ML20137X383
| ML20137X383 | |
| Person / Time | |
|---|---|
| Issue date: | 10/31/1985 |
| From: | Serkiz A Office of Nuclear Reactor Regulation |
| To: | |
| References | |
| REF-GTECI-A-43, REF-GTECI-ES, TASK-A-43, TASK-OR NUREG-0869, NUREG-0869-R01, NUREG-869, NUREG-869-R1, NUDOCS 8512100503 | |
| Download: ML20137X383 (130) | |
Text
,
e' b
NUREG-0869 Rev.1 USI A-43 Regulatory Analysis O Regulatory Analysis for USl A-43,
" Containment Emergency Sump Performance" O Proposed Resolution O Summary of Public Comments Received and Action Taken O CRGR Minutes (Ref. USI A-43)
O Draft Generic Letter 1
Office of Nuclear Reactor Regulation A. W. Serkiz, Task Manager N(hYk.,$
8512100503 851031 PDR NUREG PDR 0869 R
=
4 4
NOTICE Availability of Reference Materials Cited in NRC Publications Most documents cited in NRC publications will be available from one of the following sources:
- 1. The NRC Public Document Room,1717 H Street, N.W.
Washington, DC 20555
- 2. The Superintendent of Documents, U.S. Government Printing Office, Post Office Box 37082, Washington, DC 20013-7082
- 3. The National Technical Information Service, Springfield, VA 22161 Although the listing that follows represents the majority of documents cited in NRC publications, it is not intended to be exhaustive.
Referenced documents available for inspection and copying for a fee from the NRC Public Docu-ment Room include NRC correspondence and internal NRC memoranda: NRC Office of Inspection and Enforcement bulletins, circulars, information notices, inspection and investigation notices; Licensee Event Reports; vendor reports and correspondence; Commission papers; and applicant and licensee documents and correspondence.
1 The following documents in the.NUREG series are available for purchase from the GPO Sales Program: formal NRC staff and contractor reports, NRC-sponsored conference proceedings, and
' NRC booklets and brochures. Also avaliable are Regulatory Guides, NRC regulations in the Code of Federal thquistions, and Nuclear Regulatory Commission issuances.
Documents available from the National Technical Information Service include NUREG series reports and technical reports prepared by other federal agencies and reports prepared by the Atomic Energy Commission, forerunner agency to the Nuclear Regulatory Commission.
Documents available from public and special technical libraries include all open literature items, such as books, journal and periodical articles, and transactions. Federal Register notices, federal and state legislation, and congressional reports can usually be obtained from these libraries.
Documents such as theses, dissertations, foreign reports and translations, and non-NRC conference l
proceedings are available for purchase from the organization sponsoring the publication cited.
Single copies of NRC draft reports are available free,to the extent of supply, upon written request to the Division of Technical Information and Document Control, U.S. Nuclear Regulatory Com-l mission, Washington, DC 20555.
Copies of industry codes and standards used in a substantive manner in the NRC regulatory process are maintained at the NRC Library,7920 Norfolk Avenue, Bethesda, Maryland, and are available there for reference use by the public. Codes and standards are usually copyrighted and may be purchased from the originating organization or, if they are American National Standards, from the American National Standards Institute,1430 Broadway, New York, NY 10018.
L l
NUREG-0869 Rev.1 USI A-43 Regulatory Analysis O Regulatory Analysis for USl A43,
" Containment Emergency Sump Performance" O Proposed Resolution O Summary of Public Comments Received and Action Taken O CRGR Minutes (Ref. USl A-43) o Draft Generic Letter Manuscript Completed: June 1985 Date Published: October 1985 A. W. Serkiz, Task Manager Office of Nuclear Reactor Regulation U.S. Nuclear Regulatory Commission Washington, D.C. 20555
,e=~=m, r
i
)
- \\,,,,,/
i
1 ABSTRACT This report consists of:
(1) the regulatory analysis for Unresolved Safety Issue (USI) A-43, " Containment Emergency Sump Performance"; (2) the proposed.
resolution; (3)-a summary of public comments received and action taken; (4) the Committee to Review Generic Requirements (CRGR) minutes related to this USI; and (5) appendices that summarize assumptions, calculational methods, conse-quence analyses, and cost estimates used.in this regulatory analysis.
1 4
i l
l 4
i 4
NUREG-0869, Revision 1 111 October 1985
CONTENTS P,.aget ABSTRACT..............................................................
iii FOREWORD..............................................................
ix 1
STATEMENT OF PROBLEM.............................................
1 1.1 S umma ry o f Sa fe ty I s s ue.................................
1 1.2 Technical Findings..........................................
1 2
OBJECTIVES.......................................................
2 3
A LT E RN AT I V E S.....................................................
3 4
CONSEQUENCES.....................................................
4 4.1 Estimated Consequences Associated with Different Containment Types...........................................
7 4.1.1 PWR D ry Co n ta i nme n t s.................................
7 4.1.2 PWR Subatmospheric Containments......................
9 4.1.3 PWR Ice Condenser Plants.............................
9 4.1.4 BWRs with Mark I and Mark II Containments............
10 4.1.5 BWRs with Mark III Containments......................
11 4.2 Estimated Occupational Exposure.............................
11 4.3 Impact on NRC Operations....................................
12 4.4 Impact on Other Government Agencies.........................
13 4.5 Public Impact...............................................
13 4.6 Other Constraints...........................................
13 5
DECISION RATIONALE...............................................
13 5.1 Comparison of Regulatory Alternatives.......................
13
- 5. 2 Rationale for Selecting the Recommended Resolution..........
14 5.2.1 Options 1 and 2...................................
14 5.2.2 Option 3...........................................
18 5.2.3 Option 4............................................
18 5.3 Recommended Regulatory Action...............................
18 6
PLAN FOR IMPLEMENTATION..........................................
19 7
STATUTORY CONSIDERATIONS.........................................
20 7.1 NRC Authority...............................................
20 7.2 Need for NEPA Statement........................
20 8
BIBLIOGRAPHY.....................................................
20 NUREG-0869, Revision 1 v
October 1985
CONTENTS (continued)
Appendix A-
SUMMARY
OF PUBLIC COMMENTS RECEIVED AND ACTION TAKEN Appendix B BACKGROUND AND
SUMMARY
OF MINUTES OF MEETINGS OF CRGR REGARDING USI A-43 RESOLUTION Appendix C ESTIMATION OF PIPING FAILURE PROBABILITY Appendix D ESTIMATION OF PWR SUMP FAILURE PROBABILITY Appendix E CONSEQUENCES OF LOSS OF RECIRCULATION CAPABILITY
. Appendix F CONTAINMENT SURVIVABILITY Anpendix G ESTIMATION OF COSTS TO REPLACE INSULATION Appendix H DRAFT GENERIC LETTER L
l NUREG-0869, Revision 1 vi October 1985 y
y 3-p
+-9+
3y--y-p-ver,w-w-w--
s%
m-.---.y--
--++ - - --
w--
y-+
-e
--.--m9-.--.~6 y
-~ -.
i LIST OF TABLES Page 4.1 Assessment of sump blockage probability..........................
6 4.2 Types of nuclear plant containments and their license review status...........................................................
7 4.3 Radiation exposure to workers during insulation replacement (in person-rems).....................................................
11 4.4 Radiological effects of backfitting..............................
12 5.1 Summary of calculated values and impacts associated with various containment designs for resolution of USI A-43.....................
15 d
I i
i NUREG-0869, Revision 1 vit October 1985
f~~
-in v-.
-,a
+~m,-
~_,,,,n_,
"'*smA ga_
' -- A 4 < A & g g t
y=
_p.
k 9
I e -'_.,.
h T
y I
t-r
- f. d I;.
f 1
(i A
4 t
4.
t S
5 a
i t.
k t
l-A II.
t i<
s t
i 1
I i
1 t
- I t_
l I.
I 1
s-f l
e i
F 6
b l
t t
F f
t
'v-wo.
emwm
_...mmm-,-e_._.
aw w n
=
FOREWORD This report has been prepared to provide a self-contained reference that in-cludes the regulatory analysis for Unresolved Safety Issue A-43, " Containment Emergency Sump Performance"; the proposed resolution; public comments received on the for comment issuance of NUREG-0869 and actions taken in response; minutes of meetings of the Committee to Review Generic Requirements regarding this issue; and summaries of assumptions, calculational methods, consequence ana-lyses, and cost data utilized.
I This report was originally issued cs NUREG-0869 for public comment in May 1983.
Comments received were reviewed, and those of substantive technical or informa-tional content were incorporated into this revision.
These are summarized in Appendix A.
For historical purposes, it should be noted that two draft versions of this report--NUREG-0869, Revision IA, May 1984, and NUREG-0869, Revision 18, July 1585--were prepared for use in the predecisional review process.
This formal issuance--NUREG-0869, Revision 1, October 1985--is the concluding regulatory analysis applicable to the resolution of Unresolved Safety Issue A-43.
It should also be clearly noted that this report is not a s0bstitute for re-quirements set forth in General Design Criteria 16, 35, 36, 38, 40, and 50 in Appendix A of Title 10 of the Code of Federal Regulations Part 50, nor is it a substitute for guidelines set forth in NRC's Standard Review Plan (SRP, NUREG-0800), regulatory guides, or other regulatory directives.
The draft generic letter in Appendix H is for completeness of record and will be imple-mented through a normal regulatory issuance.
NUREG-0869, Revision 1 ix October 1985
1 l
l I
i N
' l i
I i
e 9
s i-f-
1 m
t k
e I
a L
i i
4
~* '" ' * -
w W ? 'M'M**W'Fw-erw " w m m
_sm,_4,,.
J
^
REGULATORY ANALYSIS FOR USI A-43,
" CONTAINMENT EMERGENCY SUMP PERFORMANCE" 1.
STATEMENT OF. PROBLEM 1.1. Summmary of Safety Issue Unresolved Safety Issue (USI) A-43 deals with the concerns about the availabil-f
-ity of adequate recirculation cooling water following a loss-of-coolant acci-dent (LOCA) when long-term recirculation cooling from the containment sump in-a pressurized water reactor (PWR) or the residual heat removal (RHR) system suction intake in a boiling water reaction (BWR) must be initiated and main-tained to prevent core melt.
These safety concerns can be summarized as follows:
r.
In the recirculation cooling mode, will the sump design (PWRs) or the RHR' suction intakes (BWRs) provide sufficient water to the RHR and containment spray system (CSS) pumps, and will this water be suffi-
+
ciently free of LOCA generated debris and ingested air so that pump 1.
performance is not impaired to the point'of seriously degrading long-term recirculation flow capability?
The USI A-43 safety concerns can be separated into three parts:
(1) the effects of potential air ingestion and elevated temperatures and break flow on sump (or suction intake) hydraulic performance under post-a LOCA adverse conditions (2)' the effects of LOCA generated insulation debris (resulting from a pipe break jet that destroys-large quantities of insulation) that is trans-l
' ported to the sump debris screen (s) and blocks the sump screen (or suc-
-tion strainer), reducing NPSH margin below that required for the recircu-lation pumps.to maintain long-term cooling i
(3) the effects of air or debris ingestion or other problems (such as the effects of particulate ingestion on pump seal and bearing systems) on the
{
capability of RHR and CSS pumps to continue to function Although USI A-43 was derived principally from concerns about PWR containment emergency sump performance, the concern about debris blockage applies to BWRs i
as well. The RHR suction strainers in a BWR are analogous to the PWR sump debris screen, and both BWRs and PWRs must have adequato recirculation cooling capacity to prevent core melt.
f
- 1. 2 Technical Findings l
The staff investigated these safety concerns on a generic basis, and reported its technical findings in NUREG-0897, Revision 1, October 1985.
