ML20212E540
| ML20212E540 | |
| Person / Time | |
|---|---|
| Issue date: | 04/15/1999 |
| From: | Wessman R NRC (Affiliation Not Assigned) |
| To: | Travers W NRC OFFICE OF THE EXECUTIVE DIRECTOR FOR OPERATIONS (EDO) |
| Shared Package | |
| ML20212E401 | List: |
| References | |
| NUDOCS 9909270034 | |
| Download: ML20212E540 (24) | |
Text
fe o l <
[ UNITED STATES NUCLEAR REGULATORY COMMISSION 8' WASHINGTON, D.C. 20066-0001 l'
- April 15,1999-MEMORANDUM TO: William D. Travers l Executive Director for Operations FROM: Ichard H.Wessman, Deputy Director Division of Engineering ,
Office of Nuclear Reactor Regulation
SUBJECT:
DPO ON NDE OF SPENT FUEL STORAGE CANISTERS In a memorandum dated November 30,1998, Mr. Ron Parkhill of the Spent Fuel Project Office (SFPO), NMSS, expressed his differing professional opinion (DPO) regarding the non-destructive examination of welds in spent fuel storage canisters. Your memorandum of January 5,1999, appointed Mike Modes and myself (chairman) to cn ad hoc panel to review Mr.
Parkhill's DPO. Dr. Ed Hackett (RES) was appointed as the third member of the panel; his selection was from the suggested panel members identified by Mr. Parkhill.
The DPO Panel reviewed documents transmitted by Mr. Parkhill in this November 30,1998 memorandum and various documents utilized by the DPV Panel. Mr. Parkhill also provided a number of documents as part of his direct interaction with the DPO Panel. Other relevant documents included industry correspondence relevant to the fuel storage cask issue, meeting summaries, standard review plans and interim staff guidance developed by the SFPO, and various relevant ASME consensus documents (including proposed code cases). The DPO panel held discussions with Mr. Parkhill, representatives of Nuclear Energy Institute (NEI),
representatives from the Electric Power Research Institute (EPRI), representatives of the Offices of Nuclear Material Safety and Safeguards, Nuclear Reactor Regulation, and Nuclear i Regulatory Research, and selected representatives from industry and ASME.
The DPO Panel has completed its review and its report is attached. ,
I s
Attachment:
As stated i
l 9909270034 990719 PDR ORG NE TOP PDR f?0?$)Uf f j
Differing Professional Opinion Panel Report Concerning the November 30,1998 1 Differing Professional Opinion Submitted by Ron Parkhill l
T ).l . % 4/tS[99 Richard H. Wessman, Chairman Date ~l I Ef I ;
Dr. Ed Hackett, DPO Pa'ne(Member '
Date' l
$ ff Michael C. Modes. DPO Panel Member Date l
i 1
1 l
i
i -
DIFFERING PROFESSIONAL OPINION PANEL REPORT l
. CONCERNING THE NOVEMBER 30,1998, DIFFERING PROFESSIONAL OPINION SUBMITTED BY RON PARKHILL L Introduction 4
This report discusses the review of the differing professional opinion (DPO) dated November 30, 1998, submitted by Ron Parkhill to the executive director for operations (EDO). The DPO
- (attachment 1) concerns certain aspects of the nondestructive examination and regulatory requirements regarding the examination of welds in austenitic stainless steel containers for storing spent fuel, Mr. Parkhill, who is on the staffof the Spent Fuel Project Office (SFPO) in the Office of the Nuclear Material Safety and Safeguards (NMSS), has technical review responsibilities for matters concerning fuel storage containers. Following receipt of the DPO, a panel was asked to review the DPO in accordance with NRC Management Directive 10.159,
" Differing Professional Views or Opinions." The panel members were Richard H. Wessman (Chainnan), Dr. Ed Hackett, and Michael C. Modes.
IL Background (Portions extracted from Mr. Parkhill's -morandum) j l
Dual-purpose dry cask storage systems (DCSSs) are designed and fabricated to be in service for j many years until a permanent repository for spent fuel becomes available. The NRC seeks a )
high level of assurance that DCSSs will maintain confinement, suberiticality, radiation shielding, and retrievability of the fuel under all credible loads for nonnal and off-normal accident conditions and natural phenomena. Associated regulatory requirements are found in 10 CFR Part
- 72. In recent years a number of cask designs have been developed; however, no single structural code covers the design of all spent fuel storage systems. As stated in NUREG-1536," Standard Review Plan for Dry Cask Storage Systems"(SRP) the acceptability of any given structure will be contingent upon a combination of adherence to applicable portions of multiple codes and a review of the functional performance of the stmeture taken as a whole. This approach allows the designer to request relief and the reviewer to impose additional restrictions when warranted by specific design features.
In NUREG-1536, NRC staff states that the structural design, fabrication, and testing of the confinement system should comply "with an acceptable code or standard, such asSection III of the Boiler and Pressure Vessel Code (B&PV) promulgated by the American Society of Mechanical Engineers (ASME). (The NRC has accepted use of either subsection NB or subsection NC of this code.) Other design codes or standards may be acceptable depending on their application." The NRC cited this code because no specific code exists for spent fuel storage canisters and because this code was principally utilized for nuclear power plant components. i Applicants for a fuel canister license have, typically, identified areas in which they do not comply with code requirements, such as closure weld configuration, no hydro test, no volumetric examination of the closure weld, and no use of an authorized nuclear inspector.
p4 .,
O -
2 .
W It should be noted that the regulations (10 CFR Part 72) that govem the storage of spent nuclear fuel are largely performance based. This is reaffirmed in the introduction to NUREG-1536.
Over the past several years, instances of cracking were identified at three reactor sites during the
. welding of the closure lids of a Ventilated Storage Cask Model No. 24 (VSC-24) carbon steel canister. The cracking was discovered during visual and surface nondestructive examination L (NDE) and by helium leak testing.- Corrective actions included improvements to the welding process and ultrasonic testing (UT) of the outer carbon steel closure weld.
During 1996-1998, the, staff and industry engaged in extensive activities regarding generic issues relating to the transport and storage of spent nuclear fuel. There has been a strong need to deal with these issues as many licensees are approaching situations in which existing spent fuel storage capacity is limited. Several cask designs are currently under review by the SFPO; ASME code groups are considering consensus requirements for fuel storage casks: NEI and the industry have developed approaches to dealing with cask issues; and the staff continues its interactions with the stakeholders. A significant milestone was reached in January 1997, when NUREG-1536 _was published. More recently, the staff and industry addressed cask issues in several widely attended meetings. Two significant meetings were workshops held July 1-2,1998, anu August 26,1998. A more recent workshop was held in Washington, D.C. on March 2-3,1999.
A principal topic discussed at the workshops was the UT of closure welds on austenitic stainless steel cask closures and the inclusion of the ASME Code requirements in the licensing of casks for storing and transporting spent fuel. Attributes of austenitic stainless steel casks were considered and some of the difficulties in the various NDE methods were discussed. The industry emphasized the ductile and flaw-tolerant properties of stainless steel casks and their closure welds. The industry recommended the following:
(1) Adopt the ASME Code Section III, Division 1, as a materials, design, fabrication, enmination, and testings standard.
' (2) Recognize that the cask vendor and utility owner are ultimately responsible for ensuring regulatory compliance.
. (3) Defer implementation of ASME Code third-party inspection procedures and Code stamping ofitems covered by ASME Code rules.
(4) Commit to further development of consensus codes and standards.
At the August 26,1998 workshop, the staff reminded the industry that the SRP explicitly suggests UT as the alternate method to radiography and suggests that the industry develop consensus guidance regarding alternative technical methods and propose alternatives to the staff.
Historically, it has often taken several years to develop consensus codes and standards. The
. SFPO faced near-term licensing decisions for several stainless steel cask designs and had s
y..
3 J
, identified several areas in which it was felt that NUREG -1536 (the SRP) did not contain the latest guidance. On October 8,1998, the director of SFPO issued several interim staff guidance (ISG) documents to address this situation. One of the ISG documents was ISG-4, " Cask Closure Weld Inspections," dated October 2,1998. ISG-4 recognized that radiographic testing (RT) is preferred for inspecting cask closure welds. The ISG also recognized that such an inspection was not practical. UT was recognized as the next preferred method, but the ISG acknowledged that )
UT of stainless steel welds for certain closure configurations may pose considerable difficulty.
ISG-4 recognized that dye penetrant testing (PT) only identifies surface flaws and stated the staff
. position that, if FT tests were performed at sufficiently small weld depths, they can provide reasonable assurance of weld integrity. In sum, the ISG-4 position was that the closure weld for
' the outer cover plate for austenitic 'stainless steel designs may be inspected using either volumetric or multiple-pass dye-penetrant techniques, subject to certain conditions as noted in the ISG. ISG-4 also stated that NRC regulates to the standard of adequate protection, not absolute assurance,' concluding that PT (if conducted in accordance with the ISG guidelines) provides reasonable assurance that flaws ofinterest will be identified.
During the development of the SFPO positions that were published in the ISG documents, Mr.
' Parkhill concluded that his concems regarding the NDE method were not being appropriately addressed. On September 8,1998, he submitted a formal diff: ring professional view (DPV) regarding the technique for examining welds on closures to spent fuel storage canisters (attachment 2). In summary, Mr. Parkhill stated his belief that " volumetric examination should be performed on the confinement closure welds since all other confinement welds are volumetrically inspected, it is required by the governing code, and would verify that there is not a welding process problem (i.e., single failure)."
A DPV panel, comprising Thomas Martin (Chairman), Geoffrey Homseth (NRR), and Deborah Jackson (RES), reviewed Mr. Parkhill's DPV. The panel sent its report to the director of NMSS on November 10,1998 (attachment 3). The DPV panel concluded that " employing PT of the weld periodically after a specified deposit depth will provide adequate assurance that no weld flaws greabr in depth than the specified PT interval (depth) would be plausible." In its conclusion, the DPV panel emphasized the importance of determining critical flaw size. The j DPV panel also recommended that (1) an appropriate code specific to the construction and use of these casks be endorsed and applied and (2) if PT is used to verify closure weld integrity, priority should be given to providing NRC inspection oversight of this process. ;
III. DP_O Summary I
Mr. Parkldll did not agree with the conclusions of the DPV panel. His DPO, consisting of two issues, was transmitted on November 30,1998 (attachment 1). The issues in Mr. Parkhill's DPO are summarized below:
' The first DPO issue concems the type of nondestructive examination (NDE) to be used on the closure weld for dual-purpose spent fuel storage canisters made of austenitic stainicas steel.
l 4
, . Specifically, Mr. Parkhill believes a volumetric examination, as opposed to a surface
' examination technique, is necessary to identify welding process problems that may othenvise go undetected.
He second DPO issue concems the goveming code requirements that are not being complied with.' NRC is allowing the industry to apply certain exceptions to the code in the construction
- (specifically for the closure welds) of the austenitic stainless steel dual-purpose spent fuel storage canisters. As described in Mr. Parkhill's DPO, the NRC has determined that ASME Code
~
Section III, Class 1 or 2 requirements, with certain exceptions, are acceptable for all the storage canisters licensed to date. The following specific exceptions are of concem to Mr. Parkhill:
closure weld configuration is partial penetration, no hydro test, no volumetric examination of the closure weld, and no use of an authorized nuclear inspector (ANI). Since these exceptions are allowed, no certification (N-stamp) is performed.
In his DPO, Mr. Parkhill also commented on the DPV panel report of November 10,1998.
These comments are summarized as follows:
(1) Evaluation of other volumetric examination methods, such as radiographic test, was not addressed.
(2) The DPV panel recommendation that an appropriate code specific to the construction and use of dry cask storage casks be developed is impractical for near-term licensing. The
- solution is to apply the requirements of the goveming code (ASME Section III, Subsections
' NB or NC).
(3) The DPV panel inappropriately suggests that surface examination may be acceptable for canister shop welds, if supported by a fracture mechanics analysis. (Shop welds currently receive a volumetric examination.) In the interest ofpublic confidence, industry should conform to the requirements of the governing code.
(4) In recognition ofpotential difficulties in performing surface examinations on closure welds, the DPV panel recommends NRC provide inspection oversight. Mr. Parkhill correctly
. points out that "NRC is not... an in-process quality control organization..." He suggests the appropriate solution is to comply with goveming code requirements and utilize ANIS.
(5) The DPV panel's report states that the goveming code is not mentioned in the Part 72 regulations and that this gives NRC leeway to differ from the code's requirements. NRC required volumetric examination'on the VSC-24 (carbon steel) cask as a result of the VSC-24 welding problems; the lesson leamed is to follow the consensus standard to the maximum extent possible and require volumetric examination.
-(6). The DPV panel states that it was unable to detennine if UT could be effectively used for storage casks. Mr. Parkhill believes that experts have agreed that UT can be done for the closure weld and that the cask joint configuration should be redesigned to more appropriately support a UT examination. L
O 5
, The DPO panel considered these comments as part ofits review and its views on these comments are presented in section G of the DPO summary ofissues reviewed.
IV. Summary ofIssues Reviewed by DPO Panel The DPO panel reviewed documents transmitted by Mr. Parkhill in the November 30,1998, memorandum and various documents utilized by the DPV panel. Mr. Parkhill also submitted a number of documents as part ofhis direct interaction with the DPO panel. Other relevant documents were industry conespondence relevant to the fuel storage cask issue, meeting summaries, standard review plans and interim staff guidance developed by the SFFO, and various relevant ASME consensus documents (including proposed code cases). The DPO panel held discussions with Mr. Parkhill, representatives of Nuclear Energy Institute (NEI),
representatives from the Electric Power Research Institute (EPRI), representatives of the offices ofNuclear Material Safety and Safeguards, Nuclear Reactor Regulation, and Nuclear Regulatory Research, and selected representatives from industry and ASME. Members of the DPO panel attended portions of the NUPACK subgroup meeting during the ASME meetings in Birmingham, Alabama, on February 23,1999, as this group is developing a code case dealing with cask NDE approaches.
Key issues considered are summarized in the following paragraphs.
A. Safety Significance The DPO panel considered the safety significance of the issues raised by Mr. Parkhill in the overall context ofmanagement of storage of spent fuel in austenitic stainless steel casks. Mr.
Parkhill states (page 3 of the DPO) that he does "not feel that surface examination of the canister closure weld is a safety issue as long as the materials of construction are ductile stainless steels, which can generally tolerate relatively large flaws without causing brittle fracture." The DPO panel also explored perceptions of safety significance with other staff and industry representatives during its various interviews. No one expressed a concem that surface examination practices, if appropriately controlled, represented a safety significant issue.
Individuals familiar with ISG-4 agreed that the allowed approach to NDE did not represent a safety issue, even if they personally felt that UT would be preferable. For reasons discussed further in the following paragraphs, the DPO panel agrees with Mr. Parkhill's statement and with the other staff and industry representatives regarding safety significance.
