ML19309B124
| ML19309B124 | |
| Person / Time | |
|---|---|
| Issue date: | 02/02/1977 |
| From: | Volgenau E NRC OFFICE OF INSPECTION & ENFORCEMENT (IE) |
| To: | |
| Shared Package | |
| ML19309B125 | List: |
| References | |
| SECY-77-064, SECY-77-64, NUDOCS 8004030166 | |
| Download: ML19309B124 (86) | |
Text
7 TL e 1 DFFHCHAL USE ONLI
' February 2, 1977 SECY-77-64 N g'
-w UNITED STATES NUCLEAR REGULATORY COMMISSION
- f'
- qi; mu CONSENT CALENDAR ITEM i
For:
The Coninissioners
~
From:
Ernst Volgenau, Director Office of Inspection and Enforcement Thru:
Executive Director for Operations
Subject:
C' LICENSEE CONTPACTOR AND VENDOR INSPECTION PROGRAM (LCVIP)'
Purcose:
To obtain Commission approval of the LCVIP as a continuing pt.rt of the NRC inspection program conducted by the Office of Inspection and Enforcement (IE) and _
approval of a two-year trial program to be sponsored by IE and conducted in conjunction with the American i
Society of Mechanical Engineers (ASME), to test the feasibility of recognizing third party inspection j
systems as a means of supplementing the curr nt LCVIP.
l Category:
This paper covers major policy questions.
~
Issues:
1.
Should NRC formally recognize the need for a continuing vendor
- inspection program to be conducted j
by IE?
l t
- 2. _Gi_v_en that NRC should have a veILdor inspaclio.n '
f program, how should the current LCVIP be expanded to l
_ provide adeguate inspection coverage of all vendors?
..a
Contact:
G. W. Reinmuth, IE X27461 l
M term to describe organizations supplying products and services to NRC U
- For purposes of' this paper, the word " vendor" is used.as an all inclusive M
reactor licensees, including nuclear steam sis'tein s'uppliers and architect engineering organizations.
owancnan nur nun v 8004030 16 6
.f i-2 Decision Criteria:
Issue 1 Does the LCVIP detect and assess problems which may affect public health and safety and effect corrective actions?
Are the results of LCVIP of sufficient merit to justify its continuation as a formal effort?
Issue 2 Is it likely that the implementation of adequate, effective vendor QA programs can be verified through third party inspection ' systems?
Will the benefits of utilizing third party inspection systems outweigh the costs?
Discussion:
Issue 1 Events in mid to Tate 1960's, particularly those asso-ciated with quality problems identified at the Oyster Creek facility during the late stages of construction, led to a recognition by NRC (AEC) that to adequately protect public health and safety, effective quality assurance (QA) programs must be applied to all functional activities associated with a nuclear power plant, including design, procurement, fabrication, construction, testing, and operation. The criteria for QA programs to be i
established and imolemented by licensees were subsequently I
developed and issued as Appendix B to 10 CFR Part 50.
Under these criteria, a licensee must impose the necessary quality assurance requirements on his vendors and then f
must evaluate the effectiveness of his vendors' quality j
assurance program.
The procedure of relying on licensees to assure that vendors were imp 1'ementing quality assurance programs j
was employed by NRC (AEC) during the period prior to 1973.
i Under this procedure inspections of licensees by IE
" 'iC !..
I I
=.
. included ochasions on which IE inspectors accompaaied licensee representatives on their vendor inspections.
Experience with this approach demonstrated that relying on licensees to assure vendor performance was not adequately effective in identifying and resolving.
quality problems.
This was true because a ~ licensee's.
authority over a vendor is limited to the scope of his individual contract and because a licensee tends to review vendor quality assurance programs from a limited perspective, f.e., whether the program results in the product (s) the licensee wants. These limitations inhibited the identification and correction of generic-type problems and.the feedback of information to other affected vendors and licensees and to the NRC.
- Further, placing responsibility on each individual licensee for evaluating performance of all vendors with whom he has
. contracts results in duplication of QA program evaluations since many licensees are required to audit the same vendors.
A trial inspection program covering only fuel fabricators -
was conducted by IE during 1973 to evaluate a direct approach to vendor inspectionJ This demonstrated that greater conformance to quality standards and a subsequent I
reduction in major quality problems could be achieved through an effective direct inspection program.
Appendix A of the enclosed report discusses the results of the trial program for inspection of fuel fabricators.
An analysis of c1e causes of Licensee Event Reports received during thi.s time period. showed a continuing need for improved vendor performance as evidenced by a high incidence of vendor-related problems reported by licensees. This analysis and the success of the direct inspection approach in inspection of fuel fabricators led to the initiation of the Licensee Contractor and Vendor Inspecticn Program (LCVIP) in 1974 as a trial program covering all types of vendo rs.
. g....
c r
_4_
The LCVIP has been conducted to determine if impr5ved vendor confornance to quality standards could be achieved, to identify methods and procedures by 1
which to conduct routine vendor inspections, and to identify NRC resource implications. An inspection unit was established in the Region IV office to conduct the LCVIP on a nationwide basis.
During the two-year period this unit has performed approximately 300 inspections of nuclear steam system suppliers, architect.
engineering firms, nuclear fuel suppliers, and several types of product vendors.
Tnis trial program reaffirmed the need to inspect quality assurance p. for impiecentation of effective,. _
rograms_at the v_endor level and demonstrated that a limited NRC inspection program can materially improve vendor performance.
Typical examples of inspection findings are summarized in Appendix C of 3
the enclosed report. Since there is a> lag of several years between program changes which could affect product i
quality and the use of a product affected, it is not yet possible to identify improvements in product performance.
j i
Based on impnsvements in quald < achieved as a result of upgrading licensee QA progra..~, by analogy, there is a reasonable expectation that improvements in vendor QA programs will similarly improve vendor products and i
therefore better assure public health and safety.
The organizations presently being inspected under LCVIP include five nuclear steam system suppliers, fifteen architect-engineering firms and approximately 120 com-ponent manufacturers.
The component manufacturers include the major suppliers of most ASME Class 1 type equipment including pressure vessels, steam generators, pumps, valves and major piping including support hardware. Manufacturers j
of non-ASME code components that are inspected include sucoliers of nuclear fuel, control rods, control rod drives, valve operators, and electrical and instrumentation com-ponents.
The frequency of inspection prescribed varies from one to four times a year and is determined by the safety significance of the product, the volume produced by a specific supplier and the performance history of the supplier and his prodtic.'. '.
i
]
4
~
q
, The basis used for inspecting the nuclear steam system suppliers,.srchitect-engineers and nuclear fuel suppliers is the QA Topical Report submitted by these organizations to NRC. These reports, which describe the vendor's quality assurance program, are submitted to NRR for review, an.d when approved, are used by the IE inspectors as the basis.for veHfying proper QA program implementation.
Where there is ail NRR-approved QA Topical Report and IE has verified QA program implementation, lice.nsees have been given the option of accepting IE inspections.as a basis for determining acceptability of a vendor's QA program. This procedure provides a basis for reducing some of the duplication of vendor inspection. While a licensee may accept IE inspections of a vendor's QA pro-gram, the function of product acceptance, whether the product. is. hardware or a design package, remains the sole responsibility of the licensee.
.Tne. basis. for inspecting component manufacturers, where.
~there is no NRR-approved QA Topical Report has varied.
In the case of an ASME code camponent, the basis has been the ASME-approved QA Manual.
For other components, inspections are made directly against the criteria of 10 CFR 50, Appendix 3.
Utilizing the ASME code as basis for inspection is not always satisfactory because the scope of the ASME code is limited to components in the pressure boundary and because questions raised regarding ccmaatibility of code QA requirements with Appendix B criteria have not been resolved.
It is now planned that all LCVIP inspections will be based on Appendix 3 criteria when there is no.NRR-approved QA Topical Report.
In these' cases, no basis is presently provided for reducing duplication of inspection by licensees.
Procedures for inspection to determine compliance with 10 CFR Part 21 wfll be incorporated into the LCVIP when proposed Part 21 becomes an effective rule.
The pros and c:ns of the alternatives under Issue 1 are identified below.
Issue 1
- g.,
Should NRC formally recognize the need for a continuing vendor inspect, ion program to be-c:nducted by IE?
9 m
L
1
~
[.,
i 4
Alternativ'e.1 Rely on licensees to verify vendor quality assurance
'j programs and conduct no direct NRC inspection program for vendors.
Pro: 1) Minimal NRC resource commitment (inspect only licensee QA programs).
- 2) Clearly assigns a licensee the responsibility for verifying performance of his vendors.
- 3) Avoid; any legal questions regarding direct enforcement authority over vendors.
j Con:
1)
Previous surveil. lance of vendor performance by Ticensees was not effective.
- 2) _ Minimal,.NRC_ ins _pection_ impact on TincoF
. 9A..p_ro. grams.
- 3) Safety-signifi. cant probleas and their generic impact may not get proper attenticr.
4)
Duplication of licensee effort, many licensees required to inspect the same vendors.
Alternative 2 Continue the'LCVIP as an approved part of the NRC inspection j
program.
)
Pro: 1)
Extends NRC inspection impact to total design, procurement, and manufacturing process.
2)
Problems can be assessed for safety and generic impact and information transferred to other vendors and licensees.
g 4
m
-re,,
=
l,
, 3)' - Positive effect on public confidence. __
.)
Potential for reducing duplication of 4
i f
vendor inspections.
Con:
1)
Requires comitment of NRC resources.
~Aut}irltE.for_df rect ewe'r'cAndOI_~_E _._
~
~
~
2)
. requirements on vendors. may.be questioned since vendcrs not Ticensed.
3)
May be perceived as relieving a licensee of his responsi'aility.
Both alternatives under Issue 1 provide a positive effect on pubTic health and safety; however, Alternativa 2 has a much stronger influence because of the generic capability of a direct inspection program.
Experience has shown that i
the direct approach detects vendor quality problems which go unnoticed by licensees and that this approach can be i
effective in bringing about corrective action. Alternative 1 would continue to place an unreasonable duplicative inspection burden upon industry; whereas under Alternative 2, the LCVIP as presently implemented provides a basis for partial relief from this bui-den and offers the potential for providing additional relief.
This potential can be further increased through u;ilization of third party inspecticn systems as discussed below.
On balancing alterna-tives, Alternative 2 dominates Alternative 1 and is the recomended course of action.
Approval of Alternative 2 will mean continuation of the present program for inspection of organizations which have NRR-approved QA Topical Reports (nuclear steam system suppliers, aichit5ct-enginears, ~and nuclear fUef 'sli;iisilers)~.
~
"to verify implemen~tation of the Q program 'desiiriSed in their !.
~
~
Topical Reports, and inspection of other vendor organizations to verify implementation of a QA program wnich meets the necessary criteria of Appendix 8.
Selection of vendor organizations for inspection will be based en the importance of the safety-related.prodO5::;.the repcr.ed prcblems with the '
vendor and his products, and the volume of his work.
In all cases, detenntning compliance with 10 CFR Far: 21 will be a part of the inspection program.
4
7
_a
~
i Issue 2 Tne NRC resources currently applied to the LCVIP provide inspection coverage for enly a limited nuccer of licensee contractors and vendors.
Last year under LCVIP there were
_foscections_o.f_.f.t ve. nuclear.. steam system sucoliers,..ei_qht_
, architect-engineers, five fuel fabricators, 73 ASME product vendors and. eighteen non-ASME creduce vendors.
This
~ ipectioil c:iverage of only th'e da3ar'vendari out iif 7 he in t
approximately 1000 industry vendors provides positive results, but is insufficient coverage if the vendor problems -
identified under the LCVIP are representative of those which exist at venders not inspected under LCVIP.
- Further, because of its limited coverage the current LCVIP does not provide a basis for significantly reducing duplication of vendor inspections required df the industry.
An alternative method for increasing the scope of vendor inspection is through the expansion of directly e=pioyed staff.