These find-ings can be summarized as follows:
NUREG-0869, Revision 1 1
October 1985
(1) Extensive, full-scale sump hydraulic tests generally show that low levels of air ingestion (less than 1% to 2%) will occur.
These tests also demon-strate that vortex observations alone cannot be used to quantify levels of air ingestion (as has been done in the past). The test results have been used to develop PWR sump and BWR suction intake hydraulic design guidelines for minimizing, or eliminating, air ingestion and have eliminated the need for plant-specific sump tests or model tests.
(2) Plant insulation surveys, development of methods for estimating debris generation and transport, debris transport experiments, and information received in public comments show that the effects of debris blockage de-pend on the types and quantities of insulation employed, the layout of the primary system within containment, post-LOCA recirculation patterns and velocities, and post-LOCA recirculation flow rates.
Thus, the staff concluded that a single generic solution is not possible, but rather that debris blockage effects are governed by plant-specific design features and post-LOCA recirculation flow requirements.
The test results also show that the 50% screen blockage criterion in Regulatory Guide (RG) 1.82, " Sumps for Emergency Core Cooling and Con-tainment Spray Systems," should be replaced with a requirement that debris blockage effects be assessed on a plant-specific basis.
The 50%
screen blockage criterion does not require a plant-specific evaluation of the debris-blockage potential and may result in a nonconservative analysis.
(3) Reviews of available data on the effects of air ingestion on pumps and discussions with the U.S. manufacturers of RHR and CSS pumps show (a) that low levels of air ingestion (2% or less) will not significantly degrade pump performance and (b) that the types of pumps employed in nuclear plants will tolerate ingestion of the insulation debris and other types of post-LOCA particulates that can pass through PWR sump screens or BWR suction strainers.
In summary, these findings show (1) that the potential safety problems that could result from vortex formation and air ingestion are less than previously hypothesized, and (2) that the potential for a loss of recirculation cooling capability as a result of LOCA generated debris blocking a screen is potentially a more significant concern.
These technical findings were in NUREG-0897, which was issued for comment in May 1983.
Information received during the public comment period included tech-nical data, plant-specific data, cost information, ard cost impact data.
Apen-dix A summarizes the comments received and how the staff addressed them.
Those comments that were principally technical in nature were incorporated into NUREG-0897, Revision 1, October 1985.
Plant-specific design information and backfit cost data received during the comment period were used in preparing this report. Appendix B contains minutes of meetings of the Committee to Review Generic Requirements on USI A-43.
2 OBJECTIVES The general objective of the proposed regulatory actions discussed below is to provide assurance that the safety concerns associated with USI A-43 will be NUREG-0869, Revision 1 2
October 1985
adequately addressed in the licensing process.
The goal is to meet the re-quirements of General Design Criterion (GDC) 35, " Emergency Core Cooling," and GDC 38, " Containment Heat Removal" (in Appendix A to Title 10 of the Code of Federal Regulations Part 50).
The technical findings developed as the result of the work performed to resolve USI A-43 have established a need to revise current licensing guidance en these matters.
Therefore, the staff's technical findings (NUREG-0897, Revision 1, October 1985) have been used to revise RG 1.82, Revision 0, and Standard Review Plan (SRP, NUREG-0800) Section 6.2.2, " Containment Heat Removal Systems." The issuance of these revised regulatory documents would change staff review prac-tices in light of current technical findings.
The issuance and need for imple-mentation of the revised regulatory guide and SRP section are discussed in Section 3 below.
3 ALTERNATIVES The following approaches were considered as alternatives to resolve USI A-43:
(1) Issue RG 1.82, Revision 1 and SRP Section 6.2.2, Revision 4, and require all licensees and applicants to evaluate the potential effects of dabris blockage (per RG 1.82, Revision 1) for_ confirmation of adequate n.
posi-tive suction head (NPSH) margin.
(2)
Issue RG 1.82, Revision 1, and SRP Section 6.2.2, Revision 4, and require an evaluation of the potential effects of debris blockage (per RG 1.82, Revision 1) for confirmation of adequate NPSH margin from only those licensees of and applicants for plants where loss of recirculation could lead to core melt and loss of ccntainment integrity.
The risk from off-site radioactive exposure depends both on the probability of core melt from loss of recirculation and on the effectiveness of the containment in mitigating offsite exposure.
If a licensee / applicant can demonstrate that the containment integrity of a plant will be maintained even if the core should melt because of a loss of recirculation, that licensee / applicant could be exempt from a requirement to evaluate debris blockage effects because public risk would be low.
This alternative involves the staff's determinating-on the basis of value-impact analyses for different types of containments (e.g., PWR dry con-tainments, ice condenser plants, Mark Is and IIs and Mark IIIs)--whether the benefits outweigh the cost impacts.
This approach is used to assess both Options (1) and (2) and is discussed in Section 4.
(3)
Issue a generic letter for information only to all licensees and applicants describing the potential safety concerns associated with insulation debris blockage of the sump screen, which could lead to the loss of adequate NPSH margin during recirculation.
Make clear that the 50% blockage criterion provided in RG 1.02, Revision 0, is not a necessarily conservative way to perform an asseament of debris blockage, particularly if fibrous insula-tion is used on t*.o primary system piping and components. Provide (with this generic letter) NUREG-0897, Revision 1, October 1985, which contains the staff's techy cal findings, for information.
State clearly that there is no requirement for analysis or modification for any operating plant or NUREG-0869, Revision 1 3
October 1985
plant now under construction.
Issue RG 1.82, Revision 1, and SRP Sec-tion 6.2.2 Revision 4, for implementation on Standard Plant designs and new construction permit (CP) applications only, to be effective 6 months from date of issuance of those documents.
In summary, under this alternative the significant safety information derived under the A-43 program would be provided to all licensees and ap-plicants but there would be no requirement for any action from licensees of operating plants or applicants for plants now under construction.
Standard Plant or new CP applicants would be required to address RG 1.82, Revision 1, and SRP Section 6.2.2, Revision 4, within 6 months after the date they are issued.
(4) Make no revision to RG 1.82 or SRP Section 6.2.2; publish NUREG-0897, Revision 1 (the staff's technical findings for USI A-43) as an informa-tion only document.
The need for the proposed actions can be summarized as follows:
(1) Issuance of NUREG-0897, Revision 1 (the staff's technical findings) will provide a comprehensive description of the technical issues, along with an extensive data base for designing and assessing PWR sump designs and BWR RHR suction inlets.
These findings (which show that vortex observa-tions do not quantify air ingestion) will replace the assumptions that lead to previously required inplant tests (or model tests); they also provide a common technical data base for licensees, applicants, and the staff to use, thereby reducing the regulatory burden in future assessments.
(2) Revising RG 1.82 will bring that guide into conformance with more than 3 years of experiments and generic studies related to sump design and performance, and will remove the 50% blockage criterion, which is an arbitrary assumption and not necessarily conservative from the viewpoint of sump debris blockage effects.
(3) Revising SRP Section 6.2.2 will make the review considerations consistent with the USI A-43 technical findings and with RG 1.82, Revision 1.
4 CONSEQUENCES This section assesses the consequences (values versus impacts) of each of the alternatives given in Section 3 above. Alternatives I and 2--which would re-quire licensees of or applicants for all (Alternative 1) or selected plants (Alternative 2) to perform analyses of potential debris blockage to determine if recirculation capability might be lost, and to undertake necessary plant modifications to reduce such potential risks--are the subject of this section.
Alternative 3--issaance of a generic information letter and implementation of revised licensing criteria in future reviews--would not impact current licen-sees or applicants.
The impact (cost) of including the revised licensing cri-teria in new designs is considered to be very small, and the resulting value/
impact would be favorable; therefore no detailed quantitative analysis is deemed to be necessary.
Alternative 4--do nothing--does not involve any impact.
Because of the difficulty of treating all reactors as a homogeneous group (both because of r. heir design differences with respect to the sump and type of in-sulation used, as well as because of the differences in the capability of the i
October 1985 NUREG-0869, Revision 1 4
4 containment to survive), the staff has developed averted risk.and value/ impact analyses for each of the five major types of plant containments. These are PWRs with large, dry containments; PWRs with subatmospheric containments; PWRs with ice condensers; BWRs with Mark I and II containments; and BWRs with Mark III containments.
For many plants, the staff expects that evaluations (per RG 1.82, Revision 1) would show that adequate NPSH would exist, despite the poten-tial for debris blockage.
For others, a plant-specific value/ impact analysis would not support imposing backfit requirements.
Without performing individual plant assessments, the staff cannot determine the number of plants in these two categories.
Therefore, rather than attempting to develop a value/ impact (V/I) assessment for the total plant population, the staff developed estimates of releases and value/ impacts for each plant and associated containment type, assuming that a significant probability of sump blockage exists.
For each such class, the staff considered:
(1) the potential reduction in core melt frequency, (2) the potential reduction in public risk if backfit is required (estimated averted releases in person-rem for the remaining plant life time),
(3) the costs of such backfits, and (4) the resulting V/I ratio of such modi-fications.
The results of these analyses are given below.
Blockage of the PWR containment emergency sumps or of the BWR RHR suction in-takes during the recirculation phase following a LOCA can lead to core melt and containment overpressurization unless alternate recirculation cooling water sources can be made available.
Core melt accompanied by containment failure l
can lead to a release of radioactivity and public radiation exposure.
The probability of these effects stems from (1) pipe failure (LOCA) probability i
(2) the probability of a sump blockage that would lead to loss of NPSH margin and a loss of recirculation capability (3) the probability that the containment structural can survive over-pressurization The A-43 pipe break probabilities were calculated from a 1977 data base that included piping failures of all types known at that time, including materials not used in nuclear plant piping (see Appendix C).
The estimated probabilities were 3 x 10-8 Rx/yr for large pipes (> 28 inches) to 3 x 10-4 Rx/yr for small 3
j pipes (2 to 6 inches).
The more recent experimental and analytical work, which is based on mechanistic fracture mechanics,'results in probabilities of the rupture of large-size ductile piping (unaffected by IGSCC) significantly lower (better by several decades in magnitude) than those employed in the A-43 ana-lyses.
Therefore, if pipe failure probabilities are extremely low--because of such considerations as leak-before-break, etc.--these calculations would result in very low estimated releases, and backfits would not be supportable on the l
basis of value/ impact criteria.
I The probability that the sump would be blocked depends on the amount and type of debris that could be generated by a LOCA.
The amount and type of LOCA-generated debris, in turn, depends on the size and location of the break, the layout of the containment, sump location and design (size of the debris screens),
~
recirculation flow requirements, available NPSH margins, and type of insulation employed.
Because these vary greatly from plant to plant, estimating sump j
I i
NUREG-0869, Revision 1 5
October 1985
blockage probabilities becomes highly plant-specific and arriving at a single, or generic, value is not possible.
Table 4.1 shows the factors needed to esti-mate sump blockage probabilities.
Appendix D provides a more detailed discus-sion and gives examples cf the calculations used to develop the sump blockage probabilities.