The DPO panel also sought to consider, from a general standpoint, the extent of any evaluations of risk perspective of stainless steel casks with closure welds that received only a surface examination. Although there have been extensive studies of various accidents and failure modes of cask storage systems, there has been relatively little work involving probabilistic risk assessment (PRA). The SFPO is currently preparing an integrated safety analysis (ISA) of the VSC-24 cask; this study is a systematic examination of a facility's processes, equipment structures, and personnel activities to ensure that all relevant hazards that could result in unacceptable consequences have been adequately evaluated and that appropriate protective
p 7 p '<
6'
~
measures have been identified. This study includes consideration of shield-lid weld failures and b
. the potential accident scenario of the accumulation of precipitation in the canisters over a period
. of yeam, possibly leading to criticality or hydmgen generation. The study is incomplete. It !
suggests, for the postulated conditions for a VSC-24 cask, further analysis of a criticality
- ' accident is not recommended and further analysis of the hydrogen generation scenario appears l warranted. ,
r No such analysis for a stainless steel cask has been developed.
B. Materials and WeldingIssues (1) Characteristics ofAustenitic Stainless Steels Characteristics of austenitic stainless steels that have primary relevance to use in dry cask storage systems are (a) superior ductility, toughness, and fracture toughness when compared with carbon and low-alloy steels in the same strength range; (b) excellent resistance to general corrosion, and; (c) an atomic structure that allows rapid diffusion / dissipation of embrittling species such as hydrogen. Characteristics a and c are particularly relevant to the DPO issue.
Since the stainless steels and their weldments generally have superior fracture toughness, l structures fabricated from them can typically tolerate very large flaws before structural integrity becomes questionable.- This generous flaw tolerance facilitates nondestructive examinations (NDEs) of all types, since the target flaws are reasonably large and hence easier to detect / size.
With regard to characteristic c, the atomic structure of austenitic stainless steels (being conducive to rapid diffusion of small atoms) is essentially in.mune to delayed cracking in weldments due to hydrogen. Delayed cracking due to hydrogen was an issue with the weld cracking in the VSC-24 !
carbon steel system and was a key reason for the staff requirement of a volumetric examination for the structural closure weld in the VSC-24.
(2) Welding of Stainless Steels The report of the DPV panel (November 10,1998) contained an excellent discussion of austenitic ' stainless steel weld flaws. The DPV panel concluded the following:
(a)' The most likely linear-type flaws in austenitic suinless steel welds are hot cracks which are predominantly surface connected flaws.
i (b) - Of the types ofprocess defects that may generate subsurface flaws, the defects would be l bounded in depth by the thickness of a single weld pass.
(c) The likelihood ofservice induced propagation of any potential flaws is negligible.
The DPO panel concurs with the conclusions of the DPV panel. However, the DPV panel did not discuss the potential for underbead cracking. The majority of austenitic stainless steels are susceptible to underbead cracking, especially in sections more than 3/4-inch thick with high restraint weld joint geometries. Underbead cracking occurs in the heat-affected zone (HAZ) of 1
fi , ~.
~. 7.
[, ' the base metal adjacent to the weld and, by definition, does not generally extend to the surface.
Therefore, underbead cracks would be missed in PT exams. However, for underbead cracking to be significant from a structural integrity perspective, the underbead cracks would have to occur p .in several successive passes and link up to form a continuous flaw. If adequate process controls l are maintained during welding the DPO panel believes such an occurrence would be highly ;
unlikely.
- C. NDE Alternatives for Stainless SteelCasks Radiographic evaluation of this weld configuration would be difficult because the source and recording media have to be in a line of sight (see Figure 1). In order to capture meaningful data, the recording medium should be as close as possible to the object ofinterest, in this case the seal
- weld, in order to avoid degrading the captured information by geometric and radiation scattering.
i
{
l Recordng Mede
\*
N
, =.-
Figure 1. i 1
This constraint requires the source to be placed at a distance from the weld, opposite the recording medium, causing the radiation to pass through at least 6 inches of material. ,
This requires very high levels of radiation in a small source size to resolve the weld defects.
Because the radiation passes at an angle to the probable defect orientations for cracks this method would not lend itself to good resolution. Generally the source of radiation should be aligned along the longest plane of the crack in order to best resolve the defect.
These problems are not insunnountable. The solution is however very expensive. The source of radiation would have to be a linear accelerator or microwave pumped x-ray source in order to obtain high frequency resolution. This source, along with a digitizer, has been packaged for use
- by SKB (Svensk Karnbranslehantering AB), the Swedish Nuclear Fuel and Waste Management Company, for radiographing the electron beam welded copper vessel closure of the Swedish i
,4, g 8
longi term radiographic waste burial container. The instrument is large, cannot be transported into the field, and emits large amounts of radiation while performing radiography. This is an expensive solution that is justified for long-term storage; an interim storage cask seal weld,-
however, hardly merits such an expensive solution. More information is available at:
L http://www. bio-imaging.com/pages/new4.htm Ultrasonic testing is not that difficult to attain (see Figure 2.). The ultrasonic beam can be directed fmm the outside of the vesselwall toward the weld. The applicable techniques are well
.known and well understood. They are not, however, routine techniques and they take skill in implementing. -
nib '
4 .
- =r:.~.
Figure 2.-
Ultrasonic testing using tip diffraction with a longitudinal 0 and 70 wave would fully interrogate the weld volume and, using tip diffraction, would cover all the possible crack orientations. This technique would avoid unnecessary radiation exposure to the ultrasonic technician, who would be to one side of the container. The test would be repeatable and the defect information can be readily recorded from the outputs of a standard portable ultrasonic
~
machine. The nuclear industry currently has a large number of qualified ultrasonic technicians who can, and regularly do, perform this test when examining pipe for intergranular stress p corrosion cracking.
j This solution is far less expensive than radiography and can be implemented in the field by hand-l carried equipment. It would require a day's work by a skilled technician, and perhaps consultation with a qualified Level III examiner if anomalous indications were found. If the
-volume of work warranted the additional expense, the process could be automated using adaptation of standard ultrasonic scanning equipment. This would increase the quality of the examination, lend itself to digitized data that would be available for future reference, and further reduce the radiation exposure to the operator. Using a computerized system, the operator would not need a high level of skill at interpreting the data because the data could be transmitted to a i
s ni
9
~
. central location for review. This technique would have a higher probability ofidentifying cracks than radiography and, unlike liquid penetrant, would get information about defects in the heat-affected-zone such as a crack that normally doesn't break the surface. Ultrasonic testing has been thoroughly performance tested anii its capabilities are well understood and documented.
. The least expensive method of evaluation is liquid penetrant testing. This method makes use of a cleaner, colored penetrating oil, and a line chalk powder (developer). It takes little training, is readilyavailable at many locations, and does not require extensive interpretation. The crack, if present, must be open to the surface for the oil to penetrate. This method cannot detect
- subsurface cracks or cracks in the heat-affected-zone of the weld unless they are open to the
- surfac~e . The surface of the weld cannot be buffed or peened, and scaling guns are prohibited because they all smear a thin layer of metal over the opening of fine cracks.
Liquid penetrant testing has never been performance tested because of the simplicity of the method and because the only ASME requirements are found in Section V, Article 6. The method is entirely dependent on the technician because there is no reliable method for recording the indications the technician finds. The technician would have to directly face the top of the weld five times to cleari, apply penetrant, clean, apply developer, and to evaluate. This could expose
{
j the technician to increased radiation, depending on the canister configuration.
l The three methods can be ranked by their relative ability to find all the possible defects present in I the weld. Liquid penetrant, despite its inability to find subsurface flaws, is a good method.
' Radiography is better, but may miss indications that are not oriented along the long axis of the radiation beam. Ultrasonic testing is best because it covers all the possible defects without the weaknesses of the other two methods. The methods can also be ranked by relative expense.
Radiography is the most expensive method by far. Ultrasonic testing is less expensive, but has to
'be done by a highly skilled technician and requires sophisticated electronic equipment. Liquid penetrant testing is the least expensive because it only needs a few cans of solution and a technician withless training.
D. Current SFPO Standard Review Guidelines As stated in section II of this paper (Background), NUREG-1536 contains the staff's guidelines
~ for reviewers performing safety reviews of applications for dry cask storage systems. These guidelines were further supplemented by the series ofISGs, including ISG-4, dealing with inspection of closure welds. Although the SFFO continues to work with industry regarding review criteria, the SFPO has not found it necessary to update NUREG-1536 and ISG-4 in the
- areas relevant to this DPO.
As summarized in the background section, RT was recognized as the preferred method for i inspecting cask closure welds, but was considered impractical. UT was recognized as the next l t preferred method, but was considered to be relatively difficult. Multiple-pass PT tests were considered acceptable, as long as they were done at sufficiently small weld depths.
i
10-E. Consensus Standard Body Activities M. Modes and E. Hackett attended a portion of the meeting of the NUPACK subgroup during l .the ASME meetings in Birmingham, Alabama on February 23,1999. The NUPACK subgroup is charged with developing a code (Section III, Division 3) to cover construction of waste transport systems. The scopeis currently being expanded to cover both waste transport and storage systems, as vendors are proposing " dual-use" casks for both purposes. During the portion of the
. meeting attended by Mr. Modes and Mr. Hackett, the subgroup also discussed code case N-595, Revision 1," Requirements for SpentFuel Storage Canisters.Section III, Division 1." This code e case is being developed independent of the overall efforts of the subgroup in order to provide nearer-term guidance for construction and inspection of spent fuel storage canisters. Revision 1 ;
of the code case now addresses optional alternative NDE requirements for the cover plate welds 1 in ferritic steel canisters. The attemative examinations described are based on ultrasonic testing {
(UT). The discuhion was focused on lack of compatibility with current N'RC staff guidance l
- (ISG-4, October,1998) and not on the issues raised in the DPO relative to austenitic stainless steels. However, in a separate discussion (see III.F.(2) below), Mr. Robert Nickell, acting chairman of subgroup NUPACK on February 23,1999, expressed his belief that the developing code guidance would ultimately require a volumetric examination for both ferritic and austenitic cover plate closure welds.
1 F. Industry Persnectives (1) Nuclear Enerev Institute The Nuclear Energy Institute (NEI) has been actively involved in the spent fuel cask issues and the recent workshops heli in 1998. NEI has sponsored an " issues task force" which comprises vendors, manufacturers, and utilities to bring industry focus to cask matters and industry coordination with NRC. On October 20,1998, NEI sent NRC a paper on industry's position on spent fuel dry cask weld cracks (October 9,1998). The DPO panel met with and spoke by telephone with representatives of NEI on January 28,1998, to gain a better understanding of NEI activities in this area and to clarify certain aspects of the industry's position paper.
NEl reaffirmed material presented to NRC as part of the 1998 workshops and the conclusions presented in the October 9,1998, position paper. In the position paper, NEI states, on behalfof industry, that the cask closure welds in stainless steel casks can tolerate large fabrication defects and that the likelihood of producing such large flaws is negligible. Unstable crack growth is resisted in stainless steel closure welds even when assuming unrealistic postulated defects and even when assuming that substantial margins against unstable flaw growth are available for more realistic weld defects. NEI stated that the postulated flaws in stainless steel closure welds could be readily detected by conventional UT procedures and by either root / final or multiple-pass PT
- methods. NEI also presented a recommended weld efficiency factor and flaw size acceptance criteria.
l L
.)
1 11 NEI representatives stated that NEI continues to work with EPRI and ASME and supports the I
- development oflong-term criteria. In the interim, and in recognition ofnear-term licensing
. activity, NEI stated its agreement with the provisions of1SG-4 regarding allowing either UT or
. PT closure veld NDE methods, with the appropriate controls. NEI stated that it is continuing to work with NRC staffand is participating in a March 2-3,1999, meeting between industry and -
the SFPO on crack issues.~ The NE! representatives did not believe there was a near-term safety -
.- issue with allowing PT examination of stainless steel cask closure welds.
NEI and the DPO panel briefly considered the concept of a performance-based concept for stainless steel cask closure welds and whether any risk-informed analysis of cask failures had
~ been done. NEI had no substantive recommendations regarding performance-based concepts for J examining cask closure welds and recognized it would be difficult to establish a process to qualify techniques and individuals for a performance demonstration process. NEI was not aware of any risk-informed analyses, other than a preliminary study being undertaken by the staff.
(2) EPRI-Interview with Robert Nickell (Consultant to EPRD Dr. Hackett discussed the DPO issues in a telephone conversation with Mr. Robert Nickell on Febmary 1,1999. Mr. Nickell is the primary author of the elastic-plastic fracture mechanics (EPFM) based flaw tolerance assessment contained in the October 9,1998, industry position paper on Dry Cask Weld Cracking. Mr. Nickell indicated that, from a safety perspective, the industry considered that, for the large critical flaw sizes in austenitic stainless steels, PT was adequate to confirm structural integrity for dry cask closure welds. However, he also indicated that the majority of those in the industry considered that UT was feasible, particularly given the larger target flaw sizes. Mr. Nickell is also a member of the ASME NUPACK subgroup, which is working on development of a code (Section III, Division 3) to cover design of waste transport systems. The scope is currently being expanded to cover both waste transport and storage systems as vendors are proposing " dual-use" casks for both purposes. In this capacity, Mr.
Nickell indicated that he thought the final version of this code would ultimately require volumetric examination of cask closure welds that would be considered category D corner welds under Division 3.
(3) NEI Drv Cask Storane Workshon. March 2-3.1999 NEI sponsored a workshop on dry cask storage workshop with the industry and NRC on March 2-3,1999. The members of the DPO panel were not able to attend these meetings.
- However, the SFPO submitted a summary ofmaterials presented to the DPO panel for review.
The DPO panel found that only the presentation on the ASME Code Section III, Division 3 was applicable to its work. This presentation was conducted by Albert Machiels ofEPRI and Robert l
. Nickell, consultant to EPRI. App'icable information, relevant to development of the code included the following:
- o. Scope of the subgroup is now modified to address waste storage systems in addition to waste transportation systems.
i
[ .- ,
12 o New rules will cover " confinement boundary" items and will be placed in a new Subsection WC ofDivision 3.
o At the Birmingham meeting (February 23,1999), the subgroup voted to " fast track" the Subsection WCrules.
Under the fast track pmcess, the first draft of the new Subsection WC rules would be available for the May 18,1999, ASME meeting in Greensboro, North Carolina. Therefore, the potential exists for code rules to be implemented much mom rapidly than previously considered.
G. DPO Panel's View on Ron Parkhill's Comments on the DPV Reoort (1) Evaluation of Other Volumetric Eramination Methods Not Addressed Mr. Parkhill expressed concem that the DPV panel did not address other volumetric examination methods. The DPO panel evahtated the strengths and weaknesses of the two primary volumetric examination methods: radiographic and ultrasonic. In addition, the DPO panel considered other methods that might be used to assess the imegrity of the weld.
l A simple approach is to visually inspect the weld. In order to implement this examination, each weld pass would have to be very thoroughly viewed under magnification to determine if any l cracking is present. This method is an acceptable altemate in the naval nuclear inspection program under NAVSEA 250-1500 when the weld accessibility precludes other methods of evaluation. This form ofinspection usually requires careful surface preparation such as solvent cleaning followed by an acid or alkaline etch to remove metal that might be hiding tightly closed j- cracking. -Direct visual inspection is not necessarily required because the inspection can be l performed using remote viewing and enhancement equipment such as video to avoid unnecessary L radiation exposure. This kind ofinspection is applied to welds that would otherwise mask indications by excessive penetrant bleedout or rough surfaces that cause ultrasonic scatter. When carefully applied, this method can be better than liquidpenetrant testing. The seal weld does not l suffer from accessibility problems and lends itself to ultrasonic testing. This undermines the l usualjustification for applying the visual method.