This alternative offers tighter control over inspection attivities, hence a more _
positive and faster response in achieving correct 1on
~,_
of deficiencies. Tne principal disadvantages are the high cost as compared to use of third parties and the Tower potential for reducing inspection 1
redundancy.
Tne arguments for and against this alternative and the estimated size and cost of the required inspection staff are discussed in Appendix E.
An appro'ch being considered by IE to expand LCVIP inspection a
coverage is to supplement a modest in-house inspection program with NRC-endorsed third party
- inspection systems.
This approach would provide an increase in NRC-endorsed inspection coverage of vendors without an increase in the current NRC resource cc=mitment by taking advantage of third party inspectors such as the approximately 300 ASME coda inspectors engaged in inspeccion of comconents fer nuclear power plants. A description of the ASME code
-l survey and inspection system is contained in Section IV-B of the enclosed report. As an example of the gain which might be achieved through a recognized third party inspection system, NRC-endorsed inspection coverage of the ASME com-ponent vendors could be increased from the 73 firms inspected l under LCVIP last year to about 3SO firms inspected by ASME.
This would be an approximate five-fold increase in the number of firms inspected while NRC manpower requirements for LCVIP would remain esserttial.ly the same.
e
- "Tnird party" means any drganizaticn other than NRC which provides inspection services and is. not under the direct controi of a.11censee; Ticensee-agent,, supplier,. or manufacturer.
N_ ' While at this time the third party approach has been considered only in connection witn the ASME inspection '
. system, the possibility of recognitien of other third p(e.ty inspection systems in other functional areas ar gde g., fire protection) is not ruled out.
In addition I'#83
/
to providing expanded inspection coverage, utilization of N'P0 third party inspection systems under NRC endorsement has l
sht. i the potential for providing a basis for significant re-3duc_t duction in the duplication of vendor inspections currently 7
required of licensees.
If fully effective in reducing duplicatien of inspections, flems the third party approach could reduce the overall :cos t l
of regulation since the cost benefits to the industry FP, will exceed NRC's cost of conducting the program.
Through the joint efforts of the LCVIP and a third party h;
system, the net reduction in the cost of regulation
' coufif Eslff ih'eFecouiif nc of NIC costs for LCVIP i
~~
/
mrougnTncreasec ticerise rees.
AnaiYsTs oTttie cost -
[
- cenefits. and cost recovery ascects of endorsino the
/
" ASME inscection svLtem_are. contained _iA_S_ecj;isn V-B i
. and Section VII of the_e_ne_losed. report.
While endorsement of the ASME inspection system as a
/
recognized third party appears to be an attractive approach, changes in the ASME code requirements and the inspection f
system and furtherrevaluation of the ASME system by NRC are considered necessary before recognition of the ASME system can be reccanended.
With respect to changes in code requirements, it is necessary that there be full compatability of ASME QA requirements with the NRC criteria in 10 CFR 50, Appendix B and the guidance. in regulatory. guides endorsing the N45.2 series of ANSI i-Standards. Action to; achieve this compatability has 4
already' been initiated through efforts of NRC's Office of
~
Standards Development.
Regarding other changes, the cope of the ASME survey and inspection system should be expanded to cover a complete product, including its operability, and not be limited to coverage of the pressure boundary only.
This change is essential if full potential reduction.ie duolicatien of inscections is to be achieved.
In additi~oii to 'the changes,. the ASME survey and inspection system must be evaluated to assure that the s system foiystem is eguivalent in all respects to the NRC avaliiEing and. approving QA program descri-aniffdr inspection of QA prograTf for implemen'tatidn'.piions
L
~_/
Finally, a[surance must be provided by ASME that a[Il activities under NRC endorsement will be conducted with appropriate p'ublic disclosure. -
A trial program to accomplish the changes and ~
'vfde evaluations outlined above should be undertaken in ensee program would be to provide a basis for full conjunction with ASME.
The objective of this
'were l
(qig recognition of the ASME inspection system.
Since refe"r-the success of this program will principally impact on LCVIP, IE should sponsor the program.
The effort
_Q will necessarily involve other offices of NRC.
To initiate the trial program, an exchange of
)lems.
correspondence between the participating parties i
setting forth agreed upon principles will be necessary.
A' draft.of such correspondence has er,/
l been prepared and is attached as Enclosure A to the
/
enclosed report One of the NRC required principles 7
is the granting of full audit privileges to the
/
NRC by the ASME.
It is anticipated that the results
/
of these audits would be an important part of the
/
basis for determining the equivalency of the ASME
/
system with the NRC system.
The proposed audit program would be conducted under LCVIP. Organizations which would be audited include the ASME, the National Board, the authorized inspection agencies and the state agencies that are active in the ASME code system.
The audit of state i-agencies would be necessary in those states in which the state acts as the authorized inspection agency under the ASME code and the authorized inspectors are state employees.
. r_r i
I t
(-
It is anticipated that the NRC acceptance of the ASME inspection system as a recognized third party inscection activity could be rec 0mmended en achievement of the follcwing three criteria:
1)
The ASME Code is revised as folicws:
s (a)
Scope is expanded to include requirements applicable to the total productT incl 0~ ding operability / function.
(b)
Requirements are expanded to achieve complete "fd3 compa.tibility with Apcendix B and the regulatory ag, guides endorsing the N45.2. series of standards.
N (c)
Procedures are expanded and documented regarding fC the survey, inspection and enforcement process.
du.
ct (d) Authorized inspectors are upgraded to ANSI N626 standards.
ems (e) Activities are _dqcumented[fjHpfro Fifti pUbTic j
c j
disclosure.
)
(f)
Class III components (safety-related) are appro-
../
priately addressed.
- 2) NRC reviews and formally endorses the revised Code and the ASME. inspection system procedures as achieving com-patibility with NRC's requirements and practices.
- 3) NRC audits the ASME inspection system activities under the revised Code to verify it is being implemented routinely in an effective manner.
(Until the revised Code is available, the base used for NRC audits will be NRC regulations and regulatory guides).
l The ACRS,. in its report on nuclear reactor inspection,*
I has made the folicwing statement:
"A fertile area for improving the reliability and scope of inspections is through improved interactions between Third Party (authorized inspector) and Fourth Party (NRC-IE) inspectors and acceptance by Fourth Party of Third Party inspections, subject to audit." The trial program proposed here is consistent with the ACRS recommendation.
.' 5 !. :.
- ACRS letter from Daca W. Meekler, Chairman to Marcus A Rcwden, Chair: nan USNRC dated May 19,.1976.
Subject:
Report on Nuclear i
Reactor Regulation i
1
(-
. Tne pros and cons of the alternatives under Issue 2 are identified belcw.
Issue 2 y
I Given that NRC should have a vendor inspection program, t.
how should the current LCVIP be expanded to provide adequate
/
inspection coverage of all vendors?
I Alternative 1
'ide
~ Coritihile LCVIP~~with current resourcei but inspect
~
1se:
all vendors. at lower frequency.
wern
~
Pro:
- 1) Under the current LCVIP, experienced r
-N.
fue:
personnel are available and basic inspection procedures have been developed.
r
- 2) No change in NRC resource commit =ent required.
Cc3
- 3) Con'tinues direct control by NRC of vendor inspection procedures and personnel.
Con:
- 1) Does not provide basis for increa' sed inspection coverage of vendors.
1
~
- 2) Does not provide basis for materially reducing duplicative inspections iequired of licensees.
Alternative 2 Increase cemitment ' f NRC manpcwer resources to LCVIP.
o i
' Pro:
- 1) Under the currenc LCVIP, experienced j
personnel are available and basic inspeccion procedures have been developed so the t
program can be readily expanded.
- 2) Continues direct control by NRC of vendor inspection procedures and personnel.
Con:
1)
Requires major increase in NRC manpower.
- 2) Results in duplication of inspections now performed unde qAS',d.E code system.
Alternative 3 Util.'::e third party inspection systems such as th'e ASME codeinspectionsystemtosuplemen'curreiftLCUP,fp5v'ided t
a_t.ria_1_progrant prover the. _ feasibility of the _ concept.
~
Maintairr resources at present' level.
I:
. ]
l Pro:
1)
Has potential for providing increased inscection coverage of vendors that is endorsed by NRC with no increase in NRC resource commitment.
~j 2)' Has potential for providing a basis for materially reducing duplicative inspections
/
required of licensees.
1 i
- 3) A trial program.with ASME paves the way for potential use of additiona1 third party yde inspection systems in other functional areas
'se such ~as fire protection inspections.
+ere -
4)
Provides incentive for changes in the ASME hh.t '
code requirements.
!Uci' Con:
- 1) Gives appearnace of proposing the delegation of responsibility to non-government organizations.
ems
- 2) Acknowledges a role of state agencies in regu-Tatton of nuclear ocwer plants.
Alternative 4
~
Utilize the ASME code inspection system td suppleme~nt LCVIP fpr the inspection of ASME vendors, provided a trial program demonstrates the acceptability of the ASME system.
Increase commitment of NRC manpcwer resources to LCVIP to provide adequate coverage of non-ASME vendors and contractors.
Pro:
1)
Provides timely inspection coverage of non-ASME vendors and contractors.
2)
Provides for test of third party concept.
3)
Provides same advantage as Alternative 3 but to reduced degree.
Con:
1)
Requires mod.est increase in current LCVIP staff.
- 2) May void incentive for future use of third parties to inspect non-ASME venders.
.55.. :.
3)
Provides same disadvantages as Alternative 3.
t 4
l l
f
. Alternative.1 under Issue 2 means maintaining the -
s
-/
status quo with no increase in inspection scope or-in LCVIP manpcwer levels. Alternative 2 increases
[
the inspection scope by utilizing directly employed
/
staff but incurs a significant cost in required resources. Alternative 3 has the potential for significantly expanding LCVIP impact without need ice for any increase in manecwer resources. Al terna tive r
sea 4 provides for both a test of the third. party ver,.
concept and a modest increase in directly emoloyed g?
staff. Alternative 4 is the preferred choice, hcwever because of current budgetary restraints, c-
- Ct increase in staff will not be possible before fiscal year 1979. As an interim Alternative 3 is recommended.
Recommendations; Accrove the Licensee Contractor and Vendor Inspection Program as a. centinuing part of,the NRC inspection program.
Accreve the cenduct of a two-year trial prcgram spensored by IE and conducted in conjunction with ASME to test the feasibility of utili,:ing third party inspection systems to supplement LCVIP.
Note that state agencies which act as authorized inspection agencies under the ASME cede would.be subject to audit under the trial program.
Note that the concept of including a charge in license fees for costs of LCVIP will be further developed during the trial program.
T 4
Sunshine Act Reccmmendation:
It is reccamended that this paper be considered in an open meeting.
OGC and OPE concur.
- _'_t '. '.
l I
' I t.
- IST '
~
._. Coordination:
Tne Office.of Nuclear Reactor Regulation, Office of Standards Development, and Office of Executive Legal Director concur in the recommendations of this paper.
The Office of State Programs has taken note of the proposed LCVIP-state agency interface And has no i
. objections. The Office of Administration has taken note of the license fee concept and has no objections.
See
/
OGC had no cements,.0Pg_ ccmments responded to by ver9 addition of Alternative t to Issue 2 and clarifying 3D information in Appendix E.. OPE considered Alternative (uct.
4 necessary ts give the papenbalance and to provide
,_ the Ccmission,with a full rancAgf_ckoip.es. ~
)
I
?ms.
Ernst Volg&snau -- _
t
'/
Director
/
Office of Inspection
/
and Enforcement.
'[
Enclosures:
-N
~ ~ ~. - ~ -
' ' ~
1.
IE Report on the-Licensee Contractor and Vendor Inspection Program 2.
OPE Comments Co=11ssioner ccmments or conse'nts should be provided directly to the Office of the Secretary by close of busi, ness. Tuesday, February 22, 1977.
