection, collation, indexing, and distr,1bu-I l tion of vendor QA records.
h F.
TVA would meet the requirements of HRC regulations and ANSI standards if the current requirements and intent of documented h
g procedures were followed; however, their implementa tion, pa rtic-P I
j ularly at the interface point between, various TVA organizations, h[
is less than adequate. This results in a records system that y
i meets neither the intent nor letter of the regulations. Here implementation of the procedures in their present form without M i ji j l
regard to intent and an overall cont rul of their implementation b
j by an organization capable of easily crossing organizational
,N l
boundaries would not solve the problem, d 'l VI.
JUDCHENT OF NEEDS W
A.
tical and consistent implementation of TVA policy regarding the g ',
A coordinated OEDC/POVER/0QA effort is required to assure prac-p [
identification, collection, collation, indexing, and transfer of ll vendor QA records in a form usable to NUC PR.
This effort should
[
h]l
}h begin at the office manager level with a commitment of how, when, and in what form the records transfer should occur; and the ef fort should be continued by individuals knowledgeable of vendor QA d,5 j 'j records and ADP processes. 1his cont.nuing elfort should be coordinated by an organization with the authority and capability H
to easily cross organizational boundaries.
l i l
Il B.
A coordinated effort is needed by OEDC and NUC PH to assemble l
and transmit in a usable form contract files for all currently d
contpleted contracts and purchase orders on a plant-specific
,j basis.
This should include all QEB and CONST records previously h
submitted to NUC PR.
l,
- N
,_ I
.-----e---~
' ' ~
t
- a, e
1,.
- t. s i
TTAtllHENT I
^
WELOM M S J. H. Winebrenner m
J O
~
MECHANICAL / NUCLEAR g
2 J. F. I.ewi s 8
x n
E b;[ y INSTRUMENTATION 4
0 H. D. Conner 3*
g>
d CIVIL / ARCHITECTURAL b
D. E. Smelser
%w U
ELECTRICAL I
}{. W. Knox 44 we I
J u M
S8 O
e Q
a: A C
5 TULSA REGIONAL g
INSPECTION OFFICE e
m W
ee t
a:
C. S. Bates E
c*
u a w
E6 m
C d0 ST LOUIS REGIONAL INSPECTION OFFICE 3U C. J. Tunnicliff SS M
W PROVl0ENCE REGIONAL
>.E INSPECTION OFFICE A
lf
-- 4 s
$ =E F. F. Ross 0
W2
=
h gu PITTSBURGH REGIONAL o
P v, lNSPECTloft OFFICE 4
[f D. Pelecovich M u w
s ge 3
PHILADELPHIA REGIONAL Q
33 INSPECTION OFFICE h!
V. C. Kolinger N
LOS ANGELES REGIONAL 8
INSPECTION OFFICE ky II. J. Schumacher OI w e 0; "
CHARLOTTE REGlot1AL w.3 INSPECTION OFFICE n'.
H.
A'.
Brittingham M
BIRMINGHAM REGIONAL INSPECTION OFFICE va.
._ : e _.s. a ___ = a,
.,.. n.
i N (,' 'i M*$ hkfl
[
e
. g.. g y. p v -
.'.N [
ATTACHMENT D Page 39 of 44 i
i ATTACllHENT 2
SUMMARY
OF CONTRACT Fl!.ES REVIEVED 1.
Atmospheric Relief Valves - 926977 This contract file was on microfilm in the Knoxville QEB office.
I The t file appeared to be completed, but the radiographic test (RT) reader p
sheets were illegible. A hard copy of this file had been sent to SQN r
CONST in two parts on August 16, 1977 and October 18, 1977. The CONST file for this contract had been microfilmed by HEDS.
g The microfilm had been sent back to SQN.CONST and transferred to NUC FR.
The FUC PR microfilm of this contract file included a copy of the contract, various correspondence, QEB arid MIG inspection and testing reports; but no test data or information on the RT film was available.
In a hard copy file at HUC PR copies of Cope and Vulcan procedures I
and RT procedures with the shot locations were found, but still no RT reader sheets were available.
q 2.