. Another method of examination is acoustic emission. This method has been successfully applied to real-time weld testing for simple configurations of well understood welding processes. The
! seal weld would not lend itself to this kind of examination. In addition, such tests as laser interferometry and scanning laser acoustic ultrasonic are available. These tests are very sophisticated, require extensive application development, and are costly for even simple applications.
l 1 As a result of these and other considerations, the DPO panel concluded that ultrasonic and liquid I
penetrant t'esting sliould be the only candidates considered for the evaluation of the austenitic stainless steel cask closure seal weld.
l
~
13 (2) Develonment of an Aporonriate Code ,
Mr. Parkhill stated that the DPV panel recommendation to develop an appropriate code specific to the construction and use of dry cask storage casks is impractical for near-term licensing. The solution is to apply the requirements of the goveming code (ASME Section III, Subsections NB
~
or NC).
The DPO panel agrees that the appropriate code (ASME Section III, Division 3) may not be developed in time for near-term licensing actions. However, ASME has " fast-tracked" certain aspects of the code development and, therefore, the potential exists for the new rules to be approved and implemented more rapidly than previously considered (see III.F.(3). The DPO '
panel also considered that a "goveming" code, for dry storage casks does not currently exist.
Section III of the ASME Code, which was apparently used by cask designers in an ad hoc fashion, is meant to apply to the primary pressure-retaining boundary in nuclear power plants.
Therefore, application of Section III requirements to dry cask applications would likely be conservative. However, quantification of the margins involved would be difficult at best. The recommendation of the DPO panel would be to instead rely on interim staff (ISG-4, " Cask Closure Weld Inspections". October 2,1998) and industry (" Industry Spent Fuel Dry Cask Weld Crack Position Paper", October 1998) until final guidance is available from the appropriate consensus standards body (ASME).
(3) Surface Examinations for Canister Shoo Welds Mr. Parkhill believes that the DPV panel inappropriately suggested that surface examination may ,
be acceptable for canister shop welds, if supported by fracture mechanics analysis. The DPV !
coriclusions in the panel report refer to the inconsistency of requiring volumetric NDE for cask
)
construction and then permitting surface examination for closure welds and states: "The panel believes that it would be appropriate to perform a flaw tolerance evaluation of the cask to assist in determining the appropriate NDE technique,whether volumetric or surface, that could then be applied consistently to all welds." Mr. Parkhill believes that the requirements of the governing code should be followed.
The DPO panel noted that, at this time, there is not a governing code that is directly applicable.
This has led to NRC allowing the application of different NDE techniques, whh reasonable basis,
-as described in ISG-4. The DPO panel agrees with Mr. Parkhill regarding more rigorous NDE techniques for the shop welds. Consistent with ISG-4, RT is a preferred method of NDE and can
- be more easily' applied in the shop environment.
(4) NRC Should Provide Insoection Oversieht of Closure Welds t
L Mr. Parkhill takes issue with the DPV panel recommendation regarding NRC providing L -inspection oversight.
)
l J
g- .
. I 14 I- ' 'Ihe DPO panel agrees with Mr. Parkhill's view that the NRC is not an in-process quality control L organization and does not have inspection resources to confirm that surface examinations are i
-performed on every closure weld. The NRC does, however, have regional offices dedicated to
! the inspection oflicensee activities.
In addition to the resident staff, there are specialist inspectors based at the regional offices. The inspection activities are controlled by a series ofinspection procedures. The following applicable procedures cover spent fuel storage:
60852 ISFSI Component Fabrication by Outside Fabricators !
=
60853 On-site Fabrication of Components and Construction ofISFSI l
- 60854 Preep. Gonal Testing of an ISFSI -
=
- 60855 OperationofanISFSI
=
86001 ' Design, Fabrication, Testing, and Maintenance of Transportation Packagings
=-
81001 Independent Spent Fuel Storage Installations (ISFSIs)
These inspection procedures are enveloped by Inspection Manual Chapter 2690. For ISFSIs that
- are located at a reactor site, the directive states: " activities specifically related to the ISFSI (e.g.,
DCSS fabrication, support pad construction, and loading and unloading procedures) are unique and should be reviewed in depth." Inspection procedure 60853 emphasizes this view when it directs the NRC inspector to: " Verify, by direct observation and review of records, that personnel y are constructing the ISFSI and its components in accordance with the approved design
~ specifications defined in the SAR (safety analysis report), SER (safety evaluation report), C of C (certificate ofcompliance), and, if applicable, the site specific license and technical specifications."
These are highlights taken from 64 pages ofinspection procedure details that the NRC applies to the ISFSI. These inspections are individually tailored to the site in consultation with the program office and in concert with the program manager for the site. They provide sufficient guidance for the inspection staff and there is adequate flexibility to pursue any perceived weaknesses that may exist, such as the inspection of specific seal welds.
. (5) Eollow Consensus Standard and Reauire Volumetric Examination
^
Mr. Parkhill refers to discussion in the DPV panel report regarding Part 72 being silent regarding a governing code and states: " Consensus standards are meant to be followed, not fragmented." '
The DPO panel agrees with Mr. Parkhill's view that consensus standards are meant to be followed, not fragmented. The process of developing a consensus standard is one of compromise in an attempt to balance the needs of some against the limited resources of others. As a .
l consequence, the standard becomes a web ofinterrelated requirements.
n..-- .
4 15 1
l
, If a problem is encountered in developing a standard, the problem may be overcome by l implementing a series of requhements intended to retum the component to some form of balance between the conflicting needs.' For example, if a certain type ofdefect is likely to occurin a q
design, the likelihood is reduced by introducing tests or examinations specifically sensitive to the j
defect. In some cases, the fabricator is given an option and chooses not to undertake the tests.
l The fabricator then must compensate for the omission by changing some other requirement, such as rating the component for a shorter life or installing thicker sections. Stress factors or ]
~
temperature are adjusted in response to the number of tests being done on the component. For example, pipe is upgraded on the basis of a macrotest or a hydrotest;
~
i For this reason, it is very difficult to apply part of a consensus standard. It was never intended I Lthat way. In applying the ASME Code to an operational nuclear power station, the user is strictly f prohibited from using only parts of a code. This prohibited process is called cherry picking and it is forbidden. l i
This emphasizes the need to develop a cohesive, uniform, consensus standard specifically [
eddressing the problems, design, and service of a storage cask. As discussed in Sections IV E-F, l above, the DPO panel believes that ASME and the industry are acting aggressively to develop a consensus standard dealing with DCSS.
l (6) UT Can Be Done on the Closure Weld and the Joint Should Be Redesigned j Mr. Parkhill comments that the DPV panel stated that it was unable to determine if UT could be effectively used for storage casks and suggests that experts have agreed UT can be accomplished and that the caskjoint configuration should be redesigned. t
}
As stated earlier in this DPO report (Section IV.C) it appears that UT can be done on the existing design. The DPO panel believes that no redesign should be required by NRC for the purpose ofimplementing UT.
H. Performance-Based Criteria The staff has been considering broader application of performance-based criteria in its regulatory decisionmaking. SECY-98-132," Plans To Increase Performance-Based Approaches in Regulatory Activities," was issued on June 9,1998. In a recent staffrequirements memorandum (SRM), dated February 11,1999, the Commission approved the staff soliciting input from industry and other stakeholders and pursuing performance-based initiatives.
. As discussed in SECY 98-132:
A performance-based requirement relies upon measurable (or calculable) l outcomes (i.e., performance results)'to be met, but provides more I i flexibility to the licensee as to the means ofmeeting those outcomes.
l
- A performance-based regulatory approach is one that establishes ,
performance and results as the primary basis for regulatory decision-making, !
l l
E d- Q-16 and incorporates the following attributes: (1) measurable (or calculable)
_ parameters (i.e., direct measurement of the physical parameter ofinterest or of related parameters that can be used to calculate the parameter ofinterest) exist to monitor system, including licensee, performance against clearly defined, objective criteria, (2) licensees have flexibility to determine how to I meet the established performance criteria in ways that will encourage and reward i improved outcomes; and (3) a framework exists in which the failure to meet a I
performance criterion, while undesirable, will not in and ofitself constitute or
{
resnit in an immediate safety concem."
]
l SECY98-132 also states that a performance-based approach can be implemented without the use l of a risk assessment and that such an approach would require objective performance criteria l
based on deterministic safety analysis and performance history. Allowing flexibility to the j licensee is emphasized, and an advantage of a performance-based regulatory approach is that it l allows licensees to take explicit account of system characteristics that may be more efficient and )
cost effective, yet still protective of public health and safety.
The DPO panel considereed applicability of performance-based criteria to stainless steel cask closure welds and weld inspections with a focus on fracture mechanics application to stainless steels and on the use ofPT for NDE.
.The DPV panel concluded: "The critical flaw size for the component must be detemiined from fracture mechanics or other suitable analysis. The critical flaw size then determines the
- maximum depth of weld deposit that can be made before a PT is performed." The DPO panel agrees with this finding. Use of fracture mechanics in this regard is consistent with a performance-based approach and has many precedents in operating reactor evaluations. Elastic-plastic fracture mechanics (EPFM), coupled with knowledge of the limiting stresses in the component, can be used to effectively determine " critical" flaw sizes. " Critical", as applied in EPFM, also refers to onset of stable crack growth and n91 catastrophic fracture of the component, which would be the case for a linear elastic fracture mechanics (LEFM) evaluation of a ferritic component. As discussed previously (section IV.B), the superior fracture toughness of austenitic stainless steels and their weldments can allow tolerance of relatively large flaws before structural integrity becomes questionable. As an example, generic flaw acceptance calculations performed for the NEI position paper (October 1998) show the potential for tolerance of flaws ranging from 30 percent to 50 percent of the weld thickness, depending on the flaw length and loading assumptions.
Tolerance of such potentially large flaws facilitates the effectiveness of all methods of nondestructive examination since the target flaws are easier to detect / size. This generous flaw tolerance, coupled with the absence of the potential for delayed cracking in the weldment (see section IV.B), also indicates that a surface PT examination, performed within the constraints of the fracture mechanics assessment, should be adequate for providing reasonable assurance of structural integrity for the component. This is consistent with the guidance presented by the SFPO (ISG-4), and has been employed in several SFPO safety evaluations of vendor designs for austenitic stainless steel casks.
n -
17 Although PT exams would be adequate, the DPO panel also agrees with the SFPO interim Buidance (ISG-4), in that a volumetric examination (such as UT) would be preferred. The large i target flaw sizes for austenitic components greatly facilitate the use of UT for flaw detection and l sizing. UT procedures would also have the advantage of the capability for automated inspections i perfonned remotely and the potential for detecting and sizing subsurface flaws.
Developed UT perfonnance data coupled with the large flaw tolerance of the cask weld indicate that UT can very successfully locate flaws that would affect the serviceability of the weld. Even though there are no~ performance data available for liquid penetrant testing there is enough
_ experience with the' method to say that it will detect flaws of this size that are open to the surface.
The probability of a flaw of this size not being detected because it did not break the surface is not i very high; especially if the liquid penetrant test is undertaken on an intermediate weld pass level.
V. DPO Panel Findings :
1 A. DPO Issue No. l- Regardine the Tyne of NDE To Be Used '
Consistent with performance-based concepts regarding the cask and materials, the DPO panel concludes that either UT or multiple PT examinations of the closure weld for the austenitic stainless steel canister provide a reasonable level of assurance regarding cask integrity. The Inaterial characteristics of austenitic stainless steels, the welding experience, expected flaws (should they occur), and consideration of the NDE processes lead the panel to conclude that PT examinations, conducted in accordance with ISG-4 guidance, provide a reasonable level of safety.
The DPO panel further concludes that UT examination of closure welds for austenitic stainless steel canister does provide a higher level of confidence regarding the integrity of the weld and should be applied where practical. However, the DPO panel concludes that PT is adequate to ensure the integrity of the weld. The DPO panel also concludes that austenitic stainless steel DCSS surface examination practices, if appropriately controlled, do not present a safety issue.
B. DPO Issue No. 2 - Regarding Govemine Code Reauirements The DPO panel concludes that, until a specific consensus standard regarding DCSS is developed by the indastry and endorsed by NRC, NUREG-1536 and ISG-4 criteria are reasonable and acceptable. Since no single structural code covers the design, fabrication, and testing of confinement casks, it is reasonable for the staff to accept casks complying with portions of several codes, with exceptions, and with consideration of the overall performance of the cask system.
Consequently, the DPO panel finds it acceptable for the staff to allow certain exceptions to 1
provisions of ASME Code Section III requirements, particularly with respect to weld penetration, a
18.
. hydm testing, volumetric examination of the closure weld and use of an authorized nuclear inspector. The DPO panel believes Sectiori 1II requirements should be adhered to, insofar as
. practicable, during the shop fabrication phase of canister constmetion.
VL DPO Panel Recommendations On the basis ofits myiew of the DPO, and in consideration of the relevant material and discussion with industry and NRC staff, the DPO panel makes the following recommendations:
(1). The NRC staff should continue to work closely with ASME and industry to speed the development of a consensus standard specifically applicable to DCSS. (This is consistent with the recommendations of the DPV panel). Following development of an industry standard, NRC should take timely action to endorse the standard, with limitations, if necessary.
(2) The DPO panel does not recommend that NRC staff provide full inspection oversight of cask closure welding and inspection process (as suggested by the DPV panel). The DPO panel agrees with Mr. Parkhill that NRC is not an "in-process quality control organization."
The existing inspection guidance should be used to focus on the important elements of DCSS design, fabrication and closure. NRC should continue its periodic sample-based inspection approach at fabrication, facilities and at licensee facilities using DCSS.
(3) Appropriate with other priorities, the SFPO should consider a risk analysis on austenitic stainless steel casks as a followup to the integrated safety analysis work on the VSC-24 cask.
Y I
l
m .
' VW g )
Nov:mber30,1998 k MEMORANDU,M TO: Executive Director for Operations gc$)
. FROM: Ron Parkhill - I
(/
. Spent Fuel Project Office -
- f Office of Nuclear MaterialSafety l
and Safeguards .
~
SUBJECT:
Differing Professional opinion (DPO) .