DISTR'IBUTION Commissioners Comission Staff Offices 1
Secretariat
., gune,,
9 e
i
01% o u i 4 W '77 - H cd5 t : v1 DISCUSSION OF THE 0FFICE OF INSPECTION AND ENFORCEMENT'S LICENSEE CONTRAC. TOR AND VENDOR INSPECTION PROGRAM
,I.
Introduction The primary objective of the Office of Inspection and Enforcement Licensee Contractor and Vendor Inspection Program (LCVIP) is to fee improve nuclear plant safety by providing NRC surveillance of the sea f.
design and manufacturing activities of:
architect-engineers, nuclear steam system suppliers, and vendors f who provide products such as a
vessels, piping, valves, electrical components and instrumentation for nuclear plants.
7, The technique for accomplishing this objective is to assure that I-vendor organizations have acceptable quality assurance programs and i
are properly implementing these programs. Among the specific j
functional activities inspected in the LCVIP are the design, procurement, fabrication, testing, and the examination of the I
products.
In addition, part of the LCVIP consists of feeding back i
experience from operating facilities in order to improve design, fabrication, and testing of new components.
I
'~a/ In defining the LCVIP, licensee contractors refers to the nuclear steam syt um supplier (NSSS) and the architect engineering (AE) organizations.
For the purpose of this paper, the word vendor is used as an all inclusive term to describe all product and service t
supplying organizations including NSSS and AE's.
I
. A secondary objective of the LCVIP is to provide a mechanism for reducing the duplicative inspection of vendors required by. RC rules. Some reduction in re'dundancy in the area of quality assurance s/
program evaluation and inspection is pos'sible by' giving licensee,s the
/
./
option of relying on the LCVIP for program ~evaluationi'however, the~~
~
~
~
~
LCVIP is not designed'to ass'ure the adequacy of individual componen'ts i 'e c
or services (product acceptance).
Full rasponsibility continues to
~
sea
- 2 be placed ~upon licensees ~ and will not be e. hanged or modified by any
~
UC5 off the proposals discussed in this paper.
II.
History p.
Initiation of the Licensee Contractor and Vendor Inspection Program by the NRC Office of Inspection and Enforcement (IE) was motivated by several events experienced by the nuclear industry during the start-up of the first large power reactors in the late 1960's.
Prior to about 1969, IE's vendor inspection effcrts were minimal, primarily because of manpower limitations.
The position taken at that time ytas that all functic.lal areas could not be inspected; therefore, the inspection effort was limited to site activities.
The philosophy was that if problems occurred, start-up testing at tha reactor site would be sufficient to identify such problems and correction could then be effected.
..[
. The need to re-evaluate this policy became evident with the problems identified at the Oyster Creok reactor during the late stages of its cons truction.
Major qualiti problems were identified in the reactor-pressure vessel and in piping systems.
The installation of many second hand, non-pedigree valves was also identified. Most of these problems were traceable back to faulty designs, a failure to impose i 'e C
see appropriate specifications and a general lack of quality assurance av g
by licensees, supplier shops, the project administrators, and the C
project constructors.
Subsequent problems having safety significance were also experienced at several other nuclear facilities during this same time period.
The consequence of these adverse experiences was the realization that new nuclear standards needed to be written, old ones upgraded,
{
and all standards enforced.
The most significant standard that subsequently evolved and which had the greatest impact upon industry, j
was the 10 CFR Part 50, Appendix B criteria which introduced the quality assurance concept.
A second significant effect was the realization that to assure that quality is obtained, inspection of work and enforcement of standards, whether performed by licensees or the AEC (NRC), cannot always wait
r
' until final assembly at the site but must occur during all stages of design, fabrication and erection, regardless of where or when that work is performed.
The primary reason for this second conclusfon
~.
is that it is frequently impossible to make a repair of a component on site without compromising the final quality of the component.
Ic'e Appendix B clearly placed the responsibility for assuring both teo cep -
program and product quality upon each licensee or license applicant.
y
'eef-Since IE was insufficiently staffed to conduct an independent vendor inspection program, regional site construction inspectors were ts directed to evaluate licensee vendor inspection programs as a part of the licensees' over-all quality assurance program and to periodically accompany selected licensees on their inspections of f
selected vendors.
This was referred to as the " host-concept".
This concept was utilized during the early 1970's but did not work well in practice for a number of reasons.
It was discriminatory in that the one licensee selected as the host from several eligible licensee candidates for a specific inspection was expected to follow through on corrective action of generic type problems for all licensees.
Further, inspections were difficult to coordinate and administer and more importantly were ineffective in correcting problems, since the presence of the IE inspector tended to inhibit e
r
. the identification of deficiencies.
It should be noted that during the host inspections the NRC inspector did not conduct the inspections, y
rather he observed the licen'see's inspector who made the basic decisions as to the scope of the inspection, the depth of investigation and the determination of compliance with requirements by the vendor.
Under the present LCVIP this situation does not prevail since the rrce See NRC inspector conducts the inspection of the vendor 'directly without Grg.
h.e' the presence of any licensee representative.
(uc?
Since the host-concept appeared to be largely ineffective, a new Rs, approach to vendor inspection was clearly necessary.
In early 1973, l
a trial program of direct inspection of fuel fabricators by NRC f
l inspectors was initiated.
This trial program was motivated by a 4
series of reactor fuel failures attributed to quality problems at the manufacturing plants.
Inspections disclosed that some of the basic problems could also be attributed to design deficiencies.
This trial program proved to be highly successful in obtaining greater conformance to quality standards by the fabricators and a r
subsequent reduction in major fuel problems.
See Appendix A for p
additional details.
This program also served to demonstrate the soundness of the basic concepts subsequently adopted for the LCVIP.
ami,--*-==
wee +-+
4- * - - * *
' * * * =
l
l
. During this same time period (1971-1974) licensees of operating plants were required to report events and equipment malfunctions (referred to as Licensee Eve'nt Reports or LER's).
These reports continued to increase both in number, significance, and variety as additional large power reactors progressed through the start-up stage. Analysis of the causes of these events indicated that a lc'e I
j significant number could be traced to deficiencies in design or in fabrication - work that had been performed off-site during the NCO /
design and construction stages. See Appendix B for additional detail's.
gj In the same time frame a studyE was per formed by a special Regulatory I
Task Force which noted an increase in th' number of yeported e
problems and the difficulties experienced in performing inspections in the vendor area. Nsaconsequence,theTaskForcemadenumerous recommendations which supported the expansion of the trial inspection program.
Based upon the poor experience of the host-concept of vendor inspection, the success of the trial inspection program of fuel fabricators, and the recomendations of the Task Force, IE initiated b/ "The Study of Quality Verification and Budget Impact" (the Ernst Report), dated January 1974.
l t
i
=;;;;
...g
_7_
in 1974 a new two-year trial program of vendor inspection covering all types of off-site vendors and suppliers.
This program, the Licensee Contractor and Vend'or Inspection Program (LCVIP) was s
conducted to determine if enhanced public health and safety could be achieved by assuring higher quality vendor products and services;.
to identify and refine methods and procedures for conducting vendor
' C'e see inspections; and to identify NRC resource implications.
The program rer.
g f
met these goals.
It identified many weaknesses and obtained 7
?uC:-
substantive improvements in vendor quality assurance programs.
Specific details are identified in a White Book'(NUREG-0040-02)
Rs !
issued quarterly., Typical examples of inspection findings are
/
summarized in Appendix C.
Furthermore, effective procedures for
/
concentrating effort based upon the importance of the vendor's safety related product or service have been developed.
Finally, an
~
assessment of the manpower required to implement a program of vendor inspection has been made.
(See Appendix E)
III.
Description of the LCVIP The inspection of vendors is an extremely difficult operation since there are app.oximately one thousand vendors and suppliers scattered l
all over the world who are actively supplying products or services to the U.S. nuclear industry. No two vendors are alike, even though l
.:.i:- '
i
. they may be supplying a similar type product.
The standards to inspect against are frequently different depending upon the type of product, and many product lines have no suitable national standards.-
Furthermore, vendors are not licensed; therefore, the legal base for direct NRC inspection and enforcement is not entirely clear.N E
/
i From the realization of these facts and past experience, several fe, r
j se.
ver) /
basic concepts and practices have been developed and incorporated into the LCVIP.
Because of budget restraints, it is not possible
~
to inspect every product for every reactor nor even every v2ndor.
~
Therefore, rather than attempt to_ inspect all products and all vendors, priority of inspection is given 'to those vendors providing
/
important ' safety related products or services.
Furthermore, the inspections
[
are made on a generic basis; that is, attention is concentrated on the quality assurance system of the vendors because it is usually true that if this system is satisfactory, the products or services will be satisfactory.
g Section 206 of'the Energy Reorganization Act of 1974 and proposed 10 CFR Part 21 for implementation of Section 206 has prov.ided a _.
partial base for conducti.ng direct vendor inspections.
- q k
.~l
g _.
Selection of vendors for inspection is based upon the safety significance of the product or service supplied.
Operational experience shows that many s'afety deficiencies in products are-attributable to design deficiencies; therefore, vendors having a significant influence.on the design and construction of nuclear plants are inspected most often on a periodic basis.
These i 'e i
C Bee organizations include the five nuclear steam system suppliers Gr, he' (General Electric, Westinghouse, Babcock & Wilcox, Combustion Engineering, General Atomics) and the architect-engineering (AE)
}
firms.
Approximately fifteen AE's are currently active.
Examples ns' of the more prominent AE's are Bechtel, Ebasco, Stone & Webster, Burns & Roe, Gilbert Associates, and Gibbs & Hill.
In addition, approximately 120 product type vendors have been inspected or are subject to inspection based upon the vendor work load at any given time.
These include the major suppliers of most ASME Class I type equipment as well as other safety rele< ant parts or components.
Components in the ASME Class I category include pressure vessels, steam generators, pumps, valves, and major piping, including support hardware. Manufacturers of non-ASME type components that are inspected include suppliers of such items as reactor fuel, control rods, control rod drives, valve operators, electrical switchgear, instrumentation, containment cables and others.
S
'7'
, The frequency of inspection of all types of vendors varies from one to four times a year and is determined by the importance of the safety related product, the volume produced by a specific supplier
~
and the performance history of the supplier or his product.
LCVIP inspections for all types of vendors are structured around a P
fc'e three-year repetitive cycle without regard to specific manufacturing Be vera milestones such as those which exist in the building of a single h_t' fu 'i structure.
The three-year cycle is initiated by a detailed c
examination of the vendor's quality assurance program with emphasis Ws placed upon proper implementation.
Following this effort, inspections are focused on determining the effectiveness of the quality assurance program by sampling actual work quality whether that work is represented by a product, a design or a working type document.
These inspections are preplanned to cover suitable subject areas according to the type of vendor. After the three-year period of planned inspections, the cycle is repeated beginning with a re-review of the quality assurance program.
What is inspected, the bases used, and to a degree, how inspections are conducted.is largely determined by the type of vendor being
, inspected.
One major classification includes the NSSS and AE corporate organizations whose product is not hardware, but a design package consisting of drawings, specifications and procurement
l
~
. documents.
The other major classification is hardware suppliers (product vendors). This latter major classification includes the
.=
manufacturing divisions of the NSSS since specific products are produced that are manufactured to product type standards.
n The bases used for inspecting the NSSS and AE organizations are QA
/
f c'e
/
Topical Reports that are submitted by these organizations to the NRC
'See l
- ere
/
and which describe the quality assurance programs. These programs rig -
10cp /
are reviewed by the Office of Nuclear Reactor Regulat#on (NRR) for j
f content and conformance with the 10 CFR Part 50, Apper. dix B criteria ems [
and are formally approved.
IE inspectors use these approved programs as the base for verifying proper implementation.
Specifically, by l
interviewing key personnel and reviewing documents IE inspectors determine that detailed control procedures have been prepared in accordance with the approved program and are being followed.