Auxiliary Control Air Dryers - 83630-1 This contract file was on microfilm in the Knoxville QEB office.
The file appeared to be complete. All informqtion in the ille was in chronological order with the latest document first. A hard copy of
- i this file was sent to SQN CONST on August 16, 1977. After the HSRS review at SQH, CONST personnel found the hard copy sent by QEB.
j file for this contract had been microfilmed and the film transferred f }k.
It in storage in filing cabinets in another building. The CONST was to HUC PR in the same manner as the previous file.
a
] p The HUC PR microfilm did not contain manufacturer's specifications, data packages, or certificates of compliance. This microfilmed copy Q
was in a more orderly format than the QED microfism. That is, a copy g i, of the contract was first and information was gathered into groups of conwnon types.
In the SQN NUC PH hard copy file were data packages
],
g but not certificates of compliance. The vendor-related QA data was incomplete at SQN NUC PR.
[
3.
Auxiliary Control Air Dryers, Dewpoint Alarm - 83630-2 i
This contract file was on microfilm in the Knoxville QED office.
hl i The l
i f
ffle appeared to be complete. A hard copy of this file was sent to ld SQN CONST on August 16, 1977. The CONST file for this contract had E
been microfilmed and the film transferred to NUC PR in the same manner as the previous files.
The NUC PR microfilm contained a copy of the contract and the QEB inspection reports but did not contain the certificates of compliance l
or test data packages.
P
.. w s
<r
..e
, e.o 3
g ATTACIDiEMT D g Page 40 of 44
.t 4
Vertical Turbine Pumping Units - 92609 This contract file was on microfilm in the Knoxville QEB office. The information was in two different lorations on the microfilm and divided by another contract. The image numbers given for the con-tract were incorrect. The reel numbers had been changed in ink, and the wrong reel was given to the reviewers initially. The file was not formated in an orderly manner. There was a large amount of test data, certificates of compliance, and miscellaneous information available, but the method in which the contract was set up made it impossible for the reviewers to determine its completenesa in a reasonable length of time. A hard copy of the file was sent to SQN in two parts on August 16, 1977 and October 18, 1977. The CONST file for this contract had been microfilmed and the film transferred to NUC PR in the same manner as the previous files.
The NUC PR microfilm contained no test data packages.
It did contain the usual information such as HIG receiving reports and several copies of the contract. The file was in a less orderly format than other files reviewed at NUC PR.
7 5.
Steam Generator Safety Valves - 92696 Part of this contract file was on microfilm at QEB in Knoxville and part was in hard copy. The microfilm section included a copy of the contract and changes to it and a few types of required data and certifications. The hard copy, which was is,i the Records Unit files being prepared for microfilming, contained data packages for all 40 valves. The combined contract file appeared complete. The hard copy of what QEB had microfilmed had been sent to SQN CONST in two parts l
on August 16, 1977 and October 18, 1977. The CONST files had been microfilmed and the film transferred to NUC PR in the same manner as the previous files.
The NUC PR microfilm contained a copy of the contract, QEB inspection and testing reports, some correspondence, but no data packages. The SQN NUC PR hard copy file contained manufacturer's procedures, a copy of the contract, and test results for 6 valves (34 were missing). No reference was found at NUC PR in the contract files concerning the location of the RT film or the reader sheets.
6.
Ice Condenser Seals - 82064 This contract file was in hard copy form in the Knoxville QEB office.
It was well organized with an index in the front of the' file. There was also a matrix form checklist with data requirements versus data received fur each component in the contract. This file had not been sent to the QED Records Unit for microfilming preparation. Also, there was no record of it having been 'sent to SQN CONST. The CONST file for this contract had been microfilmed and the film transferred to NUC PR in the same manner as the previous files.
The NUC PR microfilm was poorly organized, but it did appenr complete.
It had more data than the QED file.
N u
2 h
~~
.c g
~ '
h Page 41 of 44 ATTAClaiENT D There were test data sheets and certificates of compliance available which QEB did not have.
7.