Table 4.1 Assessment of sump blockage probability Event Technical consideration (s)
Safety implication Pipe break Break probability If break probability is ex-Break size tremely low, debris blockage B eak type potential is negligble.
Debris Break size and location Small pipes (< 10-inch diam-generated Target (s) location eter) generate small amounts Type (s) of insulation of debris; therefore debris Extent of jet damage (L/D) blockage effects produced by small pipes are not significant.
Transport Break location If U < 0.2 ft/sec, transport of debris Plant layout and sump location is not likely to occur; Type of debris therefore blockage would not 4
Recirculation velocities (U) occur.
Debris Amount transported Fibrous insulation debris screen Available screen area transports and coats total blockage Type of blockage screen area.
If debris screen areas are large, pressure drop is minimized.
NPSH impact Type of blockage If AHB > NPSHA, loss of hP*
recirculation can occur.
r to flow required Blockage head loss (AH )
g The survival of the containment structure and the maintenance of containment integrity also are key factors in determining the potential consequences of sump blockage (see Appendices E and F).
Even though core melt can be postu-lated to result from loss of recirculation cooling flow, the survival of the containment structure and the maintenance of containment integrity (or the ability to withstand an overpressure transient) would significantly limit the levels of radioactive release.
Recent studies dealing with containtrent struc-tural capabilities have shown that overpressure design margins do exist; there-fore containment integrity can be maintained through structural overpressure design margins, alternate containment cooling means, or controlled venting (as is done in BWRs).
However, the variability in the containment design (e.g.,
f NUREG-0869, Revision 1 6
October 1985
PWR dry containment versus PWR ice condenser design, BWR Mark I versus Mark III design) precludes a single conclusion.
Table 4.2 summarizes the number and types of containments in use.
Table 4.2 Types of nuclear plant containments and their license review status Type of containment Number of plants OLs issued!
PWR dry w/SGFC2 57 51 PWR dry w/o SGFC 14 PWR subatmospheric 7
5 PWR ice condenser 10 7
PWR TOTAL 88 63 BWR Mark I 23 21 BWR Mark II 10 7
BWR Mark III 8
1 BWR TOTAL 41 29 INDUSTRY TOTAL 129 92 1As of December 1984.
2SGFC = safety grade fan cooler.
4.1 Estimated Consequences Associated with Different Containment Types 4.1.1 PWR Dry Containments The PWR dry containment design concept is used extensively in U. S. nuclear power plants.
Currently 71 PWR dry containment plants are licensed or in final license review (of approximately 125 plants docketed).
Fifty-seven of these 71 plants use safety grade fan coolers (SGFCs) to help control LOCA containment pressures and temperatures.
As noted above, pipe break probability is the first key factor to consider in evaluating sump blockage probability because the LOCA is the debris generator.
A range of estimated pipe failure probabilities was developed for USI A-43; these are discussed in Appendix C.
The estimated probabilities were 3 x 10-8/
Rx yr for large pipes (> 28 inches) to 3 x 10-4/Rx yr for small pipes (2 to 6 inches).
These estimated pipe break frequencies do not include more recent leak-before-break considerations.
NUREG-0869 Revision 1 7
October 1985
Salem Unit 1 was used as a reference PWR to modeI piping layouts, weld loca-tions, break locations, insulation distribution, and major insulated plant com-Because of the variability of plant and sump designs and operational ponents.
requirements, the following design specifications were analyzed parametrically:
i Recirculation Flow = 6,000 to 10,000 gpm Available Debris Screen Area = 50 to 200 ft2 Available NPSH Margin (w/o blockage) = 1 to 5 ft H O 2
These design ranges are representative of 19 PWR designs for which the staff has detailed information.
The calculational methods and results are reported in NUREG/CR-3394.
The principal purpose of calculating sump failure probability is to estimate core melt frequency.
For the range of parameters given above, the estimated PWR sump failure probabilities (see Appendix D) ranged from 3 x 10-8 to 5 i 10'5/Rx yr.
Assuming that sump failure leads to core melt (which is the result if other actions cannot be or are not taken), core melt frequencies from a blocked sump would be 3 x 10-8 te 5 x 10-5/Rx yr.
Two signifiCAnt points to be noted are: (1) these sump failure probabilities were based on the assumption that all fibrous debris was transported to the sump and led to blockage (this would not be the case for PWRs that have recirculation velocities less than 0.2 f t/sec, which many large dry containments are known to have), and (2) no credit was given for detection of blockage buildup and operator corrective actions in these sump blockage estimate!..
To estimate the frequency of core melt resulting from loss of NPSH due to sump blockage, it was assumed that in 50% of the cases this loss would lead to core melt.
For the remaining cases it was assumed that the operator would detect the onset of blockage and take action to maintain recirculation flow. Thus the estimated core melt frequency is 1.5 x 10-8 to 2.5 x 10-5/Rx yr for this type of plant.
As noted above, the second approach to assessing the safety significance of this USI is based on estimating the ef fects on the public of potential releases of radioactivity associated with sump failure (see Appendices E and F).
The estimated conditional consequences from core melt for PWRs with large dry 5
containments without safety grade fan coolers are estimated to be 5 x 10 Using the estimated crire melt frequency of 1.5 to 25 x 10-8 and person-rem.
an outstanding reactor life span of 25 years results in an estimated averted risk range of 20 to 300 person-rem /Rx.
These aro low levels of averted risk and indicate %e safety issue to be of moderate-to-low significance.
The value/ impact (V/I) ratio range that can be calculated for PWRs with large dry containments (based on an estimated cost of $1.5M/Rx for replacing in-sulation) is 10 to 200 person-re:m/$million.
If less severe backfit actions are required (at an estimated cost of $0.4M/Rx), the V/I range is 50 to 800 person-rem /$ million.
Plant backfit t;ost estimates are discussed in Appendix G, and are based on a composite average of industry estimates received during the pubile comment period.
The V/I is in all cases less than the criterion of 1000 person-rem /$ million.
NUREG-0869, Revision 1 8
October 1985
PWR dry containments with SGFCs have an additional safety system capable of rejecting post-LOCA containment heat loads.
SGFCs are designed to operate in-dependently in the post-LOCA environment and would, therefore, not be directly affected by loss of the sump or containment sprays. This independent heat rejection capability will ensure that the containment would not fail because i
of overpressure. Thus, although core melt could still be postulated, mainte-nance of containment integrity (by the SGFCs) would ensure that the radio-nuclides were contained and the public risk would be very low (see Appendix E).
4.1.2 PWR Subatmospheric Containments Only seven PWRs have subatmospheric containments, and five of these have re-ceived an OL.
Most of the discussion above on PWR large dry containments applies to this plant class as well, except that subatmospheric containments do not have safety grade fan coolers.
The estimated range of core melt freque,cy resulting from sump blockage i
discussed in Appendix D applies to these plants, and is in the range from 1.5 x 10-8 to 2.5 x 10-5/Rx yr.
As noted above, these numbers were derived assuming fibrous debris would transport to the sump and that the operator would take corrective action to maintain recirculation in 50% of the cases.
The estimated averted risk (for those plants where sump design problems may exist) is 20 to 310 person-rem /Rx.
The value-impact ratio is the same as for PWR large dry containments without safety grade fan coolers (10 to 200 person-rem /$M (at $1.5M/Rx) or 50 to 800 person-rem /$M (at $0.4M/Rx).
4.1.3 PWR Ice Condenser Plants Ice condenser plants are the most prone to eventual overpressure failure if loss of recirculation occurs (see Appendix E).
Although hydrogen igniters would protect against deflagration effects, the containment could fail as a result of steam production when the reactor vessel fails or within a few hours thereafter.
Therefore, the consequences of sump blockage in an ice condenser plant are higher than the consequences of sump blockage in PWR dry and subatmo-spheric containment plants.
Three different utilities have built (or are building) 10 ice condenser plants I
as five twin units. Consequently, these units have many similar design features, including the details of insulation, sump design, and interior layout.
Infor-mation on these design features was specifically obtained by the staff and is I
discussed further in Appendix E; however, the major points of similarity are summarized as follows:
(1) All ice condenser plants use reflective metallic insulation (RMI) on pri-mary system piping and major components.
Sump blockage effects associated i
with RMI are less severe than those associated with fibrous insulation debris.
1 (2) The majority of these plants have approach velocities in the vicinity of the sump of less than 0.2 ft/sec.
Therefore debris transpr and block-age are not likely.
i I
NUREG-0869, Revision 1 9
October 1985
(3) NPSH margins for the majority of these plants are in excess of 5 feet of water; however, the PWR sump failure probabilities were derived on the basis of a 1-to 5-foot blockage loss criterion.
The net effect is that the sump failure probabilities employed for the PWR dry and subatmospheric containments should be reduced when applied to the ice con-denser plants.
It is estimated that sump blockage probability leading to-loss of NPSH could be reduced to 1 to 9 x 10-8/Rx yr, or lower.
The estimated core melt frequency is 0.5 to 4.5 x 10-8/Rx yr.
The averted release would then be based on an estimated consequence value of 5 x 108 person-rem, and for a 25 year plant life would be 60 to 560 person-rem /Rx.
This estimate is in the same range as for other PWRs discussed previously.
The estimated V/I ratio for ice condenser plants is 160 to 1400 person-rem /$M (at an estimated cost of $0.4M/Rx) and is 25 to 380 person-rem /$M (at estimated cost of $1.5M/Rx).
4.1.4 BWRs with Mark I and Mark II Containments The potential blockage of BWR RHR suction intakes is similar to that estimated for PWRs, particularly because BWRs are being reinsulated with fiberglass and newer BWRs are being built with fiberglass insulation on the primary pressure boundary piping.
BWR intakes have suction strainer areas of typically 50 to 150 ft2 (which is on the lower side of the sump debris screen area in PWRs) and somewhat higher suction flows (8,000 to 12,000 gpm/ train).
On the other hand, suppression pool velocities are generally low (<0.2 ft/sec for bulk pool velocity) and the drywell-versus-wetwell design and separation tend to inhibit insulation debris transport.
Although no detailed BWR RHR in-take blockage probability analysis has been done, the staff has estimated that the probability of BWR intake blockages will be somewhat lower; therefore, the staff estimated a core melt frequency (equivalent to half the intake blockage probability) of 2 to 10 x 10-8/Rx yr in the calculations that follow.
Such a reduction is supported by analyses done for Limerick 1 (Philadelphia Electric, 1984) (using the proposed RG 1.82, Revision 1 analysis guidelines), which showed that plant had adequate NPSH margins (principally because the plant layout resulted in high NPSH availability).
The estimated conditional consequence associated with core melt for Mark I and Mark II containments (see Appendix E) is 5 x 108 person-rem. The use of a re-maining reactor life of 25 years and the estimated RHR intake blockage proba-bilities noted above results in averted releases of 250 to 1250 person-rem /Rx; these values apply only to those plants where blockage leading to loss of NPSH has been determined.
The estimated value/ impact ratios for Mark I and Mark 11 containments (assuming containment failure) are therefore approximately 630 to 3100 person-rem /$M (at an estimated cost of $0.4M/Rx) and 170 to 830 person-rem /$M (at an estimated cost of $1.5 M/Rx).
If credit is given for containment spray recovery and containment venting (which would ensure containment integrity despite loss of recirculation as a result of NUREG-0869 Revision 1 10 October 1985
1 blockage), the consequences are reduced to 5 x 105 person rem (see Appendix E).