The inspector does not attempt to review the adequa.cy of specific designs, rather he determines that the designer or design group is properly qualified, that the appropriate design input information (codes, standards, specifications) is being applied, that designs are independently verified (ANSI N45.2.ll)U and that the design d/ ANSI N45.2.11 Quality Assurance Requirements for the Design of Nuclear Power Plants.
This standard has been endorsed by Regulatory Guide 1.64 which accepts with certain exceptions the standard as an adequate
~
basis for complying with the pertinent quality assurance requirements of Appendix B to 10 CFR Part 50.
T
. activities are documented in accordance with approved prog > am procedures.
The primary objective is to determine that the design process is properly managed'and controlled.
The same technique is -
w
/
used in inspecting the preparation of drawings, specifications or f
procurement packages.
./
T 'e Inspection of product vendors is somewhat different in that hardware C
Tse Qt ~
is available to inspect and observe for quality thereby providing a
- er ducp more specific means for evaluating the effectiveness of the manufacturer's quality assurance program.
Once again, however,.
eos.
emphasis is placed upon determining proper implementation of a quality assurance program which meets the necessary provisions of a
10 CFR Part 50, Appendix B criteria. Additionally, most products are manufaccured to specific quality level or functional standards which provide a positive method for measuring quality.
While the IE inspector does not perform accept-reject inspections (product acceptatice), he will examine products and processes on a sampling basis as a technique for verifying the proper implementation of a quality assurance program, thereby enabling a conclusion which can generically apply to all contracts serviced by the specific manufacturer.
As a point of emphasis the direct inspection of vendors by IE does not divert responsibility for assuring product quality from the
71
, )
licensee to the NRC.
It is true that in the interests of promoting uniformity in the interpretation and application of quality standards, IE has given licensees the dption of accepting IE's inspections of -
quality assurance programs.
This option, however, applies only to those organizations (NSSS, AE's and Fuel Fabricators) who have made I
comitments directly to the NRC via a Topical Report.
Ir.iplementation
'ic'e l
?sec of this procedure reduces some of the current duplication of effort.
Were he' IE has made it clear to the industry, however, that the function of luc' ~l product acceptance, whether the product is hardware or a design package, i
must be pe'rformed by th_e licensee.
Each licerisee is expected 2ms to verify that he has received specified quality in each item in accordance with'his individual contracts.
At the present time the scope of the LCVIP is well defined and primarily limited to inspecting the activities of firms performing work away from facility construction sites.
One additional type of company that could be appropriately inspected under the LCVIP are the large construction firms (J. A. Jones, Kaiser Engineers, Inc.)
that are active nationwide and who electTo~ submit Topical"
~
Repbrt's fdF NRR7evis'TMa'ny 6f tlie ~activitiesif Tu~c'h firnis ~~
~
~
~
~
w may be generic to several projects ind future chns~iderition~ wil'1 b'e
~
~
~
given to inspecting these firms under the LCVIP.
4 l
l
, Another area where the LCVIP will be modified involves the bases used for the inspection of ASME product x ~
manufacturers.
To date subs'tantial reliance has been placed upon the ASME approved QA Manuals.
It is recognized that these Manuals do not always address all the commitments made by licensees (the buyers) to the ANSI N45.2 Standards or the NRC's Regulatory Guides.
,ide j
73c0 Additionally, the Manuals may not satisfy all the necessary provisions r
=ero Th e '
/
of the 10 CFR Part 50, Appendix B criteria.
Changes in the inspection duct ['
procedures utilized by the LCVIP inspectors are needed to require o
verification that the buyer's commitments c,onsistent with the necessary ems !
Appendix ~Bcriteriahavelbeenspecifiedincontractual'docum'entsand-
~
/
that the ' endors are satisfying these requirements.
v
/
-l A third area where LCVIP modifi. cation will be required is to provide for verification of 10 CFR, Part 21 implementation.
These modifications will be made when Part 21 becomes effective.
IV.
Third Party Inspection A.
Considerations for the Utilization of Third Parties The limited LCVIP has shown that there are weaknesses in vendor QA programs which can be improved.
This observation, along with a continuing high rate of equipment malfunctions evidenced by Licensee Event Reports (LERs) and traceable in part to vendor
. activities, demonstrates the need for more effective and broader contractor and vendor surveillance.
While 10 CFR Part 50, Appendix B requires lice'nsees to evaluate and monitor supplier -
quality programs and performance, this has not been effective and has resulted in substantial duplication of effort since each
/
/
licensee is inspecting the same group of suppliers.
To achieve 73ee i
the broader, more effective coverage required to provide more ter;
'br ~
adequate protection of public health and safety, NRC can broaden
~
uc its in-house inspection program or can search for alternate ways to extend NRC influence and surveillance.
.ms While in-house inspection may be desirable,*-( contiriuing restrictions
~
~
~
4
[
on numbers of directly employed staff suggest that NRC alone cannot provide as much coverage of the industry as desirable.
This coverage can, however, be provided through the use of authorized third party inspection systems.
By third party is meant any organization (other than NRC) which provides inspection services and which is not under the direct control or jurisdiction of a licensee, licensee agent, supplier, or manufacturer.
The principal advantages in considering the utilization of third party organizations are the potential for minimizing the size of e/ See Appendix E for discussion of broadened in-house inspection and the estimated staff requirements.
, directly employed staff, the broadening of the NRC's influence upon a larger number of vendors, a reduction in the duplication of effort currently requ' ired and the achievement of greater unifonnity in the interpretation and enforcement of requirements.
Ultimately these factors should result in the delivery cf higher
'fde quality products and services to facility sites by vendors.
'?seo
['.t
- 9
/
The principal disadvantages of utilizing third party organizations A
I due? '
are that there may be a perception of dilution of responsibility.of
~
~
licensees, less. direct contro1 of vendor quality assurance progran
'03 -
evaluation and _ iris ~pectiori activities, and the need to divert some
- ~
~
~
~
~
NRC resources t'o' the auditi_ng oflthe third paRf ac'tivities.
~
o 9
In comparing
--e potential advantages against the disadvantages, substantive benefits both in cost and safety can be obtained not only by the NRC but also by the nuclear industry and the public through the NRC utilization of third party organizations.
The American Society of Mechanical Engineer's (ASME) code inspecti. system is closely associated with the evolution of the third par ty concept.
This system could meet the NRC's basic requirements for third party organizations and appears to offer a high pote1tial for integration into the NRC's regulatory process.
, B.
Description of the ASME Code System Among the several codes developed in this country, the ASME's series of " Boiler and Pr' essure Vessel Codes," referred to here -
/
'/
after as the Boiler Code or code, is perhaps the most sophisticated, widely recognized and adopted code system in the world.
It is
/
also unique in that it has a built-in quality assurance program nsee I
evaluation and certification process as well as an inspection Wce:
i
.qhy -
and enforcement system.
Mu?
c I
j The ASME Boiler Code was developed not only as a safety measure
'em,//
but also to provide standards by which any State could be assured of the suitability of boilers for installation within its
,/
borders.
It was originally developed from the rules, regulations, j
.i or laws that had been established by several of the states in the fifty or more years before its first publication in 1914.
Since that time 44 of the 50 states and 10 municipalities have passed laws adopting one or more sections of the code.
These laws provide enforcement authority which makes application of the code mandatory in those jurisdictions for the design and
, manufacture of components and systems that are covered by the code..f]
.f.
As industrial growth occurred, the code was expanded to cover y
several types of boilers and other associated manufacturing operations. Additionally, since the late 1950's, the code was fic'e expanded to include nuclear activities.
Today's nuclear version esec Wre covers the manufacture of many of the major components of reactor Eht '
?du i systems having safety importance including the functions of c
I inservice inspection and quality assurance.
The NRC has
' ems recognized the importance of the code by incorporating its requirements for Section III, Division I, Class 1 components
[
and for inservice inspection for Class 1, 2 and 3 components J
into NRC regulations (10 CFR Part 50, Paragraph 50.55a).
Among the several attractive features of the ASME system is a manufacturer certification system, the "N-Stamp" authorization process.
Tied in with this process is an inspection and enforcement function that utilizes a staff of over 3000
.f] It should be noted that the ASME Boiler Code is generally limited to the design and manufacture of vessels, pumps, valves, piping and other associated mechanical parts.
Furthermore, on dynamic components (pumps, valves) the code addresses only the pressure boundary and not internal parts nor external attachments such as drive motors or operating mechanisms.
. authorized shop and field inspectors, approximately 10% of whom are engaged in nuclear work.
The N-Stamp authorization process was self-initiated in 1968 in response to the adverse experiences previously discussed, and is proving to have an increasing
/j influence in upgrading quality performance at the manufacturing l
level.
Ufde
?nse, cf.
To manufacture or install a code stamped product or to provide gn~ ;
'dh a code covered service in accordance with the ASME system, a e
company must have a contract with an authorized inspection agency
- Tem, (AIA) which provides code required inspection services and must have a Certificate of Authorization, more commonly referred to
/
as the N-Stamp.
Application for a Certificate of Authorization is made to the ASME.
Upon receipt of the application, the manufacturer is subjected to a quality assurance program review and implementation inspection by an ASME directed survey team made up of the following:-
ASME Team Leader ASME Member (Consultant skilled in specialized areas)
National Board representative Utility representative Authorized Inspection Agency representative Jurisdictional Authority representative (State) o/ The ASME Code provides for product manufacture or installation to several levels of quality.
For organizations performing Class 3 work only, the lowest quality level, less rigid procedures than described here are employed for issuance of a Certificate of Authorization.
20 -
The team reviews the applicant's quality assurance program, as documanted in his QA Manual, for compliance with the code and
=.. :.
inspects manufacturing o'perations against the Manual.
A typical-
~
,j team survey requires approximately three days to complete.
Upon j
i, completion, a team report of the survey is written and forwarded to the ASME.
The report is then reviewed by the ASME Subcomittee ins,
/
e on Nuclear Certification (SCNC) which may accept or reject the k' era j
, b.e ^
S survey team recommendations.
Upon acceptance a Certificate of Mu ?
c 7
j Authorization (the N-Stamp) is issued to the arplicant by the
/
/
ASME.
Renewal is required every third year, including a full team survey.
i
/
The SCNC is an officially authorized subcommittee of the ASME, established in accordance with the rules of the Society's charter.
/
It is made up of representatives from all phases of industry to assure impartial decisions on the issuance of certificates.
Utilities, states and NRC are also represented on the comittee, although the NRC member does not vote.
In addition to the accept-reject responsibility for Certificates of Authorizatian, the committee supervises the survey teams and establishes policies regarding their working procedures.
q d
e rm
. The effectiveness of the code is enhanced by several requirements in the code relating to inspection and surveillance of work performed.
The N-Stamp holder must have a contract with an
/
authorized inspection agency which is usually an insurance company authorized by a jurisdictional authority (state or vfg municipality) to write boiler and pressure vessel insurance.
.f*8 The authorized inspection agency provides qualified inspectors
. Grg M.C' who assure that the N-Stamp holders conform to code rules.
>du i~
c p
Insurance companies were requested to provide this service during the early days of code development because of' their unique status of being independent of both the manufacturer and the user of a product, yet having a financial interest in the product's
/
integrity and performance.
It should be noted that the state jurisdictions may also elect to act as authorized inspection agencies.
Another agency involved in the ASME code system is the National j
Board of Boiler and Pressure Vessel Inspectors (often referred j
to as the National Board).
The National Board is composed of f
I chief inspectors of states and municipalities in the United States and of the Provinces of Canada that have adopted the code.
In addition, the National Board has a full time staff that administers and directs the activities performed by the National Board.
4
f r-
, The principal functions of the t{ational Board are to promote the uniform administration of the code among the several jurisdic-
'l tional authorities; act 'as the central enforcement and investi -
[
gative agency for the code system; represent the jurisdictional
,/
j authorities in all code matters; conduct inspector training
/
courses and issue inspector "comissions".