The attached DPO augments my Differing Professional View forwarded to SFPO management on September 8,1998 and is being forwarded to the EDO for review in accordance with NRC Management Directive 10.159. -
A differing professional view panei docu'mented their findings in a report forwarded to Carl Paperiello on November 10,1998 and on November 24,1998, Dr. Paperiello documenteif his .
evaluation of the DPV panel's recommendations. .noth are attached
{
l Attachments:
- 1) DPO
- 2) DPV,9/8/98
- 3) Memorandum from DPV Panel forwarding Report,11/10/98
- 4) Memorandum from Carl Paperiello,11/24/98 cc W.Kane, SFPO W.Hodges, SFPO' F. Sturz, SFPO T. Martin, Res J. Thoma, EDO -
Attachment 1
- . c/,C7pVf.20/M
^
O ATTACHMENTi ,
This differhg professional opinion ( DPO) is in regard to the type of nondestructive examination (NDE) to be used on the closure weld for dual-purpose spent fuel storage canisters that are currently in the near term licensing process or recently licensed. Specifically, a volumetric examination, like ultrasonics (UT) or radiography (RT), would identify welding process problems that could go undetected by a surface examination technique, like liquid penetrant (PT), even if that surface examination was perform.ed after, some intermediate welding passes. Additionally, the governing code (as identified below) for these spent fuel storage canisters has very specific requirements for the construction of these canisters that are not being followed.
Dual purpose dry cask storage systems are designed and fabricated to be in service for many years until a permanent repository is available. In recognition of the canister's long life expectancy and to provide a secure boundary to prevent radiological releases, the NRC decided that the confinement boundary,be constructed in accordance with ASME Code Section lli Class 1 or 2. requirements, which has been the standard for all storage canisters licensed to
, date. The goveming code is identified for this application by the regulator in standard review .
plans, regulatory guides and historical licensing documents. The ASME Code Section 111 was chosen since it is the standard princip~ ally utilized for nuclear power plant components construction and since a specific code did not exist for spent fuel storage canisters. To aid the staff in identifying areas were the Code cannot be met, the applicants are requested to identify all deviations from the Code. Typically, spent fuel storage canisters have the following major areas where they do not comply with Code requirements: closure weld configuration is a partial penetration in lieu of full penetration weld, no hydro test (howev. 3 vendors are now proposing to perform one after fuel loading), no volumetric examinaut, vf closure weld, and no <
use of an authorized nuclearinspector. Since these are exceptions to the Code requirements, no' certification (i.e. stamping) is performed.
The ASME Section lli Code applies to the " construction" of nuclear components which means it 1 contains requirements for material selection, design, fabrication, examination (i.e. NDE to ensure that fabrication processes are under control and meet acceptance standards), testing
. (e.g. hydrostatic for demonstration of structural integrity), inspection (i.e. by an authorized nuclear inspector who is an independent third party with significant component fabrication experience and is trained and certified to ensure that the Code requirements are met) and certification (i.e. stamping with a Code symbol). For a component to be ASME certified, its materials are in accordance with Code requirements, the design has been configured and analyzed in accordance with Code requirements; and, fabrication, testing and examination activities have performed under the watchful eye of an authorized nuclear inspector who certifies .
that the component has met Code requirements. Utilization of this Code construction process has resulted in the excellent performance record for nuclear components.
Storage canisters have had a poor fabrication record that has resulted in the staff redirecting considerable resources to resolve those specific issues. But the NRC has failed to identify and
' impose adequate measures to prevent recurrence implementation of the Code process would
. go a long way to remedy this situation - but we have failed to take any decisive act:on.
Specifically actions we could have taken, but hkve not, include: redesign the canister closure weld to be a full penetration weld which would make it easier to volumetric examine; requiring Page 1 of 4
~
l fabricators, especially thosa with fabrication probisms, to possess a valid ASME C2rtificata cf
- Authorization; use of Authorized Nuclear inspr:: tors (Anis) during fabrication; and prohibiting slag producing welding processes such as shielded metal are welding or flux core are welding.
Examples of fabrication problems that could have been alleviated or avoided if construction had l been iri accordance with Code requirements include: closure weld cracking due to an improper l welding process, undocumented welds to base metal that resulted in cracking in area of closure weld, excessive weld grinding, and improper reading of radiographs. Therefore, since many aspects of the Code's proven process to ensure proper construction are not currently required by the regulator (i.e. the closure weld config'uration, fabricators are not Code certified, and no Code required independent third party inspection is utilized), and there have been historical t
I fabrication problems with the clos.ure weld, an appropriate compensatory measure would be to examine the closure weld in accordance' with Code requirements (i.e. volumetric' examination) to ensure that it has tieen made properly and is in agreement with the design drawings.
By memorandum dated November 10,1998 the DPV panel submitted a report to Dr. Carl Paperiello regarding my initial DPV I offer the following comments regarding that report:
a) The panel's report did not adequately address the DPV issue of volumetric examination for storage canister closure weld since it only considered ultrasonic examination, as was performed on the VSC-24. No evaluation of other volumetric methods was provided,like RT.
b) The panel's first recommendation states that an appropriate Code specific to the construction and use of dry cask storage casks be endorsed and developed. This recommendation is impractical and unrealistic for the near term licensing of dual purpose storage casks. ASME Section lli Division 3 was over 15 years in development and a storage subsection to that code could not be developed for many years-long after these immediate licensing actions have been completed. The problem isn't with needing a code specific for dry cask storage casks, but is to apply the requirements of the governing code (i.e., ASME Code Section lil, Subsections NB or NC) which has already been identified and utilized by the NRC as appropriate for this application. The issue at hand is that PT examinations have not been successfulin identifying weld process problems and the existing Code requirements provides guidance on how to reme'dy this problem-viz., do a volumetric examination.
c) Also, in the panel's first recommendation, it recognizes the inconsistency between the NDE method used for the canister welds made in the shop (i.e. volumetric) and the NDE method for the closure weld made at the site (i.e. surface) and further notes that this inconsistency is not conducive to promoting public confidence in the use of these canisters. However, the panelinappropriately suggests that surface examination may be acceptable for the canister shop welds if somehow supported by a fracture mechanics analysis, if the panel was truly concemed in public confidence, it should have recommended that the requirements of the goveming code be followed rather than alleviating the volumetric examination Code requirements with surface examination techniques that failed to identify a welding process problem with the closure weld until after 19 storage casks were loaded for the VSC-24.
l Page 2 of 4
d) in the second panel recommendation, the parci corrtetly recognizes that surface examinations performed in a high radiation area could lead to inaccurate results and could result in creating flaws if chemicals are not properly rernoved. However, the panel's solutionto this problem is.to provide for NRC inspection oversight. In reality, the NRC does not have inspection resources (especially in this r'egulatory downsizing period) to confimi that surface examinations on every closure weld, at every site, are performed conectly. The NRC is not and has never been an in-process quality control organization (e.g. SFPO inspections are routinely pe'rformed at fabrication facilities only once in a 12 to 18 month period). Again , the solution is quite simple- perform the
. correct examination required by the goveming code (i.e. volumetric). If the panel really
. wanted meaningfulinspection oversight of this closure welding that provided real technical and independent credibility, follow the goveming code requirements and utilize Anis. -
e) Paragraph Ill.B of the panel's repo'rt states that the goveming code is not mentioned in the Part 72 regulations for storage and this gives the NRC the leeway to differ from its requirements. However, many codes utilized in the nuclear industry are not mentioned in the regulations (e.g. AWS, ACl, HEl, TEMA, ASHRAE, ANSI B31.1, etc., etc.) but that does not give the regulator nor the industry an invitation to utilize just those portions they deem appropriate. Consensus standards are meant to be followed, not fragmented. For the VSC-24 welding process problems, initially the governing code's requirements were not being followed for volumetric examination and a surface examination had failed to identify flaws induced by a poor welding process unti! after 1g casks had been loaded. As part of the corrective action for the VSC-24 the NRC required that volumetric examination, UT, be utilized which has been subsequently demonstrated to be very effective in identifying welding process induced subsurface flaws that were not detected by surface examination. The lesson
- learned from the VSC-24 experience should have been to follow the consensus standard to the maximum extent possible and require volumetric examination.
f) Paragraph lil.C states that the pane! was unable to determine whether it was possible to effectively use UT for the storage casks. However, no expert to date has been quoted as saying that volumetric examination of this closure weld cannot be done. To the contrary, experts from ERPI, INEL, PNL and a NDE training consultant all. agree that UT can be done for this application. Additionally, since most of these canisters are in the licensing review process, the joint configuration could be redesigned to more appropriately support an UT examination providing we could be resolute in our decision. As a point of reference, before SFPO's interim staff guidance (ISG-4) was issued two applicants had committed to do a UT examination of these '
welds, which they later rescinded since the ISG procialmed PT as an acceptable attemative to a volumetric examination.
I do not feel that surface examination of the canister closure weld is a safety issue as long as the materials of construction are ductile stainless steels which can generally tolerate relatively large flaws without causing brittle fracture. However, the ASME Section ill Code for nuclear components does not give any relief from volumetric examination requirements just because the materialis a ductile stainless steel.
Page 3 of 4
- This issue is solely with the adequacy of fabrication of these storage canisters, where the ASME Section lli Code has very specific requirements and the NRC has chosen to ignore these i standard fabrication practices in light of numerous and recurring fabrication problems.
Also, it is inappropriate to arbitrarily reduce the safety margin inherent in the ductile material of the closure weld, by substituting a surface examination method that cannot detect subsurface l flaws that could be a result of a welding process problem. The licensed configuration of these closure welds is for good solid welds, not welds that may have subsurface flaws below each
~ surface examination layer. By using a surface examination method as a substitute for a volumetric examination, one is fabricating a joint that could be inherently weaker than the other confinement bounda'ry welds. .
Finally, the construction of these canisters is lacking fundamental Code quality assurance aspects that can contribute to fabrication problems (i.e. poor weld joint design configuration, ;
fabricators not Code certified, no independent third party authorized nuclear inspectors). Since .
these Code required quality assurance aspects are not being utilized, and there has been a i history of fabrication problems, it would seem appropriate to perform the Code required volumetric examination to ensure that the closure weld has been properly made in spite of these other shortcomings.
/?>n S H x/fy Ron Parkhill -
l l
Page 4 of 4
m -
. Frees - Ronald Parkhillj4 m3)
To MWH,FCS Date 9/8/98 4:55pm Subjects: Differing Professional View ,
Attached is a formal differing professional view (DPV) regarding the spent fuel storage canister closure steld examination technique.
CC: WFK,CJE,SFS
, e a e o
e 4
4 e
- ardxenit # 2.-
3Pafes nnmma s M Fi C/ '/ G / f
3 . .
This Differing Professional View (DPV)is being initiated because SFPO staff concems remain I unaddressed with regard to the nondestructive examination (NDE) method for the spent fuel
- storage canister closure weld. Proposed near term licensing actions by SFPO allow the use of surface examination for the subject weld whereas I feel that a volumetric examination is justified for the following reasons.
On 7/20/98 members of the SFPD licensing directorate forwarded a, position that did not require l volumetric examination of the spent fuel storage canister closure weld and notified some of the ,
applicants of that preliminary pvG. Basically, the aforementioned position relies on surface l examination, liquid penetrate (PT), for the' root pass, final pass and every 1/4 inch of weld, as weh as a reduction in the allowable stress for the weld. On 7/21/98 the SFPO Director instructed ;
that a complete written position and safety rational be developed by the technical review 4 directorate prior to advising any applicants. On 7/24/98,7/29/98, & 7/30/98 three members ;
technical review section sent e-mails supporting volumetric examination of these welds and t raised numerous concems with the proposed position. Last Friday I was informed that {
management had decided that surface examination was
- good enough" for this spent fuel j storage canister closure. 'However, no written position has been developed which justifies its l use. Consequently, the following issues remain unaddressed and form the basis as to why I believe that a volumetric examination needs to be performed on the subject weld.
- 1) Subsurface flaws will go undetected by sv face examination techniques unless the flaw penetrates the surface of the weld area being examined. it has already been demonstrated (via the'VSC-24 welding problems) that surface examinations alone are relatively ineffective in identifying subsurface cracks resulting from a poor quality weloing process. For the VSC 24 q
situation,19 casks were loaded before surface examination in combination with leak testing q
discovered that there was a problem with the welding process. Had volumetric examination been employed the weld process problem would have been discovered before many of these casks were loaded. ,
- 2) The goveming code for the confinement boundary is ASME Section Ill, Subsection NB or NC
'as stated in the current SRP and historical practice for storage canisters. As such, volumetric examination is required to be performed on 31! confinement welds. (Refer to NB/NC-5210 &
5220 for Category A & B welds, respectively.) Note that radiography is the volumetric method mentioned in these Code paragraphs, since it was intended that all parts of the reactor coolant pressure boundary would be accessible during fabrication. However, UT is an acceptable substitute where RT cannot be performed (e.g. closure weld). The propose use of surface l examination by SFPO is a deviation from the goveming Code's requirements. !
- 3) The ductility of stainless steel should not be a basis for avoiding volumetric examination of the canister closure weld since volumetdc examination is not done solely to determine flaw sizes have been bounded by a fracture mechanics analysis. Volumetric examination is done principally to ensure weld quality during fabrication. it is a verification that the weld meets i design requirements and a verification of the welding process. If the ductility of stainless steel were the only consideration for performance of an UT volumetric examination then the
' goveming code would have excluded ductile stainless steel from UT examinations, which it does D21 do. UT examinations of stainless steel components are performed routinely as part of Part 50 ISI programs. The exclusion of ductile stainless steel from volumetric examination is a deviation from the goveming Code's requirements.
r
~.g'*- .
- 4) The suggested approach of surface examination for root and final weld pass as well as, every .
1/4 inch of weld is what has been proposed in a draft ASME Code Case, currently in Revision I
- 12. However, this Code Case has not been adopted by ASME and may undergo significant l
, change prior to issuance. Furthermore, the basis for the 1/4 inch spacing between PT l examinations has not been pier reviewed and I suspect that it has not included consideration of additive spacings between pts being mora than the calculated critical crack size (e.g. If the.
closure weld is 3/4 inch, the fracture mechanics critical crack size would need to be more than 3/4 -(4 X 1/16) or % inch, which is 67% of the wall thickness). Prior to adopting a preliminary pos,ltion from the consensus Code group X may be prudent to await its final version, otherwise
, we are again in disagreement with the goveming Code. 4 l 5) Reducing the allowable stress to compensate for surface examination (i.e. PT) in lieu of a volumetric examination (i.e. UT) is not a method utilized in the goveming code for the confinement boundary (i.e. ASME Section lil, Subsection NB or NC). No amount of allowable ,
stress reduction will compensate for a poorly made weld. This is an example of mixing and !
matching different code requirements. The goveming Code requirements should be followed both by the industry, as well as, the NRC. Again, this is a deviation from the goveming Code requirements.
- 6) Historically, in the SOC for the VSC-24 rulemaking, Comment #45, the NRC response stated that the closure welds meet all ASME requirements except for volumetric examination and further stated that this inspection was'not possible due to radioactive fuelin the cask. Today we know that this documented basis for not doing the volumetric examination is not correct. We .
now know that the projected At. ARA dose is very low based on the work performed by the .