The estimated averted releases would then be reduced to 25 to 125 person rem /Rx.
The estimated V/I ratios would then reduce to 63 to 300 person rem /$M (at an estimated cost of $0.4 M/Rx) and 17 to 80 person-rem /$M (at an estimated cost pf $1.5M/Rx).
4.1.5 BWRs with Mark III Containments Blockage considerations for Mark III containment are similar to those discussed in Section 4.1.4, although it could be argued that they are somewhat lower because of the wetwell-versus-drywell structural design.
For these calculations, an estimated blockage probability of 4 to 20 x 10-8/Rx yr was used.
The estimated core melt frequency is 2 to 10 x 10-8/Rx yr, and the estimated release conse-quence is 5 x 105 person rem (see Appendix E).
This release is less than that for Mark I and II containments because fission products would bubble through a subcooled suppression pool, and a significantly reduced source term would result.
Therefore, the estimated averted releases for Mark III containments are 25 to 125 person-rem /Rx (assuming again an outstanding reactor life of 25 years).
The calculated V/I ratios are 60 to 310 person-rem /$M (at an estimated cost of
$0.4 M/Rx) and 17 to 80 person-rem /$m (at an estimated cost of $1.5 M/Px).
4.2 Estimated Occupational Exposure Estimates of inplant radiological exposures associated with insulation replace-ment can be derived from actual experience during the steam generator repair and replacement at the Surry and Turkey Point plants.
Table 4.3 shows the work categories applicable to insulation replacement (as reported in NUREG/CR-3540) and the attendant exposures.
Table 4.3 Radiatien exposure to workers during insulation replacement (in person rem)
Work Surry 2 Surry 1 Turkey Pt 3 Turkey Pt 4 Installing 46.5 40.9 9.95 34.19 scaffolding Removing 15.16 19.35 70.80 63.64 insulation Reinstalling 57.80 6.30 85.72 4.17 insulation Total 119.46 65.55 166.47 102.00 Table 4.3 shows both the benefit of preplanning (or learning from experience at the first plant), as well as the variation between plants.
It should also be noted that the scaffolding installed at these plants was designed to remove NUREG-0869 Revision 1 11 October 1985
steam generators, and that entire steam generators were stripped of insulation and reinsulated.
Thus, the average of these exposures, 115 person-rem, is
-considered to be higher than that expected as a result of the insulation replacement needed to resolve A-43.
Discussions with Surry site personnel during the 1983 for comment period for NUREG-0869 indicated that a 50 person-rem exposure level for insulation re-placement is realistic if the job is thoroughly preplanned.
On this basis, the occupational exposure for major insulation replacement is estimated to be 50 person-rem per plant.
Exposures associated with alternate actions (such as increasing debris screen size) would be less.
Using the estimated averted risks developed in Appendix E and the replacement exposures given in Table 4.3, the net radiological effects of backfitting shown in Table 4.4 can be developed, t
Table 4.4 Radiological effects of backfitting 8ackfit Estimated Net averted Plant expot.ure averted risk exposure typo (person-rem /Rx)
(person-rem)/Rx (person-rem /Rx)
PWR ice 50 40 to 560
-10 to 510 condenser PWR dry w/o 50 20 to 300
-30 to 250 i
SGFCs and i
subatmospheric j
Mark I and II 50 250 to 1250 200 to 1200 l
Mark III 50 25 to 125
-25 to 75 i
)
The staff estimates that fewer than 10 plants would have the higher values of net averted exposure.
4.3 Impact on NRC Operations With respect to NRC staff review time, the impact of the proposed actions will be minimal.
The guidelines in Appendix A of the revised RG 1.82 and in NUREG-0897, Revision 1, October 1985 (and its supporting references) provide the technical information and specific guidance the staff reviewer needs to perform an evaluation in a reasonable time.
It is estimated that about 2 weeks of staff and related licensing review time will be needed per plant (estimated cost $5K per plant) to review analyses submitted.
Assuming 5 to 20 such de-talled responses are received, the estimated staff cost would be $25K to $100K.
i The experimental data and Generic plant information and calculations reported in NUREG-0897, Revision 1, October 1985 (and supporting references) represent an investment of nearly $3.0M by NRC and the Department of Energy.
This 4
NUREG-0869, Revision 1 12 October 1985
information is of value to both the NRC and industry, because similar tests need not be duplicated.
In addition, this extensive-hydraulic performance data base provides the basis for eliminating unnecessary inplant or sun.p model testing to examine vortex formation, which has been required for many plants at the OL stage.
4.4 Impact on Other Government Agencies Because nuclear plant design review and acceptance are done solely by the NRC staff, no impact on other government agencies is projected.
4.5 Public Impact If the recommendations in this report are adopted, there would be no impact the public.
Rather, there would be a value to the public: added reassurance that i
adequate sump designs exist for ensuring operability in the recirculation mode following a postulated LOCA.
As discussed above, only a small number of plants are expected to be susceptible to LOCA generated debris.
Issuance of a generic letter and the staff's technical findings to licensees and OL appli-cants outlining potential safety concerns associated with insulation debris 4
would add to the " defense-in-depth" concept, which ensures that the health and safety of the public is being maintained.
Because insulation is periodically replaced in operating plants, these findings could be used in the licensee's process for insulation selection and replacement.
4.6 Other Constraints The Commission has proposed to amend its requirements governing the backfit-ting of commercial power reactors and certain licensed nuclear facilities (see NRC Notice 84-137, dated November 30, 1984).
The regulatory analysis herein is consistent with those proposed requirements.
5 DECISION RATIONALE The regulatory alternatives pertinent to the resolution of USI A-43 were identi-fled in Section 3, and the consequences (values versus impacts) were discussed in Section 4.
This section compares regulatory alternatives and considers the value/ impact analysis results.
5.1 Comparison of Regulatory Alternatives 1
The regulatory alternatives can be compared as follows:
(1) Requiring all licensees and applicants to evaluate adequate NPSH margin would give the NRC a determination of which plants are most susceptible to debris blockage problems, and would identify plants that would bene-fit from corrective action.
This option would, however, result in the i
greatest overall industry cost impact: it is expected that only 5 to 10 t
plants would be identified as having a potential debris blockage problem that required backfit actions, while all licensees and applicants would incur analysis costs, i
(2) Requesting evaluations from only the licensees for those plants that are judged to have a high probability of containment failure would focus this NUREG-0869 Revision 1 13 October 1985 1
i 7---,-.
m-.
w-.---.,
-+e-. - -- - - -
,--.-----.,---,,,---+---,----m-
--,------%-r
--<r--
-r
<-m--,w, en--.-e--=---
safety issue on minimizing (or averting) the public risk of rsdiation exposure as the result of loss-of-recirculation capability.
This option would reduce the industry impact and would concentrate the enalysis effort on those plants (a) that are subject to containment failure and (b) that would benefit from corrective action.
Identification of only those plants that have a high probability of debris blockage would require a plant-specific analysis because of the plant-to-plant variabilities discussed in Section 4; no generic conclusion (or identification) is possible.
Therefore, this option is similar to Option (1) but would involve fewer plants.
Both options are discussed as a single option in the decisionale rationale presented belcw.
(3) The forward fit of RG 1.82, Revision 1, and SRP 6.2.2, Revision 4, to Stan-dard Plants and new construction permits (cps) would address sump design ano debris blockage effects in new plants at a very low incremental cost.
Continued use of RG 1.82, Revision 0 (which has the 50% blockage crite-rion) does not adequately address this issue, and is technically incon-sistent with the technical findings developed for the resolution of USI A-43.
This option would include the issuance of NUREG-0897, Revision 1, October 1985, and of a generic letter for information, providing the industry with technical findings that would be useful for industry safety assessments regarding routine change of insulation materials.
(4) Issuing NUREG-0897 only, without using current technical findings to revise RG 1.82 and SRP Section 6.2.2, would be contrary to (a) addressing' safety concerns using current and the most reliable technical findings and (b) the need to remove the current 50% blockage criterion in RG 1.82, Revision 0.
5.2 Rationale for Selecting the Recommended Resolution The rationale for selecting a resolution position is based in part on the values and impacts discussed in Section 4, which presents the consequences of sump blockage problems for different containment design categories.
Conclusions are based on core melt frequency, potentially averted releases (AR), and value/
impact ratio, utilizing the 1000 person-rem /$M criterion for backfit considera-tion.
Table 5.1 summarizes the values and impacts developed in Section 4.
5.2.1 Options 1 and 2 Options 1 and 2 are discussed together because Option 2 is a modification of Option 1.
PWR Dry Containments PWR dry containment plants fall into two categories, those with safety grade fan coolers (SGFCs) and those without.
PWR dry containments with SGFCs have an additional capability to reject post-LOCA decay heat and prevent containment overpressurization, thereby ensuring containment intcgrity.
Furthermore, large dry containments are least susceptible to abrupt failures as a result of hydro-gen burns, steam spikes, etc.
Therefore, it is the staff's opinion that even if loss of sump should occur, current designs would contain postulated core NUREG-0869, Revision 1 14 October 1985
Table 5.1 Summary of calculated values and impacts associated with various containment designs for resolution of USI A-43 Estimated core Calculated Calculated value/
Type melt probability 1 risk averted, AR impact ratio 3 containment (1/Rx yr)
(person rem /Rx)
(person-rem /$M)
PWR dry w/o 1.5 to 25 x 10-8 19 to 310 48 to 780 SGFCs and 12 to 210 subatmospheric PNR dry 1.5 to 25 x 10-6 w/SGFCs PWR ice 0.5 to 4.5 x 10-6 60 to 560 160 to 1400 condenser 3 25 to 380 PWR dry w/o 1.5 to 25 x 10-6 2 to 31 5 to 33 SGFCs and 1 to 21 subatmospheric w/ spray recovery Mark I and II 2 to 10 x 10-6 250 to 1250 630 to 3100 170 to 830 Mark III 2 to 10 x 10-6 25 to 125 60 to 310 17 to 80 Mark I and II 2 to 10 x 10'6 25 to 125 60 to 310 w/ venting and 17 to 83 spray recovery 1The estimated core melt frequency is based on the conditional consequences discussed in Appendix E, the sump blockage frequency estimates discussed in Appendix D, and the assumption that 50% of the time blockage occurs leads to loss of NPSH and core melt follows.
This assignment of a conditional core melt probability of 0.5 is felt to be realistic from the viewpoint of poten-tial detection of flow degradation, potential operator followup action to correct this situation, and sump design variability.
2The value/ impact ratios have been calculated for an estimated cost of
$0.4M/Rx (this assumes backfit costs vould be minimal) and for an estimated cost of $1.5M/Rx (this cost assumes replacement of troublesome insulation (s));
see Appendix G for a discussion of estimated costs.
3A separate estimate of sump blockage probability made for the ice condenser plants (see Appendix E) takes into account their specific design features.
NUREG-0869, Revision 1 15 October 1985
melt effects and maintain public releases at appropriate levels.
Thus, backfit action (including analyses) for PWR dry containments with SGFCs is not justi-fiable, and Option (1), which includes such a requirement for PWR dry contain-ments, is not attractive.
PWR dry containments-(without SGFCs) have been evaluated (see Section 4.1.1) and the results are as follows:
(1) Probabalistic risk analyses (PRAs) discussed in Section 4 have concluded that the core melt probability from this class of plants for the sequence involving a blocked sump is in the range 1.5 x 10-8 to 2.5 x 10-5 For a number of reasons as discussed above, the staff believes believes the best estimate of core melt frequency for this sequence is at the lower end of this range, about 3 x 10-6 This level of core melt frequency does not support a backfit requirement.