- ivfg,
?nse, h '[e To become an authorized inspector, a candidate must meet age, ddci' education and experience requirements; obtain a certificate of competency from a jurisdictional authority; obtain a cc: mission fees from the fiational Board; and be employed by an authorized inspection agency.
A certificate of competency is issued by a jurisdictional authority (state) based upon a passing grade from a written examination.
The examination is prepared by the National Board, administered by the jurisdictional authority and graded by the fiational Board.
Upon receiving a certificate of competency, the candidate can then apply to the flational Board for a comission.
Upon payment of a nominal fee and verification of the validity of the certificate of competency, a comission is granted.
Renewals are required annually and must be requested by the inspector's employer.
If not renewed within a five-year interval, a re-examination is required.
i e
a The specific duties of the authorized inspector are spelled out in the ASME code and supplemented by a new series of ANSI N626 standards.
It should be' noted that contrary to popular belief,-
the code inspector does not perfom " product acceptance" nor Y
/
/
does he apply the "N-Stamp" to products signifying compliance i
/
with the code.
In practice, his overall duties are to audit the f3ee manufacturer's quality assurance program, the work performed, L'er ;
(7b
- and to witness certain prescribed tests and examinations.
The Muc~:
primary responsibility for the product meeting code requirements i
and for applying the N-Stamp rests with the manufacturer.
The code also requires the manufacturer to prepare a " Manufacturer's Data Report" which must be delivered with the product and be signed by the authorized inspector.
This document provides the inspector the enforcement authority needed to assure compliance with the code.
Another important point to note is the potential involvement of states in nuclear matters through their participation in the ASME Code System.
They play an important role in certification of inspectors, ar.
.s previously mentioned, may elect to act as j
the authorized inspection agency.
In states where this occurs
I'
. (e.g., California, Oregon, Wisconsin, N. Carolina, and others),
the code inspectors are state employed.
Under present NRC
'-i regulations (10 CFR Part' 50, 50.55a) application of the ASME i
Code System is required for the assembly of important safety j
.! systems and components at reactor construction sites and for inservice inspections conducted in accordance with ASME Code,
- 80 Section XI.
This means that an agressively inclined state Mr MI through the code has a legitimate path for involvement in code sc'Uck' activities at nuclear facility sites.
To date, state involvement y
at facility sites has been limited and real jurisdictional problems minimal; however, the code system offers one way by which state-federal interface can occur.
C.
Proposal for Use of ASME Code System as Third Party In conducting vendor inspections, it was recognized early that the NRC program interfaced directly with the ASME Code inspection system since many of the inspections involved the same manu-facturing organizations and facilities.
It therefore appeared appropriate to make use of the ASME system to the extent practical to minimize duplication.
Fr m observation of the ASME Code system, it was apparent several characteristics needed modification or upgrading to make the system more compatible
& with AEC (NRC) requirements.
The need for change was also increasing since the ASME Code System was tending to infringe upon AEC (NRC) jurisdict' ion by extending its coverage into additional areas including site activities.
With pressure from W,
. industry to reduce duplication of effort and NRC's desire to I'
y g.
increase safety by using properly controlled code inspectors, a
[8"e favorable climate for discussion was established.
Initiatives
..ar, 14b.0 were taken by both the ASME and the AEC (NRC) to establish a 4c'u '
c dialogue.
Four meetings were held (November 21, 1972, October 11,
,~3 1973, October 21, 1974, August 11,1976) to exchange ideas and communicate on possible changes that could be made in the ASME r,
system to make it more compatible with NRC requirements. The ASME was willing to make significant changes in policy, procedures, code rules and in organization in return for a more positive endorsement of the ASME system by the AEC (NRC).
The ASME considers NRC endorsement necessary to strengthen the ASME's capability for making the code system more effective, for obtaining the cooperation of the varied participating parties i
in the code system and for assuring the high degree of unity needed in making the system work congruently with the NRC's regulatory system:
, As a consequence of these discussions, several significant changes were made in the ASME system.
The more important ones are as follows:
'l
[
,/
1.
To assure compatibility of tM two systems, both must operate
/
/
from a common or an equivale.it standard.
This meant that the
[,#
quality assurance requirements (Chapter NA-4000 of the ASME Y
(j 9 Section III Code and 10 CFR Part 50, Appendix B) must be made Ddue?
equivalent.
Substantive changes were made in NA-4000 and near equivalency has been achieved within the scope of iecs activity covered by the code.
7,
/
2.
To assure that Certificates of Authorization would be isned
//
on an objective basis without conflict of interest, the Su - a j
comittee on Nuclear Certification (SCNC), whose respo'nsibilities
/
were previously described, was established.
To demonstrate
/
the ASME's good faith in these matters, the ASME requested
//
,/
that the AEC (NRC) be represented on this committee.
3.
The AEC (NRC) considered it essential that the basic qualification requirements of the code inspector be upgraded u
h_/ NA-4000 generally incorporates the Appendix B criteria; however, because the scope of the ASME Code is limited (design and manufacture only),
it is not an all-inclusive standard directly comparable to Appendix B.
_ =. =
.=.
i e
. and greater codification of the duties of the inspector, his supervisor and his employer be effected.
The ASME established a special committee to write such rules which has culminated.
/
in the ANSI N626 series of standards.
These standards are
/
approximately 75% complete and are being made enforceable 3g by cross-referencing in the applicable ASME codes.
Training
- e 8
of personnel is in progress and compliance with these standards Mer.
7h{
is anticipated as they become mandatory.
t du,:
c 4.
To permit recognition of the ASME's system of certifying and Iec inspecting manufacturer's quality assurance programs as a s
suitable substitute for that performed by each licensee under
/
Appendix B, the utilities as a group should be more broadly
{
represented in the ASME decision making process. A utility representative was added to each survey team to achieve this
,/
objective.
In addition, a representative from both a private and a public utility were appointed to the SCNC.
5.
To avoid real or apparent conflicts of interest, ASME survey team member qualifications and past work associations are more carefully screened.
In addition, ASME team members are required to sign contracts that prohibit them from consulting j
on a private basis with clients on code related quality i
assurance matters.
l
-l
, 6.
Under code rules, a nuclear facility owner is expected to perform certain duties.
Rather than require a utility to be subjected to the complete ASME evaluation process, the ASME has agreed to recognize the NRC's quality assurance program evaluations as evidenced by possession of a construction permit.
7.
The ASME extended its certification process to material manufacturers and suppliers.
Because of the large number and variety of such organizations direct NRC surveillance would be prohibitively costly.
Therefore, recognition of the ASME system offers a method for influencing a lower level of vendor.
These changes have been accomplished and have been actively implemented for a sufficient length of time to demonstrate their benefits.
Though meaningful and substantive changes in the ASME code system have been accomplished, additional changes and verifications are considered necessary before integration into the NRC's regulatory process can be accomplished.
The principal items that need to be addressed are as follows:
i i
1.
The scope of the ASME QA requirements needs to be broadened.
Complete compatibility with the 10 CFR 50, Appendix B criteria as well as'the guidance in Regulatory Guides endorsing the ANSI N45.2 Standard series is necessary.
Safety related_ Class III components should be appropriately addr'essed.
~
2.
The scope of the ASME requirements 'as applied to products, including pumps and valves, needs to be expanded to cover more than the pressure boundary.
These requirements should provide rules for the total product including operability or functionability as appropriate.
3.
Verification that the ASME vendor qualification system (N-stamp system) and the accompanying inspection and enforcement process will satisfy NRC requirements is needed.
This includes the following factors:
a.
Verification that tha survey process covers the entire
~
scope of a Certificate of Authorization applicant's qualifications and operations to assure conformance to the applicable QA requirements.
b.
The development of more formalized and documented procedures for carrying out the survey function. Accept-reject criteria are a necessary part of these procedures to assure uniformity of applicatio.'.
-i
30 -
,c.
Survey team reports must include sufficient evidence to support the action of granting a Certificate of Authorization (N-Stang).
d.
Upgrading of the authorized inspectors to the ANSI N626 Standards must be accomplished.
4.
All activities including documentation must be made available for appropriate public disclosure.
To accomolish these actions and verifications, IE proposes a two-year trial program to permit the ASME system to accomplish the changes outlined, to allow a sufficient length of time for the responsible NRC Offices to evaluate the viability of utilizing the ASME vendor certification and inspection system, and to enable the identification and correction of problem areas that might arise from implementation.
The ultimate objective is that at the end of the two year program the criteria for acceptance will be met. At that time companies having Certificates of Authorization (N-Stamps) which were issued on the basis of the broadened QA 1-requiremens: would be recognized as meeting NRC requirements.
Licensees would then be given the option of relying on the ASME f
system for QA program evaluation and program inspections of those companies holding Certificates of Authorization, similar to the procedures currently applied to those submitting Topical Reports. ""
~
A subsfanifaf~ reduction in the inspection re'dundancy currently ~
~
~
~
~
0
.3,
occurring can thereby be accomplished.
To initiate the trial pr,ogram, draft correspondence has been prepared (Enclosure A) which spells out the responsibilities and actions t.". : theparticipatingorganizations,(NRC,ASME, National Board) will assume.
Public announcement of the trial program following an exchange of correspondence setting forth agreed upon principles would formally mark the beginning of the program.
Noteworthy and basic to implementation of the agreement is the granting of audit privileges to the NRC.
Designated personnel currently assigned to the LCVIP would be reassigned to perform audits of those phases of the ASME system which are associated with the certification and inspection of manufacturers and other organizations performing code work.
Organizations which would be audited are the ASME, the National Board, the authorized inspection agencies, and state organizations that are active in the code process.
NRC audits would be performed of the administrative rules used by the ASME and the National Board in managing the certification and investigative activities of the code and the ANSI N626 standards for the authorized inspection agencies and the states to evaluate their consistency with NRC requirements and the effectiveness of implementation.
e
, Even though the code system has many built-in preventive measures to avoid conflict of interest situations, the NRC auditors will be instructed to place p'riority in identifying such circumstances.
Responsibility for corrective action on these types of problems if identified would rest with the ASME or the National Board.
Should correction not be obtained, the trial _ program can be unilaterally terminated at any time.
Difficulties on this type of problem are expected to be minimal, as evidenced by the attitude and performance to date of the organizations associated with the ASME system.
V.
Benefits and Costs A.
Benefits Approximately 1,000 vendors serve the nuclear industry. Tnere are five nuclear steam system suppliers (NSSS), seventeen architect engineers (AEs), five fuel fabricators, about 350 N-Stamp holders engaged in nuclear work, and approximately F '0-600 miscellaneous product vendors not covered by the ASME code.
The primary motivation for an NRC vendor inspection program of this diverse group was to identify safety significant problems, particularly those of a generic nature, as early as possible to l
8.-
.. ~.
?*
33 -
minimize impact on ultimate plant quality and, therefore, better protect public health and safety.
Such problems do exist.
They are continually manifest in reactor operating facilities by the-LER system and in vendor plants by the LCVIP.
Examples of each follow.
Significant problems identified at facility sites that a comprehensive vendor inspection program might have avoided are (1) thin wall valves, (2) over stressing of safety valve headers resulting in ejection of the valves (faulty designs), (3) use of marginal materials, (4) electrice.1 switchgear and relay malfunctions, (5) out of code weld defects in components and others.
Each of i
these events resulted in extra costs to correct.
At vendor facilities, examples of safety significant problems identified,during vendor inspections include: (1) the sensitization of reactor pressure vessel safe-ends, (2) the use of a marginal welding process (electro slag), (3) major discontinuities in pressure vessel nozzle welds (Hatch 1, Shoreham, others)',
(4) improper heat treatment of stainless steel piping resulting j
i-in excessive sensitization and in other materials, (5) failure F
to meet nondestructive test acceptance standards (valves, pumps) and others. Appendix 0 provides further details.
o ere.