VSC-24 Owners Group and that the UT examination is viable for this application. Furthermore, doing PT every 1/4 inch tvill significantly increase dose - each PT involves approximately 200 linear inches of weld involving cleaning, application of penetrant, dwell time, removal of excess penetrant, application of developer, drying, evaluation of results, recording of results and removal of PT materials prior to continuing welding.
- 7) Regulation 10 CFR 72.236(e) requires that the cask be designed to provide redundant sealing of the confinement siystem. However, the redundant sealing requirement cannot be met if a single failure would cause both seats to fail. The single failure that would bypass the redundancy requirement would be a fall'ure of the closure welding process. The single failure I'm alluding to in this case is more than just a postulated occurrence it has happened for the VSC-24. Therefore, to prevent other single fal!ures in the welding process volu' metric examination should be utilized for examination of the closure weld, just as it is utilized for all other welds in the confinement boundary.
In summary, I believe that volumetric examination should be performed on the confinement closure welds since all other confinement welds are volumetrically inspected, it is required by l the goveming code and would verify that there is not a welding process problem (i.e. single failure). It is very hard for me to appreciate why a regulator would want to abandon the UT l ~ examination for the closure weld after it had demonstrated and verified for this application, with
. very acceptable doses and in light of known welding process problems. I think a reasonable regulator would want to change the fabrication practices to prevent known problem areas from reoccurring rather than elect to keep the status quo.
4
g ase - .
~~ $
- [ UNITED STATES
- 5 NUCLEAR REGULATORY COMMISSION WASHINGTON. O,c. 308464001 November 10, 1998 MEMORANDUM TO: Carl J. Paperiello, Director - .
Office of Nuclear Material Safety and Safeguards
' FROM: Thomas O. Martin, Chief * '
- Generic Safetyissues Branch Division of Regulatory Applications h() % .
- ~
Office of Nuclear Regulatory Research
SUBJECT:
- DIFFERING PROFESSIONAL VIEWPANEL j
On September 15,1998, I received a memorandum from you p.oviding a diffe view (DPV) submitted by Ron Parkh!!! on a Spent Fuel Project Office position(
along with the other meinbers Geoffrey Homseth, NR '
- A report of the DPV panelis attached. Please feel free to contact me or thl -
' rnembers with any questions or comments on this report.
- C
Attachment:
Asstated ii CC:
4 R. Parkhill, SFPO -
W. Hodges, SFPO .
G. Homseth, NRR
' D. Jackson, RES 1
J. Craig, RES -
. 1
- i l
1 i
' Attachment 2 l
gg. s L
[ "
- r. ..
Panel Report Concerning the September 8,1998, Differing Professional View Submitted by Ron Parkhill I. Bsckground .
This report discusses the review of the Differing Professional View (DPV) dated September 8, 1998, submitted by Ron Parkhill (Attachment 1) to Carl Paperiello, Director, NMSS. Following the receipt of the DPV, s' panel was tasked to review the DPVin accordance with NRC t
rnanagement Direc'ive.19.159, " Differing Professional Views or Opinions." The panel members were Thomas Martin (Panel Chairman), Deborah Jackson, and Geoffrey Homseth. The DPV involves the nondestructive examination of closure welds on stainless steel spent fuel storage casks.
II. DPVSummary The primary concem of the DPV was with the reliance on liquid penetrant surface examination (PT) for cask closure welds as opposed to ultrasonic examination (UT). The position developed i by the Spent Fuel Program Office (SFPO) relies on PT for the root pass, final pass, and approximately every 1/4 inch of weld. Mr. Parkhillis of the opinion that UT ought to be required and provided specific concerns related to this issue.
Ill. Summary ofissues Reviewed by the DPV Panel Sufficient documentation was provided by the involved parties for the Panet to undertake a review of the DPV. The Panel reviewed portions of the Safety Analysis Report'for both the NUHOMS-MP187 and Westflex storage casks and detailed drawing for the NUHOMS-MP187 cask. Estimates of transfer cask radiation dose rates were also provided by the SFPO for the outer closure welding configuration. The panel also held a discussion with Mr. Parkhill.
In order to resolve this issue the Panel focused on the following issues:
the flaw types and causes encountered for welding this material and whether PT is adequate to identify these types of flaws a
the pertinent code and regulatory requirements a
'the difficulty of conducting UT
=
radiological and Al. ARA aspects A. FlawTypes and Causes Encountered for Welding Austenitic Stainless Steel l Based upon staff and industry experience, as supported by a literature survey, the Panel concluded:
- 1) ~ Linear type we'id flaws (hot cracks) with some propensity for' occurrence are
. predominately surface connected flaws.
- 2) Of the types of process defects that may generate sub surface flaws, the defects would be bounded in depth by the thickness of a single weld pass.
- 3) The likelihood of service induced propagation of any potential flaws is negligible due to the lack of fatigue loading during design or accident conditions.
A more detailed discussion of Austenitic stainless steel weld flaws is provided in attachment 2.
l
c B. The Pertinent Code and Regulatory Requirements Mr. Parkhill refers to the ASME Section 111 as the gov'eming code for the confineme
. This is based on the acceptance criteria in the Standard Review Plan (SRP) for Dry Cask
' Storage Systems, NUREG 1536, which specifies that the NRC staff has accepted constru of the primary confinement boundary in conformance with Section lil, Subsections NB or NC.
The SRP also states that, after careful and deliberate consideration, the staff has made exceptions to the requirement that the applicant must fully document and completely justi deviations from the specifications of Section lil. The Code edition or date is specifically not
^
addressed in the SRP. The Pane 1 noted that the sections of the Code reference were not written to be applicable to spent fuel storage casks, and the NRC has presumably chosen this code for convenience of providing a high integrity boundary. The SRP also states that the staff has relied upon Section ill to defins the minimum acceptable margin of safety The fact that the ASME code is not specifically identified in the regulations gives the NRC the Jeeway to take their present position.
The Panel also becarrye aware that there is a 1998 ASME, Section 111, Division 3, Code for containment systems and transport casks for spent fuel. This Code has not yet been endorsed.
However, unlike part 10 CFR Part 50.55a which specifically mentions the ASME Code in the regulation,10 CFR Part 72 does not specifically reference or endorse a code.
C. The Difficulty of Conducting UT
~ Austenitic materials can be inspected effectively using UT, and these type exams are perform routinely as part of ISI programs. However, there are exceptions. Effective UTinspection of austenitic materials is dependent on thu weld configuration. In some cases it's not possible and in others extremely difficult. Unlike carbon steel which has a smooth grain structure, austenitic material has a coarse grain structure. A UT exam of austenitic materials results in the s of the sound beam at the grain boundary in the weld making it difficult to distinguish the gra boundary from flaw signals. If the flaw being detected is much larger than the grain size then it may be possible to detect the flaw with high reliability but conditions such as access, existence of crown or root design need to be considered. The Panel was unable to determine whether it was possible to effectively use UT for these casks.
D. Radiologicalan'd ALARA Aspects Based on a limited review of the experience performing NDE on actual casks, radiation exposure would not be a significant factor in selecting the NDE method. There has already been a reasonable amount of experience to gauge the radiologicalimpact of performing PT a UT on cask closure welds. For the casks that are presently receiving a UT inspection, PT is performed on the root and cover weld. Specific data was not readily available because PT and UT exposures have not been tracked separately. Some specific data points from Point Beach include an exposure of 5 mR for PT on the root weld of the shield lid and an exposure of 40 m for the total UTwork on a cask. The NDE work on the Palisades casks averaged 45 mR.
2 9
h.-
[ .. ,
l5 -
l IV. Panel Conclusions and Recommendations
' ~
The Panel concludes that employing PT of the weld periodically after a specified deposit depth
, will provide adequate assurann hat no weld flaws greater in depth than the specified PT interval (depth) would be plauale. However, before the technique is employed, the critical flaw
~
size for the component must be determined from fracture mechanics or other suitable analysis.
The critical flaw size then determines the maximum depth of weld deposit that can be made before a PTis performed. This is consistent with the SFPO position.
' Based on the review of the DPV and associated information, the Panel makes the foliowing recommendations:
- 1. The Panel recommends that an appropriate Code spec!.'ic to the constructiori and use of i
these casks be endorsed and applied. The NRC has not consistently applied minimum acceptable requirements for the level ofinspection of these casks. The standard initially intended for this purpose was developed for and is applicable to reactor pressure retaining components. The lack of an NRC endorsed standard applicable to the construction of these casks is, in the opinion of the Panel, not conducive to promoting public confidence in the use of these structures. This is exacerbatsd by the inconsistency of requiring volumetric NDE for cask l
construction and then permitting a surface examination technique for the closure welds. The
! Pane 1 believes that it would be appropriate to perform a flaw tolerance evaluation of the cask to
~
Essist in determining the appropriate NDE technique, whether volumetric or surface, that.could then be applied consistently to all welds.
i
- 2. If PT1s used to verify closure weld integrity, priority should be given to providing NRC inspection oversight of this process. Performing the sometimes tedious PT activities in a high radiation environment could create a tendency to perform some of aspects of this work too fast with a potential for missing indications or contributing to the potential for creating a flaw by, for example, inadequate cleaning of the liquid penetrant before the next weld pass.
Submitted by:
l h r Thomas O. Martin, DPV Panel Chairman W d Deborah A. Jac l
l
.2 % % P % A GeoffreW. Hor $eth
\ -
i 3
,. (- ~
- )
DIFFERING PROFESSIONALVIEW l
This Differing Professional View (DPV) is being initiated because SFPO staff concems remain l
. unaddressed with regard to the nondestructive examination (NDE) method for the spent fuel I
. storage canister closure weld. Proposed near term licensing actions by SFPO allow the surface examination for the subject weld whereas i feel that a volumetdc examination is ju forthe following reasons. ;
I On 7/20/98 members of the SFPO licensing directorate forwarded a position that did not re
- volumetric examination of the spent fuel storage canister closure weld and notified some of the appilcants of that preliminaryposition. Basically, the aforementioned position relies on surface examination, liquid penetrate (PT), for the root pass, final pass and every 1/4 inch of weld, as well as a reduction in the allowable stress for the weld. On 7/21/98 the SFPO
- that a complete written position and safety rstional be developed by the technical review directorate prior to advising any applicants. On 7/24/98, 7/29/98, & 7/30/98 three members technical review section sent e-mails supporting volumetdc examination of these welds and raised numerous concems with the proposed position. Last Friday I was informed that management had decided that surface examination was " good enough* for this spent fuel storage canister closure. However, no written position has been developed whichjustifies its use. Consequently, the following issues remain unaddressed and form the basis as to whyi believe that a volumetric examination needs to be performed on the subject weld. !
- 1) i Subsurface flaws wf!! go undetected by surface examination techniques unless the flaw penetrates the surface of the weld area being examined. It has already been' demonstrated (via the VSC-24 welding problems) that surface examinations alone are relatively ineffective in identifying subsurface cracks resulting from a poor quality w process. For the VSC-24 situation,19 casks were loaded before surface examination in combination with leak testing discovered that there was a problem with the welding process. Had volumetdc examination been employed the werd process problem would have been discovered before many of these casks were loaded.
2)
The goveming code for the confinement boundary is ASME Section 111, Subsection NB or NC as stated in the current SRP and historical practice for storage canisters. As j such, volumetric examination is required to be performed on 3!! confinement welds. a (Refer to NB/NC-5213 & 5220 for Category A & B welds, respectively. Note that j radiography is the volumetdc method mentioned in these Code paragr)aphs, s! 1 intended that a!! parts of the reactor coolant pressure boundary would be accessible during fabrication. However, UT is an acceptable substitute where RT cannot be performed (e.g. closure weld). The propose use of surface examination by SFPO is a deviation from the goveming Code's requirements. i
~ 3) . The ductirity of stainless steel should not be a basis for avoiding volumetric examinatio of the canister closure weld since volumetric examination is not done solely to determi flaw sizes have been bounded by a fracture mechanics analysis. Volumetric examination is done principally to ensure weld quality during fabrication. It is a ;
j verification that the weld meets design requirements and a vedfication of the weld '
process. If the ductility of stainless steel were the only consideration for performance of an UT volumetdc enmination then the goveming code would have excluded ductile j
ATTAQDENT 1
r- - -
2 l stainless steel from UT examinations, which it does ngj do. UT examinations of 4
. stainless steel components are performed routinely as part of Part 50 ISI programs. The exclusion of ductile stainless steel from volumetric examination is a deviation fro goveming Code's requirements.
- 4) The ' suggested approach of surface examination for root and final weld pass as well 4 overy 1/4 inch of weld is what has been proposed in a draft ASME Code Case', currently in Revision 12. However, this Code Case has nemen adopted by ASME and may undergo significant change prior to issuanoe. Furthermore, the basis for the 1/4 inch spacing between PT examinations has not been pier reviewed and I suspect that K has not included consideration of additive spacings between pts being more than the
- calculated critical crack size (e.g. If the closure weld is 3/4 inch, the fracture mechanics critical crack size would need to be more than 3/4 - (4 X 1/16) or % inch, which is 67%
of the wall thickness). Pdor to adopting a preliminary position from the consensus Code group it may be prudent to await its final version, otherwise we are again in ,
disagreement with the goveming Code.
4
- 5) Reducing the allowable stress to compwasate for surface examination (i.e. PT) in lieu of a volumetric examination (i.e. UT)is not a method utilized in the goveming cork for the confinement boundary (i.e. ASME Section ill, Subsection NB or NC). No amount of allowable stress reduction will compensate for a poorly made weld. This is an example
' of mixing and matching different code requirements. The goveming Code requirements should be followed both by the industry, as well as, the NRC. Again, this is a deviation ,
from the goveming Code requirements.
~
- 6) !
Historically, in the SOC for the VSC-24 rufemaking, Comment #45, the NRC response ,
stated that the closure welds meet all ASME requirements except for volumetric examination and further stated that this inspection was not possible due to radioactive !
j fuelin the cask. Today we know that this documented basis for not doing the volumetric t
examination is not correct. We now know that the projected Al. ARA dose is very low j based on the work performed by the VSC-24 Owners Group and that the UT examination is viable for this application. Furthermore, doing PT every 1/4 inch will '
significantly increase dose - each PT involves approximately 200 linear inches of weld involving cleaning, application of penetrant, dwell time, removal of excess penetrant, ,
application of developer, drying, evaluation of results, recording of results and removal of PT materials prior to continuing welding. !
s
- 7) Regulation 10 CFR 72.236(e) requires that the cask be designed to provide redundant -
sealing of the confinement system. However, the redundant seating requirement cannot be met if a single failure would cause both seals to fall. The single failure that would bypass the redundancy requirement would be a failure of the closure welding process The single failure I'm alluding to in this case is more than just a postulated occurrence-it has happened for the VSC-24. Therefore, to prevent other single failures in the w process volumetric examination should be utilized for examination of the closure weld, just as it is utilized for all other welds in the confinement boundary.