(2) The calculated range of averted risk (assuming containment failure, as per WASH-1400 (NUREG-75/014)) is approximately 20 to 300 person-rem /Rx.
This is a low-to-moderate level.
If corrective operator action to restore containment sprays (should debris blockage be encountered) is begun before containment failure, then the estimated averted risk is 2 to 30 person-rem /Rx.
(3) The calculated V/I ratios were 50 to 800 person-rem /$M (low retrofit cost) 10 to 200 person-rem /$M (high retrofit cost)
Utilization of a 1000 person-rem /$M criterion does not support a backfit requirement.
(4)
If operator detection of the onset of blockage and taking corrective actions are included in these consequence calculations, the estimated releases noted above would be reduced by 10% and the V/I ratios would also be reduced by 10%
Therefore, based on the above assessment, it is the staff's opinion that all PWR dry containments can be excluded from any backfit requirements.
PWR Subatmospheric Containments The rclease consequences resulting from an assumed failed containment assoc-lated with PWR subatmospheric containments are estimated to be the same as those for PWR dry containments without SGFCs (see Appendix E).
Sump blockage probabilities are judged to be in the same range also.
Therefore, the averted releases and V/I estimates are the same as noted in the preceding section.
Restoration of containment spray capability by operator action reduces esti-mated releases by a factor of 10.
Therefore, backfit requirements for subatmospheric containments are not sup-ported for the same reasons cited for not requiring such backfits for PWR dry containments without SGFCs.
NUREG-0869, Revision 1 16 October 1985
i PWR Ice Condenser Plants As noted in Section 4.1.3 and Appendix E, PWR ice condenser plants are most prone to overpressure failure.
However, fewer than 10 plants have this type of con-tainment design, layout, and recirculation flow features (see Section 4.1.3).
For these plants, the staff has determined a separate value of sump failure probability based on known plant design parameters.
The result is that the probability for sump blockage is lower, thus offseting the higher estimated consequences from core melt.
The calculated averted risk is 60 to 560 person-rem /Rx (similar in value to that for the PWR dry containments) and the calculated V/I ratios are 160 to 1400 person rem /$M (low cost retrofit estimates) and 40 to 380 person rem /$M (high retrofit cost estimate).
Thus, based on the calculated averted risk and V/I noted above, a backfit re-quirement for PWR ice condenser plants is not indicated.
BWR Mark I and Mark II Plants NRC and industry PRA studies report an estimated BWR core melt frequency (attributable to all causes) of 2 x 10-5 to 3 x 10-4/Rx yr.
The LOCA con-tribution to this total frequency is 1% to 10%, with large LOCAs having the lower value.
Thus, core melt frequancy related to this safety issue is judged to be on the order of 10-7 to 10-8/Rx yr.
If backfit actions are viewed from the perspective of a reduction in core melt frequency, the gain would minimal (this would apply only to BWRs where debris blockage was identified to be significant).
Therefore, from this viewpoint, a backfit requirement is not i
indicated.
The calculated averted risk for this class of BWRs is 250 to 1250 person-rem /Rx (see Section 4.4), based on blockage probabilities derived from PWR studies.
For reasons discussed in Section 4, these values are conservative.
The corre-sponding V/I ratios are 630 to 3100 person-rem /$M (low retrofit cost estimate) and 170 to 830 person-rem /$M (high retrofit cost estimate).
On the basis of the averted risk and V/I assessment, an argument could be made for proceeding with some type of plant assessment.
However, this view must be balanced with the knowledge that only a few BWRs may fall into the higher V/I category.
In addition (as for PWRs), it is the staff's opinion that the majority of BWRs will be at the lower end of the blockage frequency spectrum.
An example is the analysis submitted for Limerick 1 (Philadelphia Electric, 1984). When this analysis was evaluated according to the guidelines in the proposed RG 1.82, Revision 1, it showed the plant had adequate NPSH margin.
Another factor to be considered is suction realignment capability if blockage should occur.
Mark I plants can be realigned to alternate water sources (e.g.,
the condensate storage tank).
Twenty-three of 41 BWRs have Mark I contain-ments. Mark II containments do not have the option of being aligned to the con-densate storage tank, although realignment to the fuel storage pool would be a l
possibility if such procedures were provided to the operators.
4 An additional significant factor is the capability for controlled venting.
Controlled venting (an option that can be submitted for staff approval on a 1
NUREG-0869, Revision 1 17 October 1985
9 plant-specific basis) provides a means to maintain containment integrity by avoiding overpressurization.
With wetwell venting, the calculated averted risk is reduced by 10% and the V/I cited above decreases accordingly.
It is the staff's understanding that the BWR Owners Group is' supporting implement-ation of controlled venting for all plants.
Therefore, given the above considerations, it is the staff's. view that backfit requirements for Mark I and Mark II plants are of marginal significance and therefore are not proposed.
MarkIIIPiants The consequences (releases) associated with Mark III containments are esti-mated to be 10% lower than those for Mark I and II plants because, should the containment fail, the Mark III containments channel fission products through the pool before they are released to the environment with or without wetwell venting (see Appendix E).
Therefore the calculated averted risk for Mark III containments is 25 to 125 person rem /Rx.
The V/I ratios are 60 to 310 person-rem /$M (low retrofit cost estimate) or 17 to 80 person-rem /$M (high retrofit cost estimate).
On the basis of these values, a backfit requirement is not supportable for the Mark III containments.
5.2.2 Option 3 Opticn 3, which is a forward fit application of RG 1.82, Revision 1, to Standard Plant and new CP applications, would have no impact on current licensees and applicants.
The incremental impact of requiring a a sump blockage analysis for a new design would be very small; therefore the V/I is favorable.
An important aspect of Option 3 is that all licensees and applicants would be informed of the technical findings of A-43 and the recommended actions in RG 1.82, Revision 1, for performing a sump blockage analysis.
This information would give them a basis for considering analysis and corrective action as they deem necessary.
This information also would give each licensee or applicant a better basis for considering changes of insulation.
(The periodic changing of insulation is a common practice at operating plants.) Therefore, because of its obvious high value and benefit amd very low impact, the staff recommends adopting Option 3.
5.2.3 Option 4 Option 4 would use no overt means to inform the industry regarding the new in-formation and understanding developed, and the existing incorrect NRC guidance with respect to sump blockage would remain in place.
Although this option has no impact, it also has no value.
The prospect of having incorrect guidance standing is unacceptable to the staff; thus this option was rejected.
5.3 Recommended Regulatory Action On the basis of the discussion in Section 5.2, the staff recommends adoption of Option 3 and the following specific actions:
NUREG-0869, Revision 1 18 October 1985
l (1) Issue the staff's technical findings (NUREG-0897, Revision 1, October 1985) for use as a technical information source.
(2) Issue SRP Section 6.2.2, Revision 4, and RG 1.82, Revision 1.
These revi-sions reflect the staff's technical findings reported in NUREG-0897, Revi-sion 1, October 1985.
This revised regulatory guidance would apply only to future CP applications, Preliminary Design Approvals (PDAs), Final Design Approvals (FDAs) that have not received prior approval, and future applications for license to manufacture.
It would be effective.6 months following issuance.
(3) Issue a generic letter for information only to all holders of an Operating License or Construction Permit outlining the safety concerns regarding potential debris blockage and recirculation failure due to inadequate NPSH.
It is suggested (but not required) that licensees utilize RG 1.82, Revision 1, as guidance for conduct of the 10 CFR 50.59 review for future plant modifications involving replacement of insulation on primary system piping and/or equipment.
If, as a result of NRC staff review of licensee actions associated with replacement or modification to insulation, the staff decides that SRP 6.2.2, Revision 4, and/or RG 1.82, Revision 1, cri-teria should be (or should have been) applied by the licensee, and the staff seeks to impose these criteria, then the NRC will treat such an action as a plant-specific backfit pursuant to 10 CFR 50.109.
6 PLAN FOR IMPLEMENTATION The proposed resolution of USI A-43 would be accomplished through the following actions:
(1) Issue the staff's technical findings (NUREG-0897, Revision 1, October 1985) for use as an information source by applicants, licensees, and the staff.
(2) Issue Revision 1 of RG 1.82 to reflect the technical findings reported in NUREG-0897, Revision 1, October 1985.
In particular, the 50% screen blockage criterion would be replaced by a plant-specific debris. blockage
~
assessment.
RG 1.82, Revision 1, provides specific guidance acceptable to the staff for assessing sump performance and RHR suction intakes in-cluding debris blockage effects.
RG 1.82, Revision 1, would apply'to:
(a) construction permit applications and Prel.iminary Design Approvals (PDAs) that are docketed more than 6 months following issuance of RG 1.82, Revision 1; (b) applications for Final Design Approvals (FDAs),
r sthndardized designs that are intended for referencing in future construction permit applications, that have not received approval 6 months follow-ing issuance of RG 1.82, Revision 1; and (c) applications for licenses to manufacture that are docketed more than 6 months following issuance of RG 1.82, Revision 1.
-(3) Issue'NRC SRP Section.6.2.2, Revision 4, " Containment Heat Removal Systems,"
to reflect the guidance in RG 1.82, Revision 1, and the te:hnical findings NUREG-0869, Revision 1 19 October 1985 s
in NUREG-0897, Revision 1, October 1985.
The revised SRP section would apply to new CP applications and Standard Plant designs only, effective 6 months after the revisions are issued.
(4) Issue a Generic Letter to all holders of an operating license or construc-tion permit outlining the potential safety concerns related to potential post-LOCA debris blockage and the inadequacy of the current 50% blockage criterion contained in Revision 0 of RG 1.82.
This action is significant because licensees should be made aware of the potential for recirculation blockage from insulation following a LOCA, and replacing insulation in operating plants is common practice.
(A draft generic letter is provided in Appendix H.) The proposed generic letter contains no requirements and no request for a response.
7 STATUTORY CONSIDERATIONS 7.1 NRC Authority Because the proposed changes are revisions to RG 1.82 and SRP Section 6.2.2, these actions fall within the statutory authority of the NRC.
Further, the recommendation to require applicants / licensees to demonstrate adequate sump performance falls within the statutory authority of the NRC to regulate and ensure the safe operation of nuclear power plants.
7.2 Need for National Environmental Policy Act (NEPA) Statement The proposed changes and potential plant backfits do not warrant a NEPA statement.
8 BIBLIOGRAPHY The following U.S. Nuclear Regulatory Commission documents were used in the preparation of this report:
NUREG-75/014, " Reactor Safety Study," 1975 (formerly WASH-1400).
NUREG-0800, " Standard Review Plan," July 1981.
Revised Standard Review Plan Section 6.2.2, Revision 4, " Containment Heat Removal Systems," available from the NRC Division of Technical Information and Document Control, 1717 H Street, NW, Washington, DC 20555.
NUREG-0897, Revision 1, " Containment Emergency Sump Performance, Technical Findings Related to USI A-43," November 1984.
NUREG/CP-0933, SAND 82-1659, " Proceedings of the Workshop on Containment Integrity," Volume II of II, October 1982.
NUREG/CR-2403:
see Reyer.
NUREG/CR-2403, Supplement No. 1:
see Kolbe.