,-u-e g
+
y
I NRC experience in inspecting quality assurance programs has shown that improvements in licensees' quality assurance programs reduces problems in plan't operation.
By analogy, correction of -
similar quality assurance program problems in vendor shops through inspection will undoubtedly minimize problems during a plant's operational life.
The resultant benefit to public health and safety is not possible to quantify.
Neither is it possible to quantify the economic benefit attributable to vendor inspection, because there is a six year lag between the time when a vendor's process is improved and the time when a resultant product is operating in a plant.
However, the magnitude of this benefit can be appreciated by considering that one lost hour of production of a 1000 megawatt capacity reactor represents between $15,000 (at 15 mills / kilowatt-hour) and $25,000 (at 25 mills / kilowatt-hour) in lost revenue.
Expressed in another way, interest costs alone on a 1000 megawatt reactor are approximately $8500/ hour (at a capital cost of $750/ kilowatt capacity at 10% interest rate).
Even without precise quantification, the foregoing analysis shows that a sound regulatory program which includes vendor inspection results in plants that are inherently safer, operate trouble free longer, and are more economical to construct, to operate, and to maintain.
The degree to which these benefits are attained
, depends primarily on the number of vendors that are covered.
While there is not necessarily the same marginal benefit for each vendor, the overall-benefit is certainly non-diminishing as vendor coverage is increased.
B.
Cost and Cost-Benefit The NRC's cost of conducting an LCVIP can be expressed in terms of manpower.
The fiscal year 1977 manpower requirements for the vendor inspection program are itemized in Table I.
As an example estimates are provided with use of the ASME as a th,ird _
~ party if'this ~ option were available~.
As can be seen'in the Table
=
-seven' people would continue to be devoted to' ins'pection of NSSS
~
jand AElbut fi've 6ffeight cirrently inspecting ASME produ~ct~'-
~~
~ ~
l
~Teridors would]ie~ Feallocated 'to monitoring the ASME program.
The LCVIP without the use of the ASME code inspection system during the year June 30, 1975 through June 30, 1976 inspected five nuclear steam system suppliers, eight AEs, five fuel fabricators, 73 ASME product vendors (N-Stamp holders), and 18 non-ASME product vendors.
In contrast, the LCVIP supplemented by the ASME system could expand coverage of the ASME product vendors from 73 to 350, an approximate fivefold increase ~.
Since increased benefit is positively related to increased coverage, b
.,e
36 -
Table I Vendor Program Manpower Reauirements Fiscal Year 1977 Functional Area W/0 ASME With ASME Inspectors of NSSS/AE 7
7 1/
2/
Inspectors of ASME Product Vendors 8
3 Inspectors of Non-ASME Product lendors 4
4 Supervision & Management 5
5 Staff Surpart 8
8 SubeTotal 32 27 3/
Monitor ASME 5
Total 32 32 1/ Eight inspectors inspecting approximately seventy (70) ASME product vendors (s20%) one to three times per year (90 inspections). Total population of ASME product vendors is approximately 350.
~~ ithi~ihpWhal,ofineThiil pFb~g}idi, NRC ids'p~ec~ti6n~6f ASME vendo.rs would W
~
~
~
~
2/
be reduced to total of 45 inspections per year of approximately 30 suppliers (s8%) of population).
These inspections would serve as check upon the performance of the ASME inspector as well as the vendor.
3/ ~ With'~ approval"of t,he trial program,' five ~ inspectors would be diverted to
~
auditing of the several organizations involved in the ASME system to
+=
assure accept:Jele level of performance.
r
~
~
I
. )
it can be seen that supplementing the LCVIP with the ASME system resources could result in significant benefits with the same NRC manpower investmenti VI.
Reducing the Redundancy of Inspecten_s_
A major licensee and vendor objection to the NRC's LCVIP is that to date it has not brought about a significant reduction in the duplication of inspections required by Appendix B.
In the case of those organizations submitting Topical Reports to the NRC (NSSS' and AE's), the option of relying on the LCVIP inspections for QA program evaluations has been permitted and there has been some reduction. O However, since product vendors are not required to make commitments directly to the NRC via a Topical Report, current practice does not provide licensees with the same option for inspection of such companies.
One possibility for reducing duplication of inspections of manufacturers of ASME coded products during the trial program period would be for i
IE to concentrate attention in those areas that are not currently j.
considered to be adequately covered by the ASME code system.
By combining the LCVIP activities with the ASME program evaluations and 1
^
i jf Credit is only allowed for the review of quality assurance programmatic
)
activities.
Product acceptance remains the responsibility of the
)
individual licensee.
1 1
m m
~
c m--e e-
l
. inspections which are considered adequate, the necessary requirements of Appendix B could be satisfied.
Licensees could then be given the option of relying upon thess joint efforts for fulfilling their responsibilities.
Similarly for manufacturers of non-ASME coded products, the LCVIP inspections utilizing the necessary Appendix B criteria as the inspection base could be utilized to reduce inspection reduncancy.
These options are presented only as possibilities.
No plans have been made to implement them nor are any such plans being considered.
The purpose in presenting them is to provide a basis for discussing the potential cost savings from reduced inspection redundancy which might be realized under current conditions.
VII.
Cost Recovery As discussed, with the utilization of the ASME Code inspection system, NRC personnel could concentrate their inspection of A.AE product vendors on that part of the vendor's quality assurance program which i
is not covered by ASME. As noted above, this combined coverage could enable NRC to grant relief to licensees from the applicable provisions of 10 CFR 50, Appendix B, which currently require the licensees to perform these reviews.
Thus, the licensees would be able to reduce their inspections and avoid a cost which they.
currently incur.
. This cost can be estimated by considering that as of August 1,1976, 40 utilities were actively engaged in the construction of 94 reactor units (CP or LWA had been issued).
Each utility, either directly or through a procurement agent is conservatively estimated to have contracts with the following number and types of vendors.
Number Type of Vendor 1
Nuclea-Steam System Supplier 1
Architect Engineer 1
Primary reactor pressure vessel manufacturer 3
Other pressure vessel or storage tank manufacturers 3
Steam generator and heat e'xchanger manufacturers 3
Pump manufacturers i
5 Valve manufacturers 5
Piping and support fabricators 3
Manufacturers of speciality equipment (Control rods and drives, fuel handling equipment, etc.)
l 1
Fuel supplier 5
Instrumentation suppliers i
3 Electrical switchgear, cable suppliers I
1 Service supplier (nondestructive examination)
Total 35 A typical three day audit or inspection by a licensee is conducted by a three man team at an average cost (travel cost and expenses only) of $1000.
Granting relief to licensees for quality assurance program review would. permit the elimination-of one audit per -year of-each
=
=
. i vendor resulting in a saving of $35,000 per year to each utility.
In addition to the travel and expense cost, a minimum of three people are required to be maintaine'd on the payroll of each utility to carry out the 35 audits per year.
Conservatively estimating this cost at $50,000/ man, the cost saving is an additional $150,000 for an overall cost avoidance of $185,000 per year to each utility.
c Since there are 40 utilities involved, this represents a $7,400,000 savings to the industry.
For NRC to conduct the LCVIP at proposed FY-77 staffing levels, a cost of $1,615,950 is incurred.
This cost represents 19 inspectors at an average cost of $85,050 which includes salary, benefits, travel, supervis~ ion and management. housing and all supporting service costs.
Since there are na additional costs incurred for the ASME program (it must be done anyway and is already passed on in the price of ASME vendor products), comparison of these costs shows that not or'v 6
does conducting an LCVIP, supplemented by the ASME system, benefit public health and safety in a cost-beneficial way, but it also reduces the overall cost of regulation by not requiring a large inhouse inspection staff.
This conclusion can be drawn even without considering the intangible savings which the vendors realize through a reduction in duplicative inspections and uniform application of U
qaality assurance requirements.
E
-. ~ _...
i
a.
By achieving this overall reduction in the cost of regulation the progr.x,a provides a basis for paying for itself.
During the two year trial program, IE propo'ses to explore various ways to recover -
NRC costs through license fees.
This task will be conducted in coordination with other affected NRC organizations.
VIII.
Summary NRC needs to conduct a licensee and contractor vendor inspection program to assure adequate protection of public health and safety.
An LCVIP can identify generic vendor problems and bring about improvements. Additional reasons for this program are:
1.
Identification of equipment problems at the facility site may be too late to effect adequate repair.
2.
Quality and reliability of reactor systems begins at the design and fabrication stages. Approximately half of.the cost of a nuclear facility is incurred in these stages and performance of the work carried out at off-site vendor locations.
3.
Total reliance upon utility surveillance of vendors has not been l
effective as evidenced by nuclear facilities failing to perform i
as predicted which has been caused by a continuing excessive i
rate of equipment malfunction.
P i
4.
public confidence in the NRC regulatory effort is dependent to some degree upon direct inspection.
Since growth of nuclear power is public policy, ' greater pressure for inspection can be -
anticipated.
Ve'ndor inspection is an inherent part of the total inspection effort.
5.
The NRC has a "need to know" of what is going on and the problems being experienced in the nuclear supply industry in order to assure that effective corrective actions are taken on a timely basis.
6.
The NRC' needs staff expertise to recognize, understand, investigate and achieve appropriate resolution of equipment problems before those problems threaten public safety.
This is best achieved through a permanent full time vendor inspection staff specializing in the technical areas associated with reactor components.
7.
The NRC needs an independent capability for identifying component problems and assuring compliance with specified standards.
8.
There is a need for NRC visibility in vendor shops to motivate improved vendor performance.
9.
Utility surveillance of vendors has been ineffective in part because their authority and access to information is limited to 5
~
, - g_-
. the scope of a contract.
Generic problems frequently cannot be identified or corrected because of this limitation.
NRC inspectors are not limited in this ' manner.
~
~
10.
NRC inspection of vendors is more objective than buyer's inspections since costs and delivery schedules do not influence findings.
Product quality and facility safety are thereby enhanced.
11.
NRC inspection assures unifonnity in the interpretation of codes and standards and compliance with those requirements.
In addition, through the NRC's evaluation and acceptance of quality
~
~
assurance programs, the licensee can be giverj the option of relying-E upon NR[for program'~ evaluation wh'ile h'e retains full responsibility.~ ~ ~ ~
^
~
~
~
~
"for_ product accepta_nce.. _ As_. discussed,. the_ cost benefit of. tihi.s ~
l
-arrarigameci,TTi~s~u~bstia~ntialI Eo~~the individual licenseeIas~well as j
~
~
~
~
to' ihe entire industry.
j
~
~
On balance, the basic question is whether a direct government inspection program should replace the individual programs conducted by the several licensees.
Based upon the severity of the problems experienced in the past as well as a continuing incidence of malfunctions, licensee programs have not been as effective as desired.
The appropriate corrective action does not appear to be the elimination of licensee
_ 44 _
programs, nor a complete take over by an NRC program, nor a delegation of the function to a third party; rather the optimum resolution appears to be a combination " arrangement which better defines the responsibilities of licensees to eliminate duplication, and the continuation of a modest NRC vendor program supplemented by third parties.
Several independent studies (by the ACRS, The Office of Policy Evaluation, the Director of Regulation's Special Task Force in 1974, and the Atomic Industrial Form, Inc.) of this subject area support the general conclusion that the NRC should have a vendor inspection program.
The size of the program when discussed by the above groups, ranges up to a total of 90 inspectors.
Our estimated long term needs are approximately half that depending upon the
~
results of the two year trial program to test the feasibility of the
~ASM.E.as a third party.
In our yiew a credible and effective program
~~ 'n be addinistered with' this number.
~~
~~
ca IX.
Recommendations
~ Approve the' Licensee Contractor and Vendor Inspection Program as
~
a continuing _part of the NRC inspection. prog. ram.