, ATTACHMENT 1
~ln summary, I believe that volumetric examination should be performed on the confinement
- closure welds since all other confinement welds are volumetrica!!y inspected, it is required by the goveming code and would verify that there is not a welding process problem (i.e. single failure). It is very hard for me to appreciate why a regulator would want to abandon the UT examination for the closure weld after it had demonstrated and verified for this application, with very acceptable doses and in light of known welding process problems. I think a reasonable regulator would want to change the fabrica:ior, practices to prevent known problem areas from reoccurring rather than elect to keep the status quo.
l l
l l
1 l
t l
ATTACHMENT 1
a y, ,' .
L . ,7 .
i l
f ,. ' Attachment 2 1
AUSTENITIC STAINLESS STEEL WELD FLAWS A literature survey of prominent flaw types and causes.
'Ar a result of theDPV submitted by R. Parkhill (NMSS/SFPO) regarding volumetric inspection requirements for cask closure wsids in austenitic stainless steel casks, a literature survey was undertaken to examine the prNominant types of weld flaws that are encountered when welding -
this material. With knowle@e y use most likely flaw types, a selection of the optimum inspection method (s) can be made.
A list of the references consulted is included at the end cf this discussion.
BACKGROUND Fourinstances of closure weld cracking were experienced during the 1993 to 1997 period.
l .
These instancea all occurred in a specific carbon steel cask design known as the VSC-24, designed by S erra Nuclear. These events prompted the NRC, the users, and the design company to investigate the root cause and determine suitable corrective actions. The most likely cause of ths observed cracking events was determined to be hydrogen induced, or '
delayed, cracking. The corrective actions involved several changes and improvements to the cask loading and venting procedures and the welding technique.
L I
Additionally, the NRC staff determined that a volumetric examination of the structural lid weld was necessary in order to confirm weld conformance with the design requirements. Previously, ,
weld examination consisted of a liquid Penetrant Test (PT) surface examination and a helium leak test. The reason for the change to a volumetric examination was the staff opinion that delayed cracks of significant size could potentially form that would not breech the top surface of the weld, thereby remaining undetectable by the PT examination method then employed. A volumetric examination method, ultrasonic test (UT), was consequently employed by the Industry.
i l POTENTIAL FLAW TYPES IN AUSTENITIC STAINLESS STEEL WELDS To determine if an examination technique is potentially useful, an understanding of the types ~of plausible flaws is necessary. Welds in stainless steel are susceptible to some similar and some
! different weld defects compared to those encountered when welding carbon steel. One major difference in possible mechanisms involves hydrogen (or delayed) cracking.
' Carbon steells potentially susceptible to hydrogen cracking. The metallurgical reasons are well understood and the problem may be eliminated by employing one or more of several rnessures.
Austenitic stainless steel, due to its different atomic structure, is well recognized as being nearly L
Immune to hydrogen cracking. Thus, the need for a volumetric method to inspect for this type of flawis eliminated.
l The most cited weld defect arising when welding austenitic stainless steels is hot cracking. Hot cracking occurs during orimmediately after solidification of the weld bead. It usually occurs as a longitudinal crack in the center of the weld, initiating at th'e top surface of the bead.
u
,'r 4 .
Attachment 2
.' a
\
Occasionally, it will appear at the edge of the weld or as short transverse cracks. Since It is l open to the surface, the cracks are generally readily visible. A penetrant test would readily detect this' type of Saw. i Hot cracking is controlled by using filler metals with a specified " ferrite number" of 4 or hig The most commonly, employed stainless filler metal (type 308) has a ferrite number suff high to meet this mquirement. Hot cracking may also be alleviated by employing certai
. techniques: maintaining a convex bead shape, avoiding high heat input, avoiding wide weld passes orweave. - .
The propensity for generating buried linear flaws rach as can occur in carbon steelis relat1 small. The most significant mechanisms that exist for producing these types of defect are well understood and readily avoided. These are: lack of ponetration and lack of fusion.
Lack of penetration is a longitudinal crack-like defect that results when the root of the weld fails to completely fuse with the base material. This is generally avoided by ensuring good weld roo fit-up. Wide gaps or uneven gaps are to be avoided. Lack of penetration flaws do not grow during subsequent welding. Thus, their potential depth is bounded by the root pass thickness.
The propensity for producing this kind of flaw is lowin a well controlled welding program. T greatest potential for causing failure during service occurs when a fatigue load exists in the affected weld. For a dry cak, no significant fatigue load exists under any of the design or accident' conditions. '
Lack of fusion (LOF)is the failure of the weld to fuse with the adjacent base material or weld bead. An excessively large weld puddle or wide weld pass would suggest possibleLOF LOF.
can also be indicated by e rolled over bead crown. This shape results from poor wetting of th adjacent material by the weld puddle. This bead shape is readily discemable visually. Also, to the less fluid characteristics of the stainless steel weld puddle (compared to carbon ste LOF defect would have a greater propensity to be surface connected and detectable by Other weld technique measures may also be employed to avoid the potential for LOF. Too low a heat input, wide passes, and large weave all contribute to the problem. Avoidance of short-circuit arc transfer when using gas metal are weiding (GMAW) will minimize the propen LOF when this process is used. The propensity for producing this kind of defect in a well controlled welding program is low. The likelihood for generating such flaws is considered to be higher in manual welding processes than in semi-automatic or automatic processes. LOF defects will be limited in size to the depth of the deposited bead. Subsequent welding pa are unlikely to make the defect grow. In service, a significant fatigue load would be required propagate the flaw as a crack.
The above potential defects are regarded as the most potentially serious due to their line crack-like morphology. Less potentially serious defects of a non-linear morphology may a occur: porosity and slag inclusions. Porosity is rounded voids within the weld bead resulting from gas trapped in the weld puddle. It may be either sub-surface or surface connacted. De
- diameter is limited by the size of the weld bead. Clusters of smaller pores may occur over the length of the bead. Generally it arises from welding contaminated surfaces, impure w or insufficient shielding gas. The conditions resulting in porosity may be readily avoide 2
,u .
Attachrnent 2
' following good welding practice. The propensity for producing this kind of defect in a well controlled welding program is low. Since porosity is rounded, its adverse structural effects are very much less than for linear defects.
' Entrapped slag may occur when flux shielded (as opposed to gas' shielded) welding processes are used. It results.fror'n inadequate slag removal from previous weld passes. Defect size is limitedbythe size of the residual slag. Slag inclusions tend to be either rounded or laminar and often create's surface connected' pore that is readily visible. With a well controlled cleaning process and ir-*yn, the propensity for this kind of defect is low.
MISCELLANEOUS CONSIDERATIONS .
A relatively common cause of problems for inert-gas s,hielded processes is the inadvertent loss of the shielding gas due to the local " environment". These are often created when the welding conditions are uncomfortable to the welders. Common causes are requirements for preheat of the work piece, uncomfortably hot conditions in the work area, or requirement for ' dressing out" in protective clothing. Often the work crew will attempt to alleviate the heat by us!ng fans, blowers, vacuum hoses, etc., at the work site. This practice has been observed to cause loss of the shielding gas at some locations around the weld joint, a result of the wind created by the air movers. The solution is obvious. Open shop doors adjacent to the work area are also frequently encountered culprits in this problem.
The effect ofloss of shielding is readily apparent during a visual examination. The weld surface will have a coarse granular or " sugared" appearance. Since the cause is oxidation of the weld, wr> ' j and fusion will also be poor and frequently manifest itselfin a more rounded bead with nowhes at the bead edge. Requiring a visual exam after each pass and again after completing one layer would assure freedom from incorporating such a defect into the finished weld.
A positive method to enhance the chances of the PT to be able to detect subsurface flaws would be to require a partial grind-out of each pass prior to performing the PT. This obviously would aid in uncovering previously subsurface flaws that would have escaped detection. It would reduce the maximum depth of an undetected flaw. For maximum detection and to aid in reducing false positives, each pass of the weld could be ground and inspected. The disadvantages are the required grinding, additional number of PT's, and extra weld time.
Verification of the welding process and controls could also be performed. This would incorporate use of a full size mock-up (or full size section) that would be welded and inspected prior to performing the production weld. The mock-up could then be sectioned or radiographed, '
etc. to verify the process quality. The probiern with the method is; to capture the complete conditions of the production weld (s), a complete lid assembly would be needed. Only that would produce the same degree of constraint and residual stress, significant factors contributing to weld cracking.
STAINLESS STEEL CASK WELD EXPERIENCE Weld cracking was experienced during the performance testing conducted during the
, development of the NUHOMS cask design, as discussed in EPRI report, NP-6941. Surface 3
f
,y [ ~
Attachment 2
[ . connected weld cracks occurred when the initial tat,k welds were placed. Tile root causes were evaluated to be due to the high degree ofjoint constraint, too high a heat input, and poor fit-up.
l
' A weld development program was undertaken that resulted in several changes to the we technique and the ' elimination of weld cracking. The improvements included a change to weld sequencing to reduce shrinkage and heat input, improved fit up, use of a high ferrite content weld metal, and, for the shieldlid, a modification of the joint design to reduce constraint.
CONCLUSION Based upon staff and industry experience, as supported by the literature survey, the Panel finds:
'1) Some linear type weld flaws (hot cracks) with a propensity for occurrence are predominately surface connected flaws. .
- 2) Of the types of process defects that may generate _sub surface flaws, the defects would be bounded in depth by the thickness of a single weld pass.
- 3) The likelihood of service induced propagation of any potential flaws is negligible due to the lack of fatigue loading during design or accident conditions.
Employing PT of the weld periodically after a specified deposit depth will provide adequate assurance plausible. Howe that no weld flaws greater in depth than the specified PT Interval (depth) would be
'er before the technique is employed the critical flaw size for the component l
~
must be deterr ,d from fracture mechanics or other suitable analysis. The critical flaw size then determini. ,1e maximum depth of weld deposit that can be made before a PTis
- performed. 1
~
REFERENCES.
Mio Weldino Handbook. Union Carbide Corp., New York,1981 l
Weldina Enoineerino, Continuing Engineering Education Course Notes, R.L. Edwards, P.E.,
given at The George Washington University, Nov.1982.
Weldability of Steels,4th ed., R.D. Stout, Welding Research Council, New York,1987.
ASM Handbook.Vol.6, ASM International, Metals Park, Oh.1993.
j Metals Handboo_k, Vol.10. 8th edition, American Society for Metats, Metals Park, Oh.,1975.
introductorv Weldino Metaflurov, American Welding Society, Miami,1979.
Stainless Steel Fabrication. Allegheny Ludlum Steel Corp., Pittsburgh,1959.
L The Procedure Handbook of Arc Wefdino,12th ed., Lincoln Electric Co., Cleveland,1973 NUHOMS Modular Soent. Fuel Storace System: Performance Testino, EPRI NP-6941, Elec Power Research Institute. Palo Alto,1990.
Weld inteority and Performance ASM Intemational, Metals Park, Ohio,1997.'
4 4
O
NUCLEAR E.NERGf 145TliUTE L e Hendricks l$[2ESLmwsa October 20,1998 Mr. William Kane Director, Spent Fuel Project OfIice Office of Nuclear Material Safety and Safeguards U.S. Nuclear Regulatory Commission One White Flint North 11555 Rockville Pike Rockville, MD 20852
SUBJECT:
Industry Spent Fuel Dry Cask Weld Crack Position Paper, October 1998
Dear Mr. Kane:
On August 26,1998, the industry and NRC held a roundtable discussion at the NRC ofIices in Rockville, Maryland, to evaluate different approaches to volumetric examination of dry spent fuel storage cask welds.
Three pre-service inspection weld examination methods are being evaluated by the i industry. Utilities would benefit from the flexibility provided by more than one examination l method. Examination methods being evaluated include ultrasonic, liquid penetrant i examination of root and intermediate, and final weld passes, and root and final passes. !
I The industry is seeking regulatory concurrence for approaches to weld examination j methods supported by our," Spent Fuel Dry Cask Weld Crack Position Paper,"(enclosed). ;
If there are any questions regarding the position paper provided please contact me at (202) 739-8109 or by e mail (1xh@nei.org), or you may contact Ahn Nelson at (202) 739-8110 or e-mail (apn@nei.org). NEI is available to meet with the NRC to discuss these issues further.
Sincerely, c
/ dA \ '- - !
i Lynnette Hendricks
.D inI0mlationfuuvW-- la " Enclosure x w~ d , ...? i / E l-l In accordance m.tsB57,ighwM,a irl!R C w- - _ - '
Act, cxel ptinnt 7 NONk
~
MA* . . , _
.n
F l
October 9, 1998 SPENT FUEL DRY CASK WELD CRACK POSITION PAPER j INITIATING EVENT:
l Incidences of cracking were identified during welding of the closure lids of a Ventilated Storage Cask Model No. 24 (VSC 24) carbon steel canister at three reactor sites: Consumers Energy Palisades, Wisconsin Electric Point Beach, and Entergy Arkansas Nuclear One. The cracking was discovered during visual and surface non-destructive examination (NDE), and by helium leak testing. l l NRC ACTIONS AND REQUESTED LICENSEE ACTION:
The NRC staff has issued Generic Safety Issue NMSS-11, Spent Fuel Dry Cask Weld Cracks to evaluate the potential for cracking of primary and secondary l confinement closure weld and adjacent heat-affected zone (HAZ) in spent fuel dry casks. The GSI was categorized'as a medium priority in NUREG-0933. The NRC i staff plans to conduct research to assess whether corrective actions implemented by industry have resolved the problem of cracking in dry cask closure welds.
The NRC staffis recommending that outer closure welds in carbon and austenitic stainless steel spent fuel dry casks be ultrasonically inspected to the acceptance criteria of ASME Section III during the spent fuel loading process.
INDUSTRY ACTIONS:
l Carbon Steel Casks: The vendor, Sierra Nuclear Corporation (SNC), the cask owner, Consumers Energy, and members of the VSC 24 Owners Group (Wisconsin Electric Power and Entergy Operations, Inc.) have committed to ultrasonic testing (UT) for examination of the outer carbon steel closure (structural lid) weld. The i time-of-flight difTraction UT examination methodology will be used, and has been demonstrated to be capable of flaw detection on a canister closure weld mockup to the satisfaction of the NRC.
Austenitic Stainless Steel Casks: Although ultrasonic examination is widely used for inservice examination of austenitic stainless steel components, such as piping welds, in nuclear power plants, it has not yet been applied to austenitic stainless steel spent fuel dry cask welds. Detection and sizing of small defects using UT would be influenced by signal deterioration in stainless steel base metal and welds due to the large grain size present in the austenitic material and is a 1 ,
j
F L ,
[-
L October 9, 1998 j u ; potential issue. An alternate volumetric examination using radiographic testing L (RT) is not considered un option because oflimited source / image accessibility.
- l. Licensees are considering alternative preservice examination approaches in light of l
, , these potential technical considerations. Some are considering progressive dye i pe'netrant testing (PT). Others are considering UT, which is expected to be feasible L for detecting th'elarge allowable flaw sizes that have been identified for austenitic l . stainless steel storage casks.
l INSPECTION STANDARDS BACKGROUND:
l No codes or standards exist that address the design, material selection, fabrication, examination, and testing of confinement boundary welds in spent fuel dry casks.