NUREG/CR-2759:
see Argonne.
NUREG/CR-2760:
see Padmanabhan and Hecker, June 1982.
NUREG-0869, Revision 1 20 October 1985
. -. - - ~. =
+
L E
NUREG/CR-2761: -see Padmanabhan, September 1981.
.NUREG/CR-2772: -see Padmanabhan, June 1982.
- NUREG/CR-2982:
see Brocard, July 1983.
NUREG/CR-2791:
see Wysocki.
NUREG/CR-2792: 'see Kamath.
NUREG/CR-3394: _see Wysocki, July 1983.
NUREG/CR-3540:
see Parkhurst, December 1983 NUREG/CR-3616:
see Brocard, December 1983.'
Regulatory Guide 1.82, Revision 1, " Water Sources for Long-Term Recirculation Cooling Following a Loss-of-Coolant Accident," July 1985.
Other documents used in the production of this-report were:
~
' Argonne National Laboratory, "Results of Vertical Outlet Sump Tests," joint ARL/Sandia National Laboratory report, ARL47-82/ SAND 82-1286/NUREG/CR-2759, 2-September 1962.
Brucard D.,' " Buoyancy, Transport and Head Loss of Fibrous Reactor Insulation,"-
U.S. Nuclear Regulatory Commission -report, NUREG/CR-2982, Revision 1, July 1983.
j
--, " Transport and Screen Blockage Characteristics of Reflective Metallic Insulation Materials," U.S. Nuclear Regulatory Commission report, NUREG/CR-3616, December 1983.
1
- Ferrell, W. L. et al., "Probabilistic Assessment of'USI A-43," Science Appli-
- cations, Inc.~ report, September 1982.
Kamath, P., T. Tantillo, and W. Swift, "An' Assessment of. Residual Heat Removal and Containment Spray Pump Performance Under Air and Debris Ingesting Condi-
-tions," Creare,.Inc., Hanover, NH, U.S. Nuclear Regulatory Commission report,
- NUREG/CR-2792,' September 1982.
-Kolbe, R. and E. Gahan, " Survey-of Insulation Used in Nuclear Power Plants and
- the Potential for Debris Generation," Burns and Roe, Inc., Oradell, NJ, U.S.
- Nuclear Regulatory Commission report, NUREG/CR-2403, Supplement 1, May 1982.
- Padmanabhan, M., "Results of Vortex Suppressor Tests, Single Outlet Sump Test and Miscellaneous SensitivityLTests," Alden Research Laboratory, Worcester Polytechnic Institute,~ Holden, MA, U.S. Nuclear Regulatory Commission report,
- NUREG/CR-2761, September 1982.
4
--, "Hyiraulic Performance of Pump Suction Inlet for Emergency Core Cooling Systems in Boiling Water Reactors," Alden Research Laboratory, Worcester ' Poly-
- technic Institute, Holden, MA, U.S. Nuclear Regulatory Commission report,
. NUREG/CR-2772, June 1982.
-NUREG-0869, Revision 1 21 October 1985
Padmanabhan, M., and G. E. Hecker, " Assessment of Scale Ef fects on Vortexing, Swirl, and Inlet Losses in Large Scale Sump Models," Alden Research Laboratory, Worcester Polytechnic Institute, Holden, MA, U.S. Nuclear Regulatory Commission report, NUREG/CR-2760, June 1982.
Parkhurst, M.
A., L. Rathbun, and D. Murphy, " Radiological Assessment of Steam Generator Repair and Replacement," Pacific Northwest Laboratory, Rithland, WA, U.S. Nuclear Regulatory Commission report, NUREG/CR-3540, December 1983.
Philadelphia Electric Company, " Limerick Generating Station, Units 1 and 2 Containment Emergency Sump Performance," April 2, 1984.
Reyer, R., et al., " Survey of Insulation Used in Nuclear Power Plants and the Potential for Debris Generation," Burns and Roe, Inc., Oradell, NJ, U.S. Nuclear Regulatory Commission report, NUREG/CR-2403, Supplement 1, October 1981.
Wysocki, J.
J., " Methodology for Evaluation of Insulation Debris," Burns and Roe, Inc., Oradell, NJ, U.S. Nuclear Regulatory Commission report, NUREG/CR-2791, September 1982.
--, "Probabilistic Assessment of Recirculation Sump Blockage Due to Loss-of-Coolant Accidents, U.S. Nuclear Regulatory Commission report, NUREG/CR-3394, Volumes 1 and 2, July 1983.
NUREG-0869, Revision 1 22 October 1985
APPENDIX A
SUMMARY
OF PUBLIC COMMENTS RECEIVED AND ACTIONS TAKEN NUREG-0869, Revision 1 October 1985
APPENDIX A
SUMMARY
OF PUBLIC-COMMENTS RECEIVED AND ACTIONS TAKEN 1 INTRODUCTION The technical findings related to Unresolved Safety Issue (USI) A-43 were pub-lished for comment in May 1983.
Notice of the publication was placed in the Federal Register on May 9, 1983.
The official comment period lasted for 60 l
days and ended on July 11, 1983.
However, comments were received into Sep-tember 1983, with followup comments received into November 1983.
A listing of those who responded during the period and afterwards is shown in Table 1.
Copies of the comment letters are on file in the NRC Public Document Room, 1717 H Street, NW, Washington, DC.
A public meeting was held on June 1 and 2, 1983, at Bethesda, Maryland, to offer additional opportunity for public comments; however, attendance was very small.. Followup discussions were held with respondees to clarify issues raised at this meeting and in the written comments.
An overview of the comments received is provided in Section 2 below.
Sec-tion 3 contains summaries of significant comments and the actions planned to resolve them.
2 OVERVIEW 0F COMMENTS RECEIVED The major written comments received addressed seven specific subject areas.
The comment categories and commentors are listed in Table 2 below.
The com-mentors are identified in Table 2 as follows:
Alden Research Laboratory (ARL);
Atomic Industrial Forum (AIF); BWR Owners Group (BWR); Commonwealth Edison (CEd); Consumers Power Co. (CPC); Creare Research and Development (CRD);
Diamond Power Co. (DPC); General Electric (GE); Gibbs and Hill, Inc. (GH);
Northeast Utilities (NE); and Owens-Corning Fiberglass, Inc. (OCF).
By category, the actions taken in response to these comments are as follows:
Categorie:; 1 and 6:
Tables have been added to NUREG-0897, Revision 1 and NUREG-086?, Revision 1 to include the additional plant insulation information provided during the public comment period.
The text of the NUREGs has been revised to reflect recommended insulation definitions and the need to evaluate the specific insulation employed.
Categories 2 and 4:
The cost estimates provided by differant industry groups L
have varied over a wide range.
With the exception of Diamond Power Company, respondees claimed that the cost estimates in value/ impact analysis were too low.
The revised value/ impact analysis reflects an averaged value derived from costs provided.
Category 3:
A detailed sump blockage probability analysis has been performed and is reported in NUREG/CR-3394.
The results were used in the revised value/
impact analysis.
These results show a sump blockage probability range for pressurized water reactors (PWRs) of 10 8 to 5 x 10.s/Rx yr and a strong depen-dence on plant design.
NUREG-0869, Revision 1 A-1 October 1985
Table 1 Persons who commented on the technical findings related to USI A-43*
Alden Research Laboratory (ARL), M. Padmanabhan, letter to A. Serkiz (NRC),
" Comments on NUREG-0897 and 0869," June 13, 1983.
ARL, M. Padmanabhan, letter to A. Serkiz (NRC), " Revision to Table A-3 in NUREG-0869," June 22, 1983.
Atomic Industrial Forum, R. Szalay, letter to the Secretary of the Commission, "NRC's Proposed Resolution of Unresolved Safety Issue A-43, Containment Emer-gency Sump Performance, Contained in NUREG-0869," July 22,1983.
Atomic Industrial Forum, J. Cook, letter to R. Purple (NRC) and enclosure
" Examples of Staf f Review Going Beyond Approved Regulatory Criteria," June 4, 1984.
BWR Owners Group, T. J. Dente, letter to T. P. Speis (NRC), "BWR Owners' Group Comments on Proposed Revision to Regulatory Guide 1.82, Rev. 1,"
October 18, 1983.
BWR Owners Group, D. R. Helwig, letter to V. Stello (NRC), BWR Owners' Group comments on Regulatory Guide 1.82, Revision 1, July 16, 1984.
Commonwealth Edison, D. L. Farrar, letter to the Secretary of the Commission, "N9 REG-0897, Containment Emergency Sump Performance; Standard Review Plan Sec-tion 6.2.2, Rev. 4, Containment Emergency Heat Removal Systems; and NUREG-0869, USI A-43 Resolution Positions (48FR2089; May 9,1983)," July 13,1983.
Consumers Power, D. M. Budzik, letter to the Secretary of the Commission,
" Comments Concerning Regulatory Guide 1.82, Proposed Revision 1 (File 0485.1, 0911.1.5, Serial: 23206)," July 15, 1983.
Crearo, W. L. Swift, letter to P. Strom (SNL), " Comments on Figure 3-6 of NUREG-0897 and Table A-9 of NUREG-0869," June 13, 1983.
Diamond Power Company, R. E. Ziegler and B. D. Ziels, letter to K. Kniel (NRC),
" Containment Emergency Sump Performance, USI A-43," July 11, 1983.
Diamond Power Specialty Company, B. D. Ziels, letter to A. Serkiz (NRC), "HDR Test Result Summary, MIRROR Insulation Performance During LOCA Conditions,"
December 6, 1984.
General Electric (GE), J. F. Quirk, letter to K. Kniel (NRC), " Comments on Emergency Sump Documents," July 11, 1983.
GE, J. F. Quirk, letter to T. P. Speis (NRC), "Couments on Proposeo Regulatory
- Guide 1.82, Rev.
1," October 17, 1983.
- Including comments on NUREG-0869, NUFEG-0897,' proposed Revision 1 to Regula-tory Guide 1.82, and proposed Revision 4 to Section 6.2.2 of the Standard Review Plan (SRP, NUREG-0800).
NUREG-0869, Revision 1 A-2 October 1985
Table 1 (Continued)
Gibbs and Hill, Inc., M. A. Vivirito, letter to the Secretary of the Commis-sion, " Comments on Proposed Revision No. I to RG 1.82," July 11,1983.
Northeast Utilities, W. G. Counsil, letter to K. Kniel (NRC),."Haddam Neck, Millstone Nuclear Power Station, Unit Nos. 1, 2, and 3, Comments on NUREG-0897, SRP Section 6.2.2 and NUREG-0869," September 2, 1983.
Owens Corning Fiberglass (0CF), G. H. Hart, letter to A. Serkiz (NRC),
" Comments on NUREG-0897 and NUREG-0869," June 23, 1983.
OCF, G. H. Hart, letter to A. Serkiz (NRC), " Updated Comments on NUREG-0897 and NUREG-0869," July 14, 1983.
OCF, G. P. Pinsky, letter to K. Kniel (NRC), " Comments on NUREG-0879 and -0896,"
July 14, 1983.
OCF, G. H. Hart, transmittal to A. Serkiz (NRC), "HDR Blowdown Tests with NUKON Insulation Blankets," February 18, 1985.
Power Component Systems, Inc., D. A. Leach, letter to A. Serkiz (NRC),
" Nuclear Grade Blanket Insulation," November 8,1984.