_Approye_the_ conduct.of a two year. trial orocram sponsored by IE.
l and conducted in conjunction with ASME to test the feasibility of utilizing third party inspection systems to supplement LCVIP.
L L..
~: -
}
Note that state agencies which act as authorized inspection agencies under the ASME code would be subject to audit under the trial program.
~
Note that the concept of including a charge in licensee fees for costs of LCVIP will be further developed during the trial program.
1
- 7.,
L
>c i
D f
b T
~
Appendix A Quality Problems Identified by Trial Inspection Procram of Fuel Fabricators (1973-74)
?..
In the early 1970's as growth of the nuclear industry occurred and significant quantities of reactor fuel were used, a series of fuel failures and deficiencies were experienced (up to s0.8% of one fuel vendor's fuel failed in service). Although the primary impact of these problems was economic rather than safety (i.e. plant restrictions, unscheduled shutdowns, lowered availability, etc.), high fuel failure rates violate a basic NRC requirement "to maintain.or reduce radioactive materials in effluents to as low as reasonably achievable".
Early in 1973, a pilot QA inspection program of fuel vendors was initiated by IE to provide a means for obtaining information relating to the problems i
and to establish an inspection capability in the fuel area.
Initial inspection findings of this trial program identified several common deviations from requirements, problem areas and weaknesses in the fuel vendor's QA programs.
The most significant deficiencies were found in the following general areas:
Failure to adequately sta'f, separate and define QA department (and 1.
f personnel) responsibilities and authorities.
2.
Failure to provide adequate and definitive instructions and procedures i
to QC and fabrication personnel at the production level.
3.
Failure to properly qualify and document special roanufacturing processes.
- .~
. f 4.
Failure to qualify personnel.
5.
Weak and deficient subvendor audit programs.
6.
Weak, deficient and inadequate design control programs.
s:
ff Numerous other deficiencies (91 specific items) were also identified but
[b are considered of less significance than the main categories listed above.
i Fuel vendors were responsive to inspection findings by making corrections 1
and overall improvements in quality assurance programs and in upgrading
+
their staffs.'
Considerable uniformity between vendor programs has been
[
achieved.
Some of the most significant improvements were as follows:
1.
All vendors have now implemented specific detailed qualification-procedures and practices for both process and personnel qualifications.
p 2.
Generic fuel design control problems and weaknesses have been resolved.
As a result of the IE inspection program, more attention has been given to design quality assurance and a better balance between design E
l-and fabrication QA has been achieved.
3.
Through IE inspection efforts, strong quality assurance requirements are being extended to all sub-levels supplying components, materials, or services.
Principal fuel vendors are currently verifying quality requirements by more extensive audits of sub-vendors.
l
. 1
l
,. 4.
As a result of [E inspection findings and emphasis, more vendor management attention has been given to their internal audit programs e+
with demonstrated improvemen'ts in this area.
5.
In cooperation with IE suggestions, efforts are underway at most fuel vendors to streamline and reduce the procedures to prevent duplication and unnecessary complexity.
This is helping to achieve better compliance with procedures.
6.
At fuel vendors where inadequate QA staffing and/or definition of authority and responsibilities was encountered, additional personnel were hired, more staff training was provided and, where necessary, the QA staff was given more authority and responsibility in the day to day operations.
Although current fuel performance experience is difficult to quantify
~
+
and compare to earlier fuel failure data, current overall fuel failure rates are estimated to be at least a factor of two lower than when this inspection program was initiated. Much of this improvement in performance can be directly related to improved QA programs motivated in part by the NRC inspection effort.
l l
l' 2.-
Appendix B Analysis of Causes of Equipment Malfunctions In March 1974 an analysis of 654' equipment malfunctions reported to the -
NRC in the preceding year was conducted to determine the relative distribution of product type malfunction and the basic sources or causes of the malfunctions. The latter were categorized into design, manufacture, construction or maintenance functions.
Malfunctions or errors involving deficiencies in calculations, stress analysis, selection of material or misapplication of a component were attributed to design.
Defective materials, leaks, cracks or failure to functionally perform as designed were classified as manufacturing deficiencies.
Errors involving alignment, orientation, handling, failure to provide adequate physical protection were considered construction deficiencies.
Operational or maintenance deficiencies included corrosion, seal problems, improper calibration and others.
The results obtained were as follows:
Major Catecories of Malfunctions 1.
Mechanical components or systems 268 or 41%
2.
Electrical components or systems 161 or 25%
3.
Electronic and instrumentation 155 or 23%
i 4.
Personnel or procedural deficiencies 70 or 11%
654 100%
=
- =
~
a=Me
- e**-<
I
^
, p Malfunctions attributable to equipment type were as follows:
,. g..
":Z:
ASME Covered Components (Mechani, cal) p 120 Valves (all types) 43 Piping h
18 Pumps 15 Pressure vessels 6.1d tanks i-7 Steam generators and heat exchangers 203 or 31 %
Non-ASME Covered Components _(Mechanical) 21 Diesel generators 35 Control' rods 6
Fuel 3
Ve'ntilating components u-65 or 10%
Electrical Components _
76 Pump motors and valve operators 85 Switchgear, relays, etc.
161' or 25%
Electronic Equipment and Instrumentation I
80 Set point drift
- .g.;
75 Instruments, alarms, circuits 155 or 23%
- .y 1
3-Personnel Errors
+=
':.y...
All types 70 or 11%
~
Malfunction Cause Categories
~"
Design 152 or 23%
i Fabrication 81 or 12%
i u
Construction 131 or 20%
Maintenance and operations 290 or 45%
^
654 100%
=
Noteworthy is that 35% of the malfunctions of all types were traceable to design and fabrication, work which is performed by a vendor away from j.:-
the facility site.
..e
[..
t t
- ,7 ;
t Appendix C Deficiencies Identified in Vendor Quality Assurance Programs by the LCVIP (1974-76)
Numerous deficiencies in vendor quality assurance programs and in program implementation have been identified by the LCVIP during the 1974-76 time g
period.
Because of the large number and diversity, individual deficiencies are not listed.
The following summarizes the principal types and categories of deficiencies identified:
1 A.
Management Deficiencies 1.
Failure to conduct audits-internal and external.
2.
Inadequate or no training programs.
i r
3.
Failure to honor or implenent commi.tments.
4.
Failure to follow through on corrective actions.
5.
Failure to establish and implement adequate control systems l-I (design, documentation, procedural development, personnel qualification,procurementprocess,etc.)
6.
Failure to impose the same requirements upon internal product divisions as outside suppliers.
B.
Procedural Deficiencies 1.
Failure to qualify special processes.
2.
Failure to follow qualified procedures.
. b$
.4
2-3.
Use of unqualified suppliers.
g 4.
Use of faulty heat treat. ment procedures (results 'n change of material properties which has potential impact upon product quality and performance).
5.
Failure to calibrate process equipment (heat treatment furnaces, welding equipment, environment controls, tools, etc.).
6.
Use of unqualified personnel (welders, nondestructive test personnel,etc.).
7.
Use of wrong material.
Failure to control materials (identify, storage, exposure to 8.
incompatible materials, etc.).
9.
Failure to perform specified measurements, tests, examinations.
C.
Record Deficiencies 1.
False, incomplete and missing records.
I 2.
Failure to update manuals, instructions, procedures.
l 8
3.
Lack of traceability of materials, work through records.
l e.
,l
Appendix D Safety Significant problems Identified by Vendor Inspection 1.
Sensitization of reactor pressure vessel safe-ends.
Because the materials used in the reactor pressure vessel and the n
connecting reactor piping are different, a transition piece called the safe-end is attached to the vessel nozzles by the manufacturer.
The safe-end is usually made of a stainless steel which is compatible with the piping material. One of the stainless materials (Type 304) frequently used is susceptible to sensitization, a condition which results in the material becoming more prone to stress corrosion cracking in the operating environment.
Safe-end sensitiiat. ion may i
be caused by improper or untimely heat treatment during vessel j-manufacture.
These conditions were identified by an AEC (NRC)' vendor inspector during the late 1960's. Approximately fifteen reactor vessels required subsequent replacement or repair of the safe-ends as a result of this finding.
I 2.
Use of a marainal weldino crocess (electro slaa).
Electro slag is a metal joining process which it m owhat analogous to casting.
Use of the process was observed in the manufacture of reactor pressure vessels.
The problem in applying it to reactor pressure vessels is that special heat treatment of the finished weld is required to assure acceptable tensile and impact properties.
I
, The effectiveness of the heat treatment is limited to vessel wall thicknesses no greater than 6".
Since vessel wall thicknesses frequently exceed 6" and tes' ting of the deposited weld metal is not -
required, application of the process to reactor pressure vessels was b
marginal.
Use of the process was subsequently discontinued.
I-3.
Major discontinuities in oressure vessel nozzle welds.
[
During a two man NRC inspection of a reactor pressure vessel manufacturer, l
i' both inspectors identified independently the fact that the manufacturer i-had been having difficulty in meeting weld acceptance standards on the main piping nozzles.
Investigation disclosed the problem to be generic to a number of vessels.
As acceptance of the welds is dependent upon interpretation of radio graphs, unacceptable defects are not always identified.
Subsequent to identification of the
{
problem, one vessel (Hatch 1) which had been shipped to the facility L
site was found to have unacceptable defects.
Extensive repairs were
[
i made to the Hatch vessel at the field site as well as to others still i
in the shop.
4.
Improper heat treatment of materials.
j Heat treatment is frequently necessary to permit fabrication or to assure that material properties in the final finished product are satisfactory.
Since heat treating is a special controlled process frequently tailored to the material or to the form being fabricated, adequate control is not always exercised or obtained.
Many instances
~
of improper heat treatment have been identified by NRC inspectors during vendor inspections.
3-5.
Failure ta meet nondestructive test acceptance standards.
A number of cases have been identified in the review of radio graphs of components where acceptanice standards have not been met.
These deficiencies are usually a.isociated with pump and valve castings and in welds.
i i.
l' i.
O e
t o
l I:.
55.-.5
f.
l '.'L 1
o sc
- l..'C.
.. k}.
Appendix E Manpower Reauirements for an Exoanded Vendor Inspection Program Conducted by Directly Employed Staff Given there is a need for an NRC vendor inspection program, then there is need for expanding the current LCVIP effort to provide inspection coverage of all vendors.
A possible option for expansion is the use of directly employed staff to perform all needed inspection activities.
The following information is provided to point out the projected manpower needs which would be required to implement an adequate vendor inspection program.
The principal argument for use of directly employed staff is the higher degree of control inherently obtained by the direct management of the l
inspection activities.
This enhanced control re,sults in faster dissemination of information and the ability to obtain a more positive response in the correction of deficiencies, hence greater assurance of
~
safety.
The arguments against use of directly employed staff for all vendor inspection is a further duplication of inspections (the ASME would continue their program under state laws independent of NRC action); the marginal legal base by which a direct NRC program could be conducted (new legislation would be required); and the high cost of staff as compared to use of third parties.
e
. 1 To carry out an adequate inspection program using directly employed NRC inspectors covering all types of vendors would require a total of 88 inspectors.
This represents an' increase of 69 over the present LCVIP staff but does not include the additional support, administrative and supervisory positions needed to maintain the expanded inspection staff.
I If The number also assumes no growth in the number of organizations who may be supplying the nuclear industry after 1980.
The increased staff size is obtained in the following manner:
F Nuclear Steam System Suppliers (NSSS)
I l-Number of companies 5
Number of inspectors needed SE
(
Architect Engineers Number of companies (active) 12 U
Number of inspectors needed 13
[
Fuel itbricators (and sub-suppliers)
Number of companies 9
Number of inspectors needed 2
~
t 1/ Because'of the importance of the NSSS to a nuclear project, one' man-year inspection effort is needed for each firm.
2] Four of the AE's have 15 or more projects, others have a single project thus the inspection need varies from one quarter man-year per company to three man-years p_er company..
f p.