' However, instructive guidance is provided in the ASME Section III, Division 3 rules for the construction of containment boundaries of spent fuel and high-leve'l waste l transport packaging. Transport cask closure lid-to-shell welds are defined as i Category C welds. Subsection WB'of Division 3 requires radiographic (RT) examination following fabrication, but permits an alternative inspection that L combines UT and surface examination [ magnetic particle (MT) or PT] when access l
for the radiographic source and image is limited. This requirement is similar to the
- Category.C ' welds requirements contained in ASME Section III, Subsection NB, for l Class 1 components. The ASME Code does not permit substitution of multiple surface examinations in lieu of the UT portion of the examination. However, Subsection NB permits progressive PT examination for partial penetration welds l' for nozzle attachments to Class I components.
l l i l
DISCUSSION:
Utilities would benefit from the availability of alternative examination methods for ]
.austenitic stainless steel spent fuel dry cask closure welds. Industry is evaluating several approaches to surface and/or volumetric examination of these austenitic stainless steel spent fuel dry cask welds.
Existing canister weld configurations preclude radiographic examination. Other j l
possible examination methods under evaluation for austenitic stainless steel spent j l
l : fuel dry cask closure welds include UT examination and liquid penetrant ]
. examination of root,' intermediate, and final weld passes, or possibly limited to root j r and final weld passes.
)
Detecting flaws in au.stenitic stainless steel welds using ultrasonic examination methods is normally very difficult, but may be possible if the critical flaw size is 2
L.. #
(.
\
- j l
October 9, 1998 l l large. When flaws are sufficiently large, the UT signal deterioration experienced in austenitic stainless steel will not compromise the detection and sizing accuracy. ,
If the critical flaw size ofinterest is sufficiently large, use of progressive dye i penetrant examination may be a practical alternative inspection method that will result in reliability of flaw detection equal to that of UT for austenitic stainless steel applications. The progressive dye penetrant method inspects the weld root
! pass and final pass or inspects the weld root pass, an intermediate pass, and final I pass. Because the critical flaw size is large, PT surface examinations can assure l identification of unacceptable fabrication flaws and would result in assuring an equivalent structural integrity and a reduced personnel radiation exposure. j Proposed Preservice Examination Alternatives: Three alternatives for inspection of austenitic stainless steel dry storage cask welds are discussed below:
(1) UT; (2) multiple-pass PT, with at least one intermediate weld pass examination; (3) root and final weld pass PT.
The specific elements of each of the three approaches include acceptance criteria for flaws detected by examination, acceptable examination methods, criteria for examination, and criteria for examination qualification.
Because of the intermittent nature of the PT examination and the deterioration of the ultrasonic signal as it propagates through the larger grain structure of the stainless steel, a large target flaw size is a pre-requisite for reliable detection.
Therefore in each approach a common element is the need to justify a large target flaw size for the preservice examination. For cample, the target flaw size for ultrasonic examination significantly affects the choice of technique, the approach used to qualify the examination, and the cost of the subsequent examination.
Similar arguments apply to liquid penetrant examination. Therefore, the target flaw size should be selected with due regard to the relatively large allowable flaw size and the characteristics of the types of weld defects expected in austenitic stainless steel weld and HAZ material. Both of these topics are addressed in the following discussion.
Weld /HAZ Defect Characteristics The weld and HAZ defect characteristics discussed here are limited to those that might occur during or immediately following the fabrication process, without consideration of service-induced defects.
For example, the potential for service-induced intergranular stress corrosion cracking (IGSCC)in sensitized materialis not addressed. However, the conditions under which crack extension by the delayed hydrogen cracking (DHC) mechanism 3
- i l
L.
October 9, 1998 have been evaluated, and it.was detennined that DHC is not of concern. An assessment 'of DHC is presented later in this document.
Four types of fabrication-induced . weld defects are addressed: ,
(1) lack of fusion from a poorly implemented welding process, procedure, or .j technique- 1
)
(2) slag inclusions from processes that utilize flux (e.g., shielded metal-arc (SMAW), submerged-are (SAW), and flux-cored (FCAW) welding processes; (3) hot tearing in the HAZ during automatic gas-tungsten arc (GTA) or gas-metal-arc (GMA-) welding processes; and
' (4) weld metal cracking caused by excessive joint restraint and relatively-high cooling rates in GTA and GMA welds.
Lack of fusion will b'e minimized through adherence to the Welding Procedure Specification (WPS), Procedure Qualification Record (PQR), and Welder / Welding Operator Performance Qualification requirements of the ASME Code Section IX. In addition, the extremely low probability for a connected lack-of-fusion defect would
- be detected by multiple surface examinations, in particular for the PT examination of the root pass of the weld.
Slag inclusions will be eliminated through the use of either GTA or GMA welding processes. In addition, consideration must be given to the potential for tungsten electrode contamination of the weld metalin the GTA process, by maintaining proper distance from the molten stainless steel puddle during the welding and using arc initiation techniques that do not require intimate contact (e.g., high frequency start) with the material.
Hot tearing in the HAZ is minimized through the use of either 304'or 304L base metal, and ER308 or FR308L as the filler metal, rather than the stabilized grades (e.g.,321 and 347).- This limits the potential for molybdenum, titanium, or columbium (niobium) carbide precipitation in the HAZ from the stabilizing element, in preference to chromium carbide precipitation. Stabilizer carbide precipitation in the HAZ can lead to hot tearing at the toe of the weld. The use of non stabilized base and filler metals essentially eliminates this concern, even though the potential for sensitization and subsequent service-induced cracking may be increased.
The most likely form of weld fabrication cracking is caused by various combinations of excessive joint restraint, the presence of excessive residual elements (e.g., sulfur, i 4 k
y L
October 9, 1998 phosphorus, or silicon), rapid cooling (from the GTA or GMA welding processes),
and ferrite content outside the range of 5 to 20 percent, as determined from the Schlaefiler diagram. The cracking associated with these combinations is characterized by microstructural defects at the ferrite-austenite boundaries during initial solidification and reheating during subsequent passes, with such defects located at or near the weld surface. The defects are expected to be very small, J l unless the effect ofjoint restraint is sufficiently pronounced to cause linkage of l grain boundary defect structures. PT examination will be very effective in detecting linked grain boundary defects, and has a reasonable likelihood of detecting small surface ferrite-austenite grain boundary defects.
Slag inclusions will be eliminated through weld procedure selection. HAZ hot i tearing will be eliminated through the use of non-stabilized grades of stainless steel base and weld metal with minimum delta ferrite content.
Continuous lack of fusion and weld metal cracks of any substantial size will be
. and at least one additional PT examination after an intermediate pass and after the ,
final surface pass. Intermittent lack of fusion will be minimized by weld procedure i specification and qualification, and by welder / weld operator performance qualification.
Allowable Flaw Size and Acceptance Criteria: Acceptance criteria for preservice examination flaw detection and sizing will be based on:
(1) lower-bound elastic-plastic fracture toughness data for the aestenitic stainless steel weld material, similar to the data used to generate end-of-evmaation period flaw acceptance standards for SAW and SMAW in Tables IWB-3641-5 and IWB-3641-6 in ASME Section XI, Subsection IWB rules for inservice inspection of Class 1 nuclear power plant components; (2) postulated flaws located in the closure weld at the lid /shell interface and oriented in the axial /circumferential plane; (3) radial tensile and bending stresses that are assumed to be equal to design-limiting values; and (4) an evaluation criterion from ASME Section XI, Appendix K (see Paragraph K-4220), that limits the elastic-plastic crack driving force to be less than the material elastic-plastic resistance curve value at a crack extension of 0.1 inches. Stability of crack extension is chosen as the basis 5
l t
4 F
October 9, 1998 Lfor acceptance, rather than crack initiation, because of the ductility shown ;
-by the lower-bound fracture toughness.
1 i
Bounding Flaw Tolerance Calculations: Acceptable geometric dimensions of l austenitic stainless steel weld defects are defined using the result for a postulated ,
flaw in a limiting weld geometry with limiting design loads as compared to lower-bound crack growth resistance (J-R) curves. The bounding resistance curve data -
are described in EPRI TR-106092 [1] and are shown in Figures 24 and 25 of the I reference. From these curves a crack-extension value of 2.5mm (0.1 inches)is '
conservatively chosen as the onset of unstable crack growth for which the limiting :
' J-R resistance curve 'value at a crack extension of 2.5 mm (0.1 inches)is approximately 250 kJ/m2 (1430 in-lb/in2).
The postulated flaw in the closure weld is shown in Figure la, and the idealized semi-elliptical geometry is shown in Figure Ib.- For a closure weld that satisfies the
' ASME Code geometric requirements for a Category C full penetration weld the ,
flaw is shown in the axial /circumferential plane at the inner boundary between the -
closure lid and shell, extending vertically along and circumferentially around the
- shell lid interface. The flaw is characterized by the non dimensional parameters 2c/W, a/c, and a/t, where 2c is the width of the flaw, a is the flaw depth, t is the thickness of the closure plate, and W is the closure plate circumference. A flaw extending completely around the closure lid is defined by 2c/W = 1. A flaw ,
extending completely through the full penetration weld is defined by a/t = 1. For a partial penetration closure weld, a discontinuity between the lid and shell may be present below the weld. The flaw definition takes any such discontinuity into account by defining a/t to encompass any discontinuity below the weld plus the extension of the discontinuity (flaw)into the weld proper.
For purposes of this evaluation, the radial stresses will be assumed to be either a
- remote radial tensile stress, acting across the closure lid cross section, equal to the yield stress, or a radial bending stress, acting across the closure lid cross section, equal to twice the yield stress.
An estimate of the crack-driving force suflicient for the purposes of this evaluation can be obtained from existing results in the literature for a geometry that approximates that of the closure weld assembly (see Figure 2). For example, Raju, et al. [3] have published stress intensity factor solutions for surface cracks in flat plates subject to remote stress fields that are uniform, linear, quadratic, and cubic functions in plane's parallel to the crack face. The solutions are given in the form Ki = a., F V n a/Q, 6
e N
October 9, 1998
~
- where Kr is theLapplied stress intensity, ao is the stress amplitude, F is the non-
- dimensional stress intensity factor, a is the crack depth, and Q is the crack'-
l geometric shape factor given by -
Q = _1 + 0.464 (a/c) , for a/c less than or equal to 1, !
iand j Q =.1~+ 0.464 (a/c)", for a/c greater than 1. -
Values of F were calculated in Reference 3 for values of 2cSV.= 0.0,0.1,0.4,0.6,0.8,'-
~
and 1.0; for values of a/c = 0.2,0.4,0.6,0.8, and 1.0; and for values of a/t = 0.05,0.2, 0.5,0.8, and 1.0 (a completely through-wall flaw).- For particular cases examined, the applied stress intensity is converted to an applied J-integral using the formula
'in ASME Section XI, Appendix K,' Paragraph K-4210 J = 1000 (KI)2/E' i where J is in units ofin lb/ing, Kr is'in ksi Vin, and E'is in ksi.
Bounding Flaw C'alculation: The typical spent fuel dry cask lid closure weld has a circumference of 300 inches and a thickness of between 5/8 and 2 inches. j
'A'ssuming the 2-inch thick lid provides conservatism in the flaw tolerance calculations. Also assumed for this bounding calculation is an axial /circumferential
. flaw that extends completely around the circumference,'with a depth equal to half of the most conservative lid thickness of 2 inches. ' Postulate a remote tensile stress ;
of 30 ksi, approximately equal to the yield stress. For such a flaw,2c/W is 1, a/t is !'
0.5, and a/c is very small (0.013). The highest stress intensity is at the "a-tip" (i.e.,
at the depth of the crack), with F = 1.8 or somewhat higher. For conservatism, we will use F = 3. Then Ki is approximately 159.5 ksi Vin, and the applied J-integral is less than 1,000 in lb/in2, below the threshold for unstable crack extension.
Appendix'A, demonstrates that the J-integral must be computed from elastic-plastic
- fracture mechanics principles for a/t = 0.5 and 2c/W = 1, but that the approximate
-- calculation of J is valid for such a flaw with a/t = 0.3 and perhaps with alt = 0.4. j
! For purposes of this set of calculations, stainless steel welds are tolerant of flaws ]
with depth a/t'= 0.3 and 2c/W = 1.
This bounding calculation' demonstrates that for conservative stress fields caused
- by extreme' design loadings, such as handling accidents, the fabrication flaw that must be detected extends completely'around the circumference of the lid and 30% to
. 40% through-wall. Such a flaw would he' easily detected and sized by current UT 7
~
October 9, 1998 examination techniques, and could be readily detected by either root-and-final or multiple-pas's PT methods, as well.
. Acceptance Flaw Calculations: The postulated flaw in the bounding calculation was selected to be extreme in order to demonstrate the ample flaw tolerance of the niaterial. However, the results for more meaningful defects are needed in order to
- establish flaw accepiance criteria.
For example, consider the same conservatively-thick closure weld, but in this case assume a through-wall flaw (a/t = 1) that hai 9: = 0.2. For a 2-inch thick lid, such -
a flaw would extend 20 inches (2c) around the circumference of the lid, so that 2c/W is approximately 0.07. Such a flaw would have a surface area of approximately 30 in2 For this case, the "a-tip" stress intensity factor F (1.98)is higher ths.n the :-
tip" value (1.50), which will give an applied stress intensity of 148.9 ksiVin and an applied J-integral of about 837 in-lb/in2 for a uniform tensile stress equal to the minimum yield strength. This applied J is of the same order as the value calculated for 2c/W = 1 and a/t = 0.3, and remains much lower than the threshold value of 1430 in-lb/in2 For the case of a thin closure weld (assume t = 0.75 inches), a through-wall flaw (a/t
= 1) with a planar area of 30 square inches would extend approximately 50 inches (2c) around the closure lid circumference, so that 2c/W = 1/6. Again, the a-tip stress intensity controls. For a/c = 0.03 (approaching zero), the a-tip stress intensity approaches 3.0, so that the applied J is between 725 and 750 in lb/in2. The margin against crack instability is 2, which is consistent with the thick-weld case. The
. effect of a more realistic bending stress would be to increase this margin further.
Two other flaws of similar flaw area to the above (30 in2) are examined in order to evaluate the effect of flaw geometry (depth and length) on the J value for a lid thickness of either % or 2 inches. Since the two thicknesses were shown above to give comparable margins, we concentrate on the %-inch thickness for the two demonstration cases: (1) a semi-elliptical flaw extending halfway through the thickness (a/t = G.3) and 100 inches around the circumference (a/c = 0.007, with 2c/W = 0.33 ); and (2) a semi-elliptical flaw extending 807c across the thickness (a/t
= 0.8) and 64 inches around the circumference (a/c = 0.02, with 2c/W = 0.13 to 0.21). The stress intensity factors for these two postulated flaws subject to uniform radial tensile loading are greatest for the a tip, and have values somewhat higher than the published values of 1.45 to 1,76 for a/c = 0.2. For purposes of these calculations, the values of F will be increased to 2.5. Then, J is 250 in lb/in2 for the i . first flaw and 400 in-lb/in2 for the deeper flaw, both well below the limiting value of 1430 in lb/in2 For the case of a 2 -inch thick lid, the J values are 425 in-lb/in2 and
[
'1070 in lb/in2 respectively for the same relative-size flaws. If the applied stresses are bending in nature, the a tip and c tip stress intensity factors are approximately 8
p October 9, 1998 equal and less than half as large as those for uniform tensile stress. In such cases, assuming maximum bending stress amplitudes equal to twice the yield stress will
- produce slightly smaller J values, again well below the limiting value of 1430 in-lb/in2 CONCLUSIONS:
. Industry concludes:
(1) Austenitic stainless steel lid-to-shell welds in spent fuel dry cask can tolerate very large fabrication defects for stresses from extreme design-basis loads, radial tensile stresses at minimum yield stress levels or more realistic bending stresses at twice minimum yield stress levels.