Table 2 Categories addressed in major written comments Comment Category ARL AIF BWR CED CPC CRD DPC GE GH NE OCF (1)
Survey of insulation used is X
X not current or complete.
(2)
Cost estimates are low.
X X
(3)
Estimates of sump blockage X
X X
X probabilities are high.
(4)
Value-impact analysis questioned.
X X
X X
(5)
BWRs should be exempt; A-43 is a X
X X
PWR issue.
(6)
Insulation material definitions X
X and descriptions need revision for clarity and completeness.
(7)
Technical comments on and X
X X
X X
X X X clarifications of subject matter in NUREG-0897 and NUREG-0869.
NUREG-0869, Revision 1 A-3 October 1985 I
l
=
By category, the actions taken in response to these comments are as follows:
Categories 1 and 6: Tables have been added to NUREG-0897, Revision 1 and NUREG-0897, Revision 1 to include the additional plant insulation information provided during the public comment period. The text of the NUREGs has been revised to reflect recommended insulation definitions and the need to evaluate the specific insulation employed.
Categories 2 and 4: The cost estimates provided by different industry groups have varied over a wide range. With the exception of Diamond Power Company, respondees claimed that the cost estimates in value/ impact analysis were too low..The revised value/ impact analysis reflects an averaged value derived from costs provided.
Category 3:
A detailed sump blockage probability analysis has been performed and is reported in NUREG/CR-3394.
The results were used in the revised value/
impact analysis.
These results show a sump blockage probability range for pressurized water reactors (PWRs) of 10 8 to 5 x 10 8/Rx yr and a strong dependence on plant design.
Category 5:
NUREG-0869 and Regulatory Guide 1.82 have been revised to speci-fically identify areas of concern for boiling water reactors (BWRs) and for PWRs.
Category 7:
Technical corrections and clarifications have been made in the appropriate sections of NUREG-0897 and NUREG-0869.
The NRC staff greatly appreciates the review and comments provided by the The time and effort they have taken to review USI A-43 has resulted respondees.
in an improved report that will reflect current findings and a balanced position with respect to this safety issue.
NUREG-0869, Revision 1 A-4 October 1985
3 COMMENTS RECEIVED AND PROPOSED ACTION (OR RESPONSE) ACTIONS The NRC staff has given complete and careful consideration to all comments received on USI A-43.
Summaries of significant comments and the actions taken by the NRC staff in response are provided in Table 3.
Comments are presented in alphabetical order, based on the name of the commenting institution.
NUREG-0869, Revision 1 A-5 October 1985
Table 3 Comments received on USI A-43.and NRC staff response
-E i5:?
Comment NRC Staff Response S$
Alden Research Laboratory
?!
$L ARL noted typographical errors and proposed These corrections and clarifications have been incorporated 5.
technical clarification to several tables into NUREG-0897 and NUREG-0869.
E Atomic Industrial Forum g
l The cost impact of $550,000/ plant used in Costs impacts were re-evaluated based on cost estimate value/ impact analysis is low by at least information received from AIF and other respondees
(
a factor of 2.
4 Economic considerations related to reduced The essence of a value/ impact analysis is that'it attempts probability of plant damage should be excluded to identify, organize, relate, and make " visible" all the from the cost-benefit balancing.
Decisions significant elements of value expected to be derived from T
should be based primarily on the value/ impact a proposed regulatory action as well as all significant ratio.
elements of impact.
The net values are compared with the net impacts.
Thus if a proposed safety improvement is accompanied by an adverse side effect, the impairment is is subtracted from the improvement to arrive at a net safety value for consideration in the value/ impact assessment.
Similarly, when the immediate and prospective cost impacts are summed, thay should include all elements of economic impact on licensees, such as costs to design, plan, install, test, operate, maintain, etc.
Plant downtime or decreased plant availability is included when applicable.
The summed R
impacts, however, should be net impacts, for comparison EI with net values. Thus, any reductions in operating costs, E[
improvements in plant availability, or reductions in the probability of plant damage are properly a factor in deter-mining net adverse economic impact.
Future economic costs l
28 and savings are appropriately discounted.
Tchle 3 (Continued)
E A9 Comment NRC Staff Response S
Qualitative differences among impact elements are respected,
,y and distinctive elements of impact (of which averted plant-r damage probability, as a favorable rather than adverse 7
impact, is a prominent example) are separately identified, for appropriate consideration in regulatory decision making.
o s
The ratio of avoided public dose to the gross cost of imple-mentation is ordinarily a major decision factor.
- However, it is not by itself always a good guide to a sound regula-tory decision.
The issues _ involved are often too complex for a decision on this criterion alone.
Other factors that enter, often in important ways, may include any economic benefits that reduce a net adverse economic impact, the safety importance of the issue, and values and impacts that 4
cannot or cannot readily be quantified; for example, jeop-ardy to a deltase. layer in the defense-in-depth concept or expected reductions in plant availability that can be foreseen but not precisely estimated.
A sound regulatory decision rests on adequate consideration i
of all significant factors.
An overly simple approach can mislead if it simplifies away complexities that are the essence of the issue at hand.
t The assumption that sump failure will occur in A detailed sump blockage probability analysis has been per-the case of 50% of the large LOCAs should be formed and is reported in NUREG/CR-3394. The results show.
justified.
a wide range of sump blockage failure probabilities (i.e.,
,g 3E-6 to SE-5/ reactor year) and a high dependency on plant g
design and operational requirements.
These results are reflected in a revised value impact analysis utilizing a o
[
range of sump failure probabilities.
e The use of PWR release categories from The containment failure probabilities and release cate-WASH-1400 is too conservative.
Containment gories used in the regulatory analysis for USI A-43 were failure probabilities used in WASH-1400 based on information presented in WASH-1400, and also on i
Table 3 (Continued)
E A
y Comment NRC Staff Response e
are inadequate to describe the nuclear other considerations.
The comments presented by an AIF E?
industry's present knowledge in this field.
subcommittee regarding the validity of continued use of fl Releases due to " vessel steam explosion" WASH-1400 assumptions, etc. are being evaluated through Si are unrealistic and should not be considered.
other activities such as:
reevaluation of source terms, 8
SASA studies, etc.
USI A-43 regulatory analyses were based on the following considerations and for the reasons noted:
g (1) WASH-1400 assumptions were utilized to provide a com-mon baseline calculations for reference plants and were used to estimate increases in releases due to a postulated loss of recirculation flow capacity.
Until revised failure mechanisms and new source terms are determined, this approach provides a consistent set of calculations.
T (2) Although using a small containment failure probability associated with steam explosion would be more appro-priate, release category PWR-1 (which includes steam explosion) was not a dominant contributor to release.
Release categories PWR-2, -4, and -6 were the dominant pathways contributing to increases releases due to a failed sump for the plants analyzed.
(3) Basing release effects on the assumption of simul-taneous failure of core cooling and loss of contain-ment sprays is conservative. If containment were not lost (as would be the situation for PWRs that have dry E?
containments with safety grade fan cooler systems),
EI the LOCA energy could be disspated without containment Ei overpressurization and failure.
Thus releases asso-ciated with PWR-2 and -4 categories could be discounted ts and PWR-6 releases only used.
Such considerations 8
have been incorporated into this revised regulatory analysis.
I
Table 3 (Continued)
E=
[
Comment NRC Staff Response e
(4) Other factors--such as containment structural design E
margins that argue against gross containment failures i
1 (as postulated in WASH-1400), realignment to alternate E
water sources, controlled venting for BWRs, etc.--have E
also been considered this revised regulatory analysis, w
The use of the CRAC Code and a "no-evacuation,"
The 50-mile radius reflects a substantial part (though not 50-mile-radius model to develop public doses all) of the total population dose, and is thus a reasonable is unrealistic.
index of the radiological effect on the public.
Standard-ization of calculations to that radius is helpful in com-paring risks associated with different issues and average such risks for use with the 1000 person-rem /$M criterion.
Evacuation of people is not considered because calculations T
suggest that, although it may sometimes be important for people directly affected, the effect of evacuation on the total population dose is likely to be small.
NRC should utilize information developed more Possible changes in the source terms are being considered recently (i.e., NUREG-0772) to reassess and by the special task force established by the Commission reduce the source terms, rather than continue to review the source-term issue.
Changes would be pre-to use the PWR-2 and PWR-3 release categories mature before this group completes its evaluation and the from WASH-1400.
new values are accepted by all parties involved.
NRC should utilize the " leak before break" Elastic plastic fracture mechanics analysis techniques to concept in evaluating the safety significance analyze pipe break potential has been used in USI A-2, with of A-43.
a limited number of PWRs being analyzed.
For USI A-2, the R
submittal of such analyses for specific break locations (on a plant-specific basis) will require obtaining an exemption o
E from the requirements of GDC4.
Submittal of such analyses to address the USI A-43 debris blockage issues would be reviewed by staff on a plant-specific basis, should
- li-4 censee or applicant elect to utilize this approa, i
L
Table 3 (Continued)
E=?
Comment NRC Staff Response w
BWR Owners Grot.p d'1 After quick review of the proposed revision to The requirement for long-term decay heat removal is i
the regulatory guide, the BWR Owners Group and applicable to light-water reactors, both BWRs and PWRs.
E GE maintain that USI A-43 is not a generic s
issue for BWRs.
The revisions to RG 1.82, which now proposes All types of insulation should be evaluated for the specific criteria for BWRs, should apply potential of debris generation, transport, and suction only to light-water reactors that have any strainer blockage.
The wide variation in plant designs potential for harmful debris generation (i.e.,
and insulation employed does not support a generic light water reactors that extensively use statement.
fibrous insulation).
T These comments and any future comments by RG 1.82, Revision 1 (along with NUREG-0897, NUREG-0869 I$
the BWR Owners Group should not substitute and SRP 6.2.2, Revision 4) was issued "for comment" in for the normal notice and comment procedure May 1983.
Only 14 responses were received as of Sep-that allows potentially affected licensees tember 1983.
Some of these comments (in particular to respond to proposed regulatory guide GE's July 11, 1983 letter) cited a need to specifically changes.
address BWR-related concerns in the RG.
This was done and copies were sent to GE and the BWR Owners Group.
Given the previous extensive distribution of "for comment" reports and regulatory positions and the rather small number of responses, the staff does not plan to reissue RG 1.82, Revision 1 for comment.
The NRC staff will incorporate additional valid technical points received from the BWR Owners Group and GE.
The most recent input from the BWR Owners Group o
E (July 16, 1984) does not provide new significant find-ings; rather this input re-expresses concerns pre-U viously voiced and stresses possible misinterpreta-tions of wording in RG 1.82, Revision 1.
Table 3 (Continued)
E A
y Comment NRC Staff Response to Commonwealth Edison E1 The Commission has not sufficiently justified the A-43 resolution does not mandate. retrofits; rather, i
need to imoose retrofit requirements on either applicants are requested to assess long-term recir-operating or near-term operating license units.
culation capability utilizing RG 1.82, Revision 1 and to then determine what corrective actions may be g
needed. The use of an information bulletin to the majority of the plants does not constitute imposition of a retrofit.
Cost estimates for surveys, design reviews, and The A-43 value/ impact evaluation has been revised retrofitting are questionable.
based on detailed sump blockage probability studies (NUREG/CR-3394) and cost estimates received from industry responses.
O The proposed RG 1.82 is overly conservative.