Ice Condenser Ilinge Blocks - 823844 This contract file was in hard copy form in the Knoxville QEB office.
It was a small contract, and it appeared to be complete.
This file had not been sent to QEB Records Unit for microfilening preparation.
There was no record of it having been sent to SQN CONST. The CONST file for this contract had been microfilmed and the film transferred to NUC PR in the same manner as the previous files.
I The NUC PH microfilm nppeared to be complete.
8.
Aluminum and Stainless Steel, lioneycombed Cushions - 82034 Part of this contract file was on microfilm at' QED in Knoxville and part was in hard copy. The microfilm portion contained, among other things, a copy of the contract, specifications for the items, certi-k ficates of compliance, and shipping memorandums. The.hard copy g
portion of the contract file contained test data reports for all cushions. There was no transmittal of the contract file to CONST or to 11EDS. The CONST file for this contract had been microfilmed and the film transferred to NUC PR in the same manner as the previous files.
The HUC PR microfilm contained a copy of the contract and changes to it, receiving reports, and certificates of dompliance for eight cushions. There were 24 aluminum and 96 stainless steel cushions bought on this contract, so there should have been 120 certificates of compliance.
9.
Missile Doors for Air Condition,ing Enclosures - 87226 This contract file was on microfilm in the Knoxville QEB office. The file contained a copy of the contract and the usual correspondence.
There was no test data and no mill certifications. There was no record of shipping the contract file to SQN CONST. The CONST file for this contract had been microfilmed and the film transferred to NUC PH in the same manner as the previous files.
The HUC PH microfilm contained a copy of the contract cotrespondente and inspection and testing reports, hnt no vendor test data or mill c e,r t i f i ca t i on s.
10.
Reactor Supports - 75018 Part of this contract file was on microfilm at QED in Knoxville and part was in hard copy. The microfilm portion of the contract ron-tained a copy of the contract, correspondence, and seme manufacturer's Information. The hard copy file, which was in the QLB Records Unit files, contained data for all components in separate data packages.
The combined contract file appeared complete.* The NSRS was told that the contract file had been sent to SQN CONST in the 1979-80 time 3
w
mm
.," Q.
e,.;;. p g
ATTACllMENT D hPage42of44 period and returned at a later date, but there was no record of this t ransisi t t a l.
The CONST file for thir. contract had been reicrofilmed and transferred to NUC PR in the same manner as the previous files.
The NUC PR microfilm contained a copy of the contract, dence, and shipping and receiving records.
correspon-There were no test data packages, no mill test reader' sheets.
reports, and no information on the RT film or 11.
Spare Diesel Generator and Exciter. Voltaar__Resulator -
825204 g
This contract file was in hard copy form at QEB Knoxville.
The file contained a copy of the contract and correspondence.
There was no test data.
CONST or to HEDS for microfilming.There was also no record of the file The CONST file for the contract had been microfilmed and the film transferred to HUC PR in the same manner as the previous files.
The NUC PR microfilm contained a copy of the contract and changes to it.
There was more correspondence at SQN than at QED.
There were also receiving reports but no test data or certificates of comp 11-
?
Some installation instructions were included.
ance.
1 In the SQN hard j
copy controlled manuals was a copy of the selsmic analysis and the procedures for performing it.
This was also present in the SQN hard L
i l
copy contract file (considered to be duplicates). The controlled J
copy came from the NUC PR Central Office in Chattanooga; the origin 1
of the other copy was not known.
W 12.
l.evel Switches
_83530-1
}j[
This contract file was in hard copy in the Knoxville QEB office.
lq Pressure tests were available. A document in the file verified that test data was on file at the vendor's plant. Certificates of compli-ance were accepted.
in Knoxville, but the reviewers could not locate it.A materials test file was No test data was available in the contract file. Since the Chicago Regional Office has been closed, their contract files had been sent from that office to Knoxville.
The reviewers looked at the Chicago file for this contract, which was located in a different area than other QEB files, and large numbers of test data were found, much more than could be located in the normal Knoxville file.
The regional office file appeared to be more complete than the Knoxville office file.
although it was less organized.