=:.
I
. ASME Vendors Number of companies 241 Class I
~
108 Class II, III Number of inspectors needed E
For Class I vendor QA program audits 18 0
For Class I vendor routine inspection 32 For Class II, III vendor-program review E
and routine inspection 7
E Total 57 3] Each Class I vendor requires 3 man team, QA program audit every third year.
4_/ Each Class I vendor requires a minimum of two inspections per year.
,5) Each Class II, III vendor requires a minimum of one inspection per year.
ll i
jf This expanded NRC inspection program would not be performed precisely in the same manner as the current ASME effort in that the NRC inspectors would not be assigned to a resident status at manufacturing facilities.
The effectiveness of the projected NRC effort is considered to provide equivalent quality assurance.
)
, Non-ASME Vendors Number of companies 500-600 Estimated
~
Number of inspectors needed E
For ele:trical equipment 4
For instrumentation N
I E
For other miscellaneous equipment 4
F Total 11 Total number of inspectors required:
Nuclear Steam System Suppliers 5
Architect Engineers 13 Fuel fabricators 2
ASME vendors 57 Non-ASME vendors 11M If the cost avoidance to licensees is determined in the same manner for p:
inspections conducted by a directly employed NRC staff as that determined utilizing third parties, no net saving can be shown.
l
_7f Estimate that up to 60 electrical equipment manufacturers would require one inspection per year.
8_/ Estimate that approximately 50 instrumentation manufacturers would require one inspecticn per year.
9] Estimate that approximately 60 inspections per year of manufacturers of miscellaneous equipment would be required.
These inspections would not be of the same 60 firms in any given year.
~
l
. A cost saving to licensees utilizing third parties was calculated as
$7,400,000.3S/ The cost of an expanded NRC staff to 88 inspectors is
$7,484,000, thus cost recovery is not a valid argument for utilizing a directly employed NRC staff.
In referring to inspection coverage of all vendors, it should be noted that the inspection of each of the approximately 1000 nuclear suppliers annually or on a fixed frequency basis is not intended.
The need for inspection is based upon the safety importance of a particular supplier's product and its performance, thus over any fixed period of time, th.e inspection of 10% or less of the total number of vendors in any'one classification may be adeouate coverage of all vendors in that grouping.
The manpower estimates above are based on that premise and explains why 57 NRC inspectors can provide the equivalent inspection coverage provided by the approximate 300 ASME inspectors.
10/ See Section VII of the main report.
=
a Vendor procram Mancower Reouirements (Issue 2)
(Alternate 1)E (Alternate 2)U (Alternate 3)
(Alternate 4)3_/
Current Total In-House Current LCVIP Expanded LCVIP LCVIP LCVIP With ASME With ASME (Trial Program)
NSS 3
5 3
5 AE 4
13 4
13 Fuel Fab 2
2 ASME 8
57 3
3 Non ASME 4
11 4
11 i
Audit ASME 5
5 TOTAL
- 19 88 19 39
-1/ Assumes LCVIP would inspect all vendors as in expanded program but at lower frequency and with no increase in staff.
y Assur.4es scope of program would be expanded to provide adequate inspection in all product and service areas utilizing directly employed staff.
y Assumes successful use of ASME as third party for inspection of ASME vendors.
Non-ASME vendors and contractors would be covered by modest increase in directly
)
i employed staff.
.. - ~ ~
~
l Enclosure A PRINCIPLES TO BE SET FORTH IN AN EXCHANGE OF CORRESPONDEFCE BETWEEN THE UNITED STATES NUCLEAR REGULATORY COMMISSION, THE AMERICAN SOCIETY OF MECHANICAL ENGINEERS, THE NATIONAL BOARD OF BOILER AND PRESSURE VESSEL INSPECTORS REGARDING THE CERTIFICATION AND INSPECTION OF NUCLEAR SUPPLIER QUALITY ASSURANCE PROGRAMS 1.
Purpose The purpose.in an exc'hange o,f correspondence is to define the
~
~
actions which will be taken and the responsibilities which will be assumed by the American Society of Mechanical Engineers (ASME), the National Board of Boiler and Pressure Vessel Inspectors (National Board) and the Nuclear Regulatory Commission (NRC) relating to the ASME Certification program and third party inspection of suppliers providing products or services to nuclear facilities in accordance with the ASME Boiler and Pressure Vessel Cede, Sections III (Division 1 and 2) and XI. The objective of this effort is to provide NRC Licensees and license applicants with a procedure for accepting the ASME certification of quality assurance programs of nuclear suppliers and the monitoring of supplier quality activities to assure compliance with NRC quality requirements.
===2.
Background===
The Atomic Energy Act of 1954 and the Energy Reorganization Act of 1974 authorize the NRC to license and regulate the manufacture, construction, and operation of atomic energy production and utilization facilities from the standpoint of the comon defense and security and a..
, the public health and safety.
In exercising its authority, the NRC has incorporated the ASME Boiler and Pressure Vessel Code into its regulations except for recog'nition of the ASME certification prograrrr and stamping.
Concurrently, several of the states and political subdivisions (defined in the constitution of The National Board as any city of the United States having a population of 1,000,000 or more) have passed laws requiring that the overall construction and
[
inservice inspection of certain components and systems, e.g. pressure vessels, piping, valves, pumps, etc., be performed in accordance
~
with the rules of the ASME Boiler and Pressure Vessel CodesSection III, that operational and inservice inspection be 'Jerformed in accordance with Section XI.
Construction, as used herein, is an all inclusive tenn comprising materials, design, fabrication, examination, testing, inspection, and certification required in the manufacture and installation of items.
The apparent overlapping of similar functions has resulted in the evolution of separate and independent inspection systems applicable to a common group of nuclear equipment suppliers and vendors.
The consequence of this overlapping has been duplication of effort in p
the areas of quality assurance program evaluation and inspection.
In the interest of efficiency and greater effectiveness in assuring the ultimate safety of nuclear facilities, the NRC, the ASME, and the National Board are engaged in a mutual effort to establish quality assurance requirements and applicable administrative procedures of each organization which are compatible.
3-The exchange of correspondence is intended to serve as an initiating vehicle for a two year trial program and to document administrative arrangements for proceeding 'with that program.
The objective of this trial program is to provide for the establishment of suitable bases and verifications to permit NRC recognition and utilization of the ASME system of certification and inspection of nuclear suppliers.
Successful completion of the trial program is expected to culminate in NRC recognition of the ASME certification and inspection system as an acceptable procedure which may be used by NRC licensees and license applicants to satisfy NRC requirements' relating to the evaluation and inspection of supplier quality assurance programs.
3.
General Understandings 1.
The trial program is to be conducted for a two year period beginning on the date of your acknowledgement of this letter.
2.
This program may be unilaterally terminated by any of the three parties upon submission of written notice giving reasons for termination to the remaining parties.
3.
To the extent practicable, each party will consult fully with the others with respect to any action concerning implementation of the program.
Copies of all correspondence and other documents
. J. '
..L-..
y t
i
. relating to enforcement or corrective action of identified H
deficiencies will be furnished to the other parties on a timely basis.
4.
The participating parties will implement the understandings set forth in this. correspondence so as to avoid duplication of effort to the maximum extent consistent with their professional and I
statutory obligations and the health and safety of the public.
To 'iis end, each party may use advice and information provided by the other parties for the purpose of carrying out its responsibilities.
5.
In order to coordinate surveys, inspections, audits and reviews of activities and efficiently implement requirements, each agency will regularly advise the others of its schedules for accomplishing assigned functions.
4.
Agency Responsibilities and Agreements a.
The American Society of Mechanical Engineers agrees to the following:
1.
To continue its leadership and coordinating role in promulgating, issuing and improving the nuclear codes and standards under its jurisdiction.
The ASME will improve and broaden the quality assurance requirements and administrative procedures of the ASME Boiler and Pressure Vessel Code, Sections III and XI and any others which may apply, so that compatibility with the NRC's rules and regulations is achieved.
Specific LL
=
' areas of the ASME code requiring modification and/or implementation include:
(a)
Expansion of scope to include requirements applicable to the total product including operability / function.
(b)
Expansion of ' requirements to achieve complete compatibility with Appendix B and the regulatory guides endorsing the N45.2 series of standards.
(c)
Expansion and documentation of procedures regarding the survey, inspection and enforcement process.
(d)
Upgrading of authorized inspectors to ANSI N626 standards.
(e)
Upgrading of QA requirements relating to Class III components (safety-related).
i l
l f
l
g T
6-2.
To continue its leadership, coordinating and administrative role in implementing the nuclear certification program of the Society's codes and' standards.
Specifically with respect tr e certification of nuclear suppliers providing products and services in accordance with Section III, the ASME will continue to manage the survey teams, including the selection of qualified team members; review survey team reports, issue or defer issue of Certificates of Authorization based upon information provided by Society representatives or agents; provide opportunity for a hearing on charges of Code vi'olations and, if necessary, recall the Certificate of Authorization involved; and maintain records of such transactions.
3.
To permit designated NRC representatives to inspect and audit the activities administered by the Society relating to the issuance of Certificates of Authorization to nuclear product or service suppliers including all survey team activities and records associated with the issuance of such nuclear certificates.
4.
To encourage and assist those organization's (National Board, Jurisdictional Authorities, Authorized Inspection Agencies) in achieving corrective actions to the NRC's satisfaction on deficient items within the scope of the Code identified by designated NRC representatives.
"::G '.'
. 5.
To prepare administrative procedures acceptable to the NRC for conducting and documenting the program of supplier certification.
Inherent in these procedures shall be accept-reject criteria for the issuance of Certificates of Authorization.
6.
To make appropriate documentation relating to supplier certification activities available for public disclosure.
b.
The National Board of Boiler and Pressure Vessel Inspectors agrees to the following:
1.
To continue its enforcement, and investigative activities i.
associated with implementation of the ASME nuclear codes and standards.
2.
To assist authorized inspection personnel in meeting or exceeding the requirements of the appropriate Amer'ican National Standards Institute (ANSI) N626 Standards by conducting suitable training courses.
3.
To permit designated NRC representatives to inspect and audit the activities performed by the National Board related to the monitoring and surveillance of third party inspection
=
as required by Sections III and XI of the ASME Boiler and Pressure Vessel Code and the ANSI N626 standards.
8-4.
To assist upon request, NRC designated representatives to arrange meetings and audits of the authorized inspection agencies including s' tate and municipal agencies acting in ASME code implementation to verify compliance with the ANSI N626 standards.
5.
To assist jurisdictional authorities and Authorized Inspectior Agencies in achieving corrective actions to the NRC's satisfaction on deficient items within the scope of the Code identified by designated NRC, ASME or National Board representatives.
t c.
The Nuclear Regulatory Commission agrees to the following:
l 1.
To encourage industry support of the ASME and Nation Board activities associated with certification of nuclear Code-stamped items under the rules of the ASME Boiler and Pressure Vessel Code Section III and with the inservice inspection of nuclear power plant components under Section XI.
i 2.
To inform the public including NRC licensees and applicants of the objectives of the trial program.
3.
To review and formally endorse as appropriate the revised Code and the ASME inspection system procedures as achieving compatibility with NRC's requirements and practices.
-l
---,-~-7
"*T:':
- 1*M:
I t
t W 4.
To audit the ASME and National Board activities associated with nuclear certification and third party inspection on a planned scheduled ba' sis, report the findings, and cooperate-in obtaining corrective actions on areas found deficient.
The NRC audits are to be conducted utilizing the revised code as appropriate to verify that the activities are being implemented in an effective manner.
(Until the revised Code is available, the base used for NRC audits will be NRC regulations and regulatory guides).
3.
To provide assistance, advice and guidar.ce to the ASME and the National Board regarding implementation, changes in administrative procedures or in code rules for activities included in this agreement to assurr: compatibility with NRC requirements.
t Director Office of Inspection and Enforcement USNRC i
--