(2) The likelihood of producing such large flaws during canister fabrication is negligible.
' (3) Unstable crack growth is resisted in stainless steel closure welds even when assuming unrealistic postulated defects (flaws that extend completely around the weld circumference and 307c to 40% across the weld thickness).
- (4) Substantial margins against unstable flaw growth are available even for more realistic weld defects, such as flaws on the order of 30 in2 in area that extend 50% to 80% across the weld thickness and 20 to 100 inches around the weld
' circumference.
(5) Substantial margin is available even for a weld defect that extends completely acreas the weld (through wall) and of the order of 20 to 50 inches around the weld circumference.
(6) Any of these flaws wo'uld be readily detected by conventional UT examination procedures and by either root / final or multiple pass PT examination methods.
For the case of multiple-pass PT examination, postulated weld defects that extend 507c to 100% across the weld would be detected by a combination of mid-pass and final pass PT. For the case of root / final pass PT, a weld efficiency factor of the order of 75% provides an equivalent level of assurance.
RECOMMENDED ACCEfFrANCE CRITERIA:
L I.
L The recommended acceptance criterion is a flaw with an equivalent area of 30 in2 This corresponds, for example, to an elliptically shaped through-wall 2-inch-deep 9
r-9 1
' October 9, 1999 flaw extending 20 inches around the weld circumference or, for a thinner weld, a %-
-inch-deep flaw extending 50 inches around the weld circumfsrence.
s 4
v 10 1
4
October 9, 1998 DELAYED HYDROGEN CRACKING (DHC) IN STAlNLESS STEEL WELDMENTS i
I l
Two necessary conditions are needed for DHC to occur, which are. (1)The presence of a sufficient amount of hydrogen in solid solution. This would depend on the !
J source of hydrogen, the temperature, and the hydrogen solubility limit. (2)
Sufficient hydrogen mobility by diffusion to the crack tip. This requires a tensile stress field or lower crack-tip temperature than the bulk temperature. As the hydrogen in solution diffuses to the crack tip, it will remain in solution until the local hydrogen concentration begins to exceed the solubility limit. It will then precipitate in the form of solid hydrides. When the hydride reaches a critical size, a few microns, it will fracture, allowing the crack to extend by that amount. The process will be repeated and the crack will extend discontinuously in time in ,
increments of a few microns at a time. The crack velocity depends on the !
temperature - the higher the temperature, the larger the velocity. ]
The scenario described above cannot exist in stainless steel weldments in a spent fuel canister cover for the following reasons: (a) the hydrogen in the material is an impurity with a concentration level of 10 to 15 ppm, which is too low to sustain the formation of damaging hydrides; and (b) at the canister-lid storage temperature, the hydrogen solubility limit is nearly zero ppm; consequently, the 10 to 15 ppm impurity hydrogen is totally immobilized and no difrusian to the crack tip can take place. This picture does not change during storage, as no corasion mechanisms or ;
environmental conditions exist that could raise the hydrogen levels in the weld or base metal above the impurity level.
The above discussion is confirmed by experimental data from the Handbook of Stainless Steels, by Becker and Bernstein, which reports that sustained-load crack growth experiments on Type 321 stainless steel showed little or no effect of 34.5 MPa hydrogen even when 25-mm (1-inch) thick specimens were loaded at the unstable fracture-stress intensity, 35 MPa6 (32 ksiE); all crack extension observed was in the fbrm of blunting. The same behavior was observed for Type 304L, tested under the same hydrogen pressure conditions, even when the ,
specimens were in gross yielding. These results are from M. R. Louthan, Effects of j Hydrogen on the Mechanical Properties of Low Carbon and Austenitic Steels,
" Hydrogen in Metals," pp. 53-75, American Society for Metals, Metals Park, Ohio, 1974.
11 ;
1 l
s
I October 9, 1998 REFERENCES
[1] Nickell, R. E. and Rinckell M. A.," Evaluation of Thermal Aging Embrittlement for Cast Austenitic Stainless Steel Components in LWR Reactor Coolant Systems," EPRI TR-106092, EPRI, Palo Alto, CA, September 1997.
[2] Mills, W. J. " Fracture Toughness of Stainless Steel Welds," in: Fracture Mechanics: Nineteenth Symposium, ASTM STP 969, T. A. Cruse, Editor, American Society for Testing and Materials, Philadelphia, PA,1986, pp. 330-355.
[3] Raju, I. S., Mettu, S. R., and Shivakumar, V., " Stress Intensity Factor Solutions for Surface Cracks in Flat Plates Subjected to Nonuniform Stresses,"in:
Fracture Mechanics: Twenty-Fourth Symposium, John D. Landes, Donald E.
McCabe, and J. A. M. Boulet, Editors, American Society for Testing and Materials, Philadelphia, PA,1994, pp. 560-580.
12 ip-
October 9, 1998 APPENDIX A ELASTIC PLASTIC FRACTURE TOUGHNESS ESTIMATION Elastic Plastic J. Estimation In order to c6nfirm the flaw stability calculations for deep flaws and/or relatively c high stresses on the remaining ligament, an elastic-plastic J-estimation approach
[A1] is applied to the case where the assumed flaw depth is equal to a/t = 0.3,0.4, and 0.5, for full circumferential flaw. In this particular case, the model shown in Figure 2 degenerates into a single-edge cracked plate (SECP) geometry subjected to a remote uniform tensile stress. The direct clastic-plastic solution is given by J4ppu.a = Jei i. + Jiu.n.. (A.1)
For this geometry Ja oc = n a F2 p2 / (E' t2), ( A.2) where a is the flaw depth, F is a geometric magnification factor, P is the applied tensile load per unit thickness, E' = E/(1-v2) with E the elastic modulus and v Poisson's ratio, and t is the plate thickness.
The plastic portion of the solution is given by Jn n = a ao co d (a/t) hi (a/t, n)(P / Po)*1 ( A.3 )
In this expression, d = t - a is the remaining ligament distance, Po is the limit load for that ligament per unit thickness, and hi is a geometric magnification factor. 4 The material parameters a, n, co, and to are derived from a Ramberg-Osgood deformation theory plasticity model for the material [A2), ,
I c / co = c / co + a (c / c,.)a. ( A.4 ) )
The uniaxial stress and strain are given by a and e, respectively.
The limit load in plane strain for the SECP geometry is given by P., = 1.455 q d c , ( A.5 )
where q = [1 - (a/d)2]v2 - a/d. ( A.6) !
l 1
~
[.
t .
4 1998 October 9, The Ainsworth Modification.
1 In order to simplify the calculations and eliminate errors caused by the Ramberg-Osgood approximation to the uniaxial stress strain curve, the EPRI J-estimation approach will be modified, following Ainsworth [A2], who suggested the introduction of a reference stress. This reference stress is defined by or r =- (P/P.,) c., (A.7) where o. is chosen to be the yield strength. Then, with the reference strain, cr.r, defined to be the to'.al strain on the uniaxial stress-strain curve corresponding to a r.r the plastic portion of the J-integral becomes Jn .. '= o,.r d hi (c,.c - or./E). ( A.8)
Equation (A.8) still retains a dependence on the Ramberg-Osgood approximation to the uniaxial stress-strain curve through the geometric shape function hi.
Ainsworth [A2] introduced an additional approximation that eliminates this dependence in favor of the expression for the elastic stress intensity factor, so that
.Jnou.= KP (E cr./or.r- 1)/E, (A.9) where p = 0.75 for plane stress and 1.0 for plane strain. This approximation introduces a dependcnce on the elastic stress intensity factor combined with the secant modulus into the expression for the p!astic crack driving force.
Limit Load Calculations.
For a/t = 0.3,0.4, and 0,5; a = 0.6,0.8, and 1.0 inches, respectisely. Then the remaining ligaments are d = 1.4,1.2, and 1.0 inches, respectively, and Equation (A.6) gives a = 0.6594,0.5352, and 0.4142, respectively. If co is chosen to be 30,000 psi, then Equation ( A.5) gives the limit loads as 40,296 lb/in for a/t = 0.3; 28,034 lb/in for a/t = 0.4; and 18,080 lb/in for a/t = 0.5. Selecting the yield strength to be the ASME Code minimum value reduces the calculated limit loads substantially. ;
. The corresponding values of or r are 22,335 psi for a/t = 0.3: 32,104 psi for a/t = 0.4; j and 49,779 psi for a/t = 0.5.
Crack-Driving Force Elastic Contributions.
2 ,
1 1
d 4
i
, October 9, 1998 From Equation (A.2), using the graphical representation from Buchalet and Bamford [A.3], the elastic contributions to the elastic-plastic crack driving force c're as follows: a/t = 0.3, Fi hmiform remote stress) = 1.5, Ki u 61.8 ksiVin, and Jsion, =
=
144 in-lb/in2; a/t = 0.4, Fi = 2.0, is = 95.1 ksiVin, and Jeon,= 341 in-lb/in2; and a/t =
0.5, Fi = 2.75, Ki = 106.3 ksiVin, and Jaoue = 426 in-lb/in2, Crack Driving Force Plastic Contributions.
For a/t = 0.3, the value of o,.c is below the yield strength, so that the plastic contribution to the crack-driving force is zero with this approach. However, for a/t ;
a 0.4, the reference stress is slightly above the ASME Code minimum yield strength, so that a small plastic contribution to the crack-driving force would be expected. For a/t'= 0.5, the reference stress is well above the yield strength, with a secant modulus that is of the order of 1.8 x 108 psi (from Reference 4, for annealed 304 stainless steel at room temperature, an engineering stress of 49,910 psi (compared to 49,779 psi in the calculation above) corresponds to an engineering strain of 0.0273; weld material would have a substantially higher engineering stress at lower engineering strain, but the annealed data is conservative). Equation (A.9) shows that the multiplier (Ec, dor r-1) on the elastic contribution to the crack driving force is of the order of 15; i.e., Jnone would be of the order of 6,000 in-lb/in2, Conclusions.
For postulated flaws that are fully circumferential and that extend across the weld
.to a epth d a/t = 0.3 to 0.4, the conversions from Kir gu.a e to J[ Applied using the expression J '= K2/E are sufficiently accurate. For flaw depths greater than a/t = 0.4, the fully-plastic contribution to the elastic plastic crack-driving force is substantial, reflecting the net section stress. For purposes of this discussion, a limit of a/t < 0.4 will apply for fully circumferential flaws.
3 1
\ q
l October 9, 1998 References t
iAl) V. Kumar, et al.,"An Engineering Approach for Elastic-Plastic Fracture,"
Report No. EPRI NP-1931, General Electric Company , Schenectady, New York (July 1981).
- [ A2] Ainsworth, R. A.,"The Assessment of Defects in Structures of Strain Hardening Materials," Engineering Fracture Mechanics, Volume 19, p. 633 (1984).
[A3] C. B. Buchalet and W. H. Bamford," Stress Intensity Factor Solutions for-Continuous Surface Flaws in Reactor Pressure Vessels," in: Mechanics of Crack Growth, ASTM STP 590, America i Society for Testing and Materials, i Philadelphia, pp. 385-402 (1976). l
[A4] J. B. Conway, R. H. Stentz, and J. T. Berling, " Fatigue,' Tensile, and Relaxation Behavior ofStainless Steels," Report No. TID 26135. U. S. Atomic .
Energy Commission, Washington, DC, (1975).
(
4
l
- I Weld Weld]
i ~2" Flawj l l
1 I
i
- .y l j
.l !
.I I
-l 1
~45" .
- l. l I
' t (a) Posmlated Flaw in Full Closure Weld ,
{
n a a o 4a4nA449 & ension T Loading = cy l l llll a a T, ^ Bending Loading 2cy
~~
A ,i &
l i ,
Y j Y '
ir p 3rt I
l l b .
~!
l j -
e l l l t M. ,
I l i- u H i
l Consider axial flaw i I w is closure plate circumference !
I t is weld / plate thickness l
I (2c)/w = 1.0. full circumferential flaw !
,l________
(b) Flaw Geomet.f n and Loading Figure 1. - Postulated Flaw Configuration
r, I
s(x) j' 1
, . . Six) l I
Il s .
1i ll Il lt I
'I %
i iY xx s s
I k
\s
- 9. 9
,m~~
J sW}! j
%w N t
! -w i ,
t: !
.. _La e-I 1 l
Figure 2 - Geometry and Coorcinate System of Surf ace Cracks (Raju. et al. - ASTM STP 1207)
I l
4 s EDRI Licensed Matenal ___
ktstenal Constoerations 500 500-400 l 400
,M .. ;
^. 300 300-s V 1
/^ .# o --
y f 4.cr-;v- - Y 7 /-C 'c/v.
200 200 g ,-
t./
Deformatia c.J.(k)/m:1). ,- .
.g/
--- 100 100 ,'d I I I e . g
.v.
I
-5 Typo 316 SAW
-c-OASS OF.au Stanc Cast FN 15%
- CASS CF 8M Staue Cast 1015% FN y
- -- 1
-w-CASS CF 8M Staue Cast FN41C*/.
f. '
! - -O 0- 4 ,
i i
i i 2 3 4 0 .1 3
C sex Extension (mm)
Figure 24 Companson ol Crack Growth Resistance U-R) Cu.wesfor Type 116 SAW ':nd Stan::.lly-C::: CT b'M Material ar !!S*F.
2 30-
'*:CO 9E-MT.Sar i
f .
. .. )
l i
- 3. '- .
i-ESRI Lleensed Matenal
~
as 500 __
l l.
I A
P 400
_/ OO fl 0
/ 0
[
ff .
A d '
y ,/> -
200
- l,.
.V ,
Deformatiq1J Kj/rr2 ., ..
b ,.'
100--
fI
,I s. ;. -
g'
-Y-Type 315 SAW
'y -O-CASS CF 8M Centr. Cast FN>15%
-O-OASS OF BM Centt,.0ast 1015% FN
' -d-CASS CF.8M Osntr. Cast FNc1C%
=
~
0 -
- i 2 3 4 0 1 C.ack Extension (mm) l Figure 23. Comparison of Crack Growth Resistance U R) Curvesfor 1ype 316 SAW and Cenmfutally Caxt CF 8M Matertal at !!S"F.
- . n i
.-4 c- %, ,.,e,