There was no record of the QED con-tract file being shipped to SQN CONST or to MEDS for microfilming.
The CONST file for this contract had been microfilmed and the fills transferred to NUC PR in the same manner as the previous files.
i h
3 The HUC PR microfilm contained several copies of the contrnet, cor-il respondence, certificates of inspection, information sheets, test data for hydrostatic tests, and certificates of compliance.
J The test dets sheets were not coroplete. The NUC PR Igard copy (considered duplicates) contained seismic certification, test data sheets, and engineering reports.
The switches had originally been al! listed as 4
jw 8
x.,
s,.
...g..
.. r.. 't
.. m.
h h
Page 43 of 44 ATTACIDiENT D s
requiring QA; but since many of the switches were in systems which did not require QA, memorandums had been written deleting some require-ments for non-QA switches.
13, 6900-Volt Switchgear and Transformers - 54495 This contract file was in hard copy in QED Knoxville. The file was in the QEB Records Unit files and there was no record of it being transmitted to HEDS for microfilming or SQN CONST. The file was well organized and appeared complete. A copy of the contract was included along with data packages, test reports, and certificates of compliance.
B Seismic certification was available, but that test data was not in the file. The CONST file for thin contract had been microfilmed and the film transferred to NUC PR in the name manner as the previous Iiles.
The NUC FR microfilm contained much duplication. Copies of the contract were included, both blank and complete.- Seismic certifi-cation and shortage reports were included, but there were no test reports. The NUC PR hard copy file contained the seismic test report and analysis but did not contain the test reports found at QEB.
- 14. 480-Volt Switchgear and Transformers - 54523 This contract file was in hard copy in the Knoxville QEB Records Unit files. This file was neat and well organized. A note was in the file stating that tent data was with the technical engineer. Otherwise, the
!!1e appeared complete.
Information was contained in the contract file that, af ter acceptance onsite, about every type of weld defect possible was found. These were corrected by the vendor. The completed contract had not been filmed or sent to SQN CONST. The CONST file for this con-tract had been microfilmed and the film sent to NUC PR in the same manner as the previous files.
The NUC FR microfilm contained a copy of the contaaet and changes to it, receiving reports, certificates of compliance, and correspondence.
There were no test data packages. The NUC PR hard copy file contained certi!! cates of compliance for scismic anlayses, transformer tests, equipment manuals, and apparently complete test data.
- 15. Accumulators Pumps. CVCS_- 826:101 This contract file was in hard copy at the Knoxville QEB offich. No test data was available in the Knoxville file; it was all still in i
the los Angeles Regions! Office. The Los Angeles office had.its files merhed as completed on Hay 7, 1981. The CONST file for this i
contract had been microfilmed and the film transferred to NUC FR in the same mannnr as the previous filen.
The NUC PR microfilm ille had a copy of the contract and three copies of appendices A, B, and C.
It also had four document packages. A note was present stating that the RT file had been returned to SQN NUC PR, but the location of that film could not be determined.
+
_.__.__m a m
.-u
' * ' "s; l' *
.;)
(.I e
vs
[e 4,'
ATTACIDiENT D Page 44 of 44
+..
16.
Dellefonte Containment - 85617-1 This contract file was reviewed to determine the means of retrieving data from a completed file.
It was in hard copy form and had only recently been reviewed and sent to the QE8 Records Unit. The reviewers asked to see infornation on a piece of a polar crane bracket, 3-56 H1-1, PA 3-55-5.
The information was found in about three hours and with a telephone conversation between the QEB Knoxville materials engineer (section supervisor) and the inspector in the field of fice to determine the vendor's method of cross-referencing. The person indexing the file also helped in locating the requested document.
The Veldsents Unit Supervisor indicated a need to add additional l
intelligence to the file for case of information retrieval.
j l
{
r l
I I
e_
1 i
s e
l 6
ATTACHMENT E HSRS Investigation Report R-85-07-NPS f
+-
G 9
e t
p
/
i r
l l
t t
I l
t 1
l l
l
[
m,.