The NRC staff acknowledges that conservatisms exist However, given the need for assurance that the in RG 1.82, Revision 1.
However, such conservatisms recirculation sump remains a reliable source are prompted by the limited amount of available infor-of cooling water, the commentor agrees that an mation regarding insulation destruction due to high evaluation of sump designs, potential for debris, pressure jets and attendant debris generation, and air ingestion, and adequate net positive suction the wide variability of plant designs and types of head (NPSH) is fully justified.
insulation used.
The commentor questions the assumption that 50%-
A detailed sump failure probability analysis was of LOCAs lead to sump loss, the value/ impact ratio performed and_is reported in NUREG/CR-3394. The given uncertainties in estimated costs, the basis
" averaged" sump failure probability was 2E-5/ reactor-for assuming 23 years remaining plant life, etc.
year with a range of 3E-6 to SE-5/ reactor year.
n l
{
Consumers Power u
)
Regarding the proposed Revision 1 to RG 1.82, the Appendix A of proposed RG 1.82, Revision I was always commentor stated (1) that Appendix A should be intended to provide additional information and/or clearly delineated as being an information and not design requirements.
Appendix A has been clearly guidance source, not as presenting design require-labeled as such.
ments, and (2) that consistency is needed with respect to NPSH terminology.
Table 3 (Continued)
EE 9
Comment NRC Staff Response E.
1 Regarding the value/ impact analysis, the commentor That 50% of LOCAs lead to sump blockage has been a
4 5
questioned the assumption that 50% of the loss-of-reevaluated (see NUREG/CR-3394), and the results li coolant accidents (LOCAs) lead to sump blockage and of that detailed study have been used in revising the-IL cites a sump failure frequency of 2 x 10 4 per A-43 release estimates.
E demand from another probabilistic risk analysis.
4
~
j The commentor questioned the direct application of The calculation of avoided accidents costs, loss-of-core melt frequency reduction for computing avoided plant costs, etc., are consistent with current NRC accident cost.
The commentor disagrees with taking staff evaluation practices.
Recalculation of the credit for loss of plant cost.
Rather, the parameters previously used will be carried out with commentor states that loss-of plant costs should be the revised blockage frequencies.
deducted from avoided accident costs.
1 Creare T
R; The beta factor used to predict a pump's Efforts were made to obtain the original data tapes required NPSH in an air / water mixture is based and calculate the data's scatter; however, this j
on data whose scatter was not reported. The information was not readily available. The suggested NUREG should note this and caution the applicant cautionary note has been added to NUREG-0897.
and reviewer to carefully consider the adequacy of the NPSH margin if it is marginal.
The use of an arbitrary minimum allowable NPSH NUREG-0897 and RG 1.82, Revision 1 no longer recommend margin, either as a fixed value (i.e., 1 foot) a minimum allowable NPSH margin.
Instead, they note j
or as a percentage value (i.e., 0.5 x margin with that whatever NPSH margin is available (after account-no screen blockage), is not justifiable.
It should ing for hydraulic and screen blockage effects) should i
be recognized that what constitutes a safe NPSH be evaluated with respect to each plant's long-term l
R margin is a plant-specific judgment.
recirculation requirements.
e*
h, Diamond Power Company t$
NUREG-0897 resolves a significant safety problem in The NRC staff concurs.
^
El a thorough and equitable manner.
I L
)
j
Ttblo 3 (Centinued)
E Comment NRC Staff Response h
E The commentor provides recommendations regarding The proposed classifications have been combined with-the classification of various insulating materials, other similar proposals to revise and clarify the (f
particularly on the need to distinguish between insulation classification and descriptions used in totally encapsulated insulation and jacketed NUREG-0897.
E insulation.
E The commentor provides listings of the types of The information has been added to NUREG-0897 and insulations purchased since 1980 and the types NUREG-0869, along with data received from other of insulations used in recent retrofittings.
manufacturers.
The commentor states that the costs in the This cost information has been reflected in the value/ impact analysis are in agreement with its revised value/ impact analysis (NUREG-0869), along costs and provides the following figures:
with other industry cost figures.
Cost of MIRROR ** reflective metallic j{
insulation = $40/ft2 for material w
alone.
Installation cost, excluding material
= $25/ hour.
Productivity = 1.24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> /ft2 of insulation.
Reflective metallic insulation is not the Information supplied by Owens-Corning Fiberglass Co.
predominant type of insulation used in newer and the Diamond Power Co. regarding types of insula-plants.
Recently insulated plants mainly tion used in existing and future reactors has been use fiberglass insulation.*
added to NUREG-0897 and NUREG-0869.
These reports EE have been revised to reflect this new information.
EI The trend appears to be toward a higher utilization of
]
{
fibrous insulations.
E$
I 8?
Letter of July 11, 1983.
1 b
'\\.
Table 3 (Continued)
E A9 Comment NRC Staff Response E
S A report on "HDR Test Results on MIRROR This report has been included as Appendix E in
\\
27 insulation performance during LOCA conditions NUREG-0897, Rev. 1.
The results of this report do
\\'
$L was submitted to provide additional information not support a hypothesis which postulates free and
!L to the existing data base used in resolution undamaged inner foils being available to transport 8
of USI A-43.*
at low velocities and to cause blockage.
- However, the limited data base precludes developing a detailed debris generation model.
General Electric Company SRP 6.2.2 and RG 1.82, Revision 1 make no RG 1.82, Revision 1 and SRP 6.2.2 have been modified distinction between BWRs and PWRs; regulatory to identify PWR-and BWR-related concerns and renamed i
^
criteria should differentiate between various
" Water Sources for Long-Term Recirculation Cooling plant designs.**
Following a Loss-of-Coolant Accident."
T Reference should be made to technical findings Based on the responses received, the A-43 technical Z
that imply that A-43 concerns do not pose a findings will be revised to reflect (1) that there is serious problem for BWRs.*
a more extensive use of fibrous insulations (i.e.,
NUKON*) than previously identified and (2) that BWRs are reinsulating with NUKON".
NUREG-0897 will reflect current findings and identify both PWR-and BWR-related concerns.
The value impact analysis utilizes a PWR for GE's point on utilizing a PWR probabilistic risk the risk assessment and PWR-oriented industry assessment for drawing conclusions for a BWR is impacts and, as such, is not directly applicable acknowledged.
Similar assessments have been made for to BWRs.*
BWRs and those results have been utilized in preparing this revised regulatory analysis.
R o
K,
- Letter of December 6, 1984.
- Letter of July 11, 1983.
P-y
- -- ~
Table 3 (Continued)
E=F Comment NRC Staff Response E8
~
General Electric has reviewed the proposed The requirement for long-term 'ecay heat removal is d
y revisions and has concluded that the design
' applicable to both BWRs and PWRs.
RG 1.82, Revision 1 requirements proposed in RG 1.82, Revision 1 Appendix A'contains a series of tables.(or guidelines) are excessively prescriptive and not generically that have been derived from extensive tests and anal-E applicable to the BWR.*
ytical studies.
This information is provided for of referral and can, or'need not, be used--at the s
1 user's option.
RG 1.82, Revision 1 is general, and I
not prescriptive.
The applicant has the responsi-
)
bility for design submittal and justification of the j
safety aspects thereof.
The proposed RG should be revised so that no The technical findings in 1983 (versus earlier findings) further requirements are imposed on designs that
'are considerably different, particularly with respect i
have already taken design precautions that preclude insulation employed currently and the transport char-T air ingestion into, or blocking of, suction lines acteristics of insulation debris.
The air ingestion j
Ifi used for long-term decay heat removal.*
potential has been experimentally quantified and found to be small.
However the 50% blockage criterion in the current RG 1.82 permitted applicants to essen-tially. bypass.the debris blockage question.
For those plants where design crecautions can be clearly demon-i strated, further actions (retrofits) are not necessary.
In addition, the proposed RG should be further The licensee and/or applicant always has the option l
revised to provide for alternative means of to propose alternate means.to deal with a particular i
i ensuring that long-term heat removal is not lost design or safety problem.
[
as a result of suction blocking or air ingestion.*
O In the SER for GESSAR, the NRC indicated that At the time the SER for GESSAR II was written, A-43 i
USI A-43 posed no problem for the Mark III concerns relative to BWRs were still under evaluation.
o
{
containment configuration.*
The staff's SER cited several elements of the e*$
- Letter of October 17, 1983.
i
Table 3 (Continued)
E=?
Comment NRC Staff Response E8 GESSAR II design that tended to reduce the probability
["
for blockage of the RHR suction inlets due to LOCA 1
generated debris.
The staff concluded that plants i
referencing the GESSAR II design could proceed pending 8
resolution of USI A-43 without endangering the health and safety of the public while completing its evalua-e tion of GESSAR.
The unique aspects of each Mark III plant design should be evaluated during plant-specific reviews of A-43 concerns.
The tests performed by Alden Research Laboratory The comment is partially correct, because BWR RHR for Reference 3 may even be very conservative for suction inlets are located at some elevated distance T
BWRs since it appears the tests utilized sump above the wetwell or suppression pool floor.
- However, M
screens directly on the sump floor.*
the insulation debris transport characteristics (see NUREG/CR-2982, Rev. 1) showed that low velocities (i.e., 0.2 - 0.3 ft/sec) can transport fragmented debris and are applicable to both BWRs and PWRs.
The proposed regulatory guide should be revised RG 1.82, Revision 1 states:
"This regulatory guide to include criteria that will allow alternative has been developed from an extension experimenal and measures for precluding loss of long-term analytical data base.
The applicant is free to select decay heat removal due to air ingestion or alternate calculation methods which are founded in blockage.*
substantiating experiments and/or limiting analytical considerations." Thus, the applicant is free to select alternate methods or measures for precluding R
loss of long-term decay heat removal.
o E
Earlier surveys on the use of insulation in light As stated above, current findings do not support the water reactors have concluded that most BWRs utilize earlier surveys or conclusions.
NUREG-0897 is being metallic insulation, which minimizes the potential revised to incorporate findings from public comments 8
- Letter of October 17, 1983.
- - _ ~
Table 3 (Continued)
E A
l(
Comment NRC Staff Response e
for formation and subsequent transport of debris received (particularly with respect to insulations
{
to the sump screens.*
currently used and the change to fibrous insulation I~
from previously used reflective metallic insulations).
4 S
Recent tests on the transport of thin stainless steel foils show that this material can be transported at e
low velocities (i.e., 0.2 to 0.3 ft/sec).
Gibbs and Hill, Inc.
Section B does not discuss the fact that sump Appendix A (page 1-9) has wording very similar to the configurations that differ significantly from commentor's suggested wording.
I the criteria of Appendix A may be equally acceptable.
Gibbs and Hill recommends adding the following concluding paragraph to Section B:
"If the sump design differs significantly from the T
guidelines presented in Appendix A, similar data t0 from full-scale or reduced-scale tests, or in plant tests can be used to verify adequate sump hydraulic performance."
Tables A-1 and A-3 are inconsistent and Table A-2
.The inconsistencies have been corrected.
has inconsistencies in water level noted.
i Northeast Utilities Tests show that gratings are as effective as Gratings were very effective in reducing air ingestion i
solid cover plate in suppressing vortices.
to essentially zero.
1 j
E?
The procedure in Appendix B is too prescriptive.
Appendix B in NUREG-0897 presents the staff's technical The NRC should allow licensees to define and' findings for A-43.
Appendix B was included to illus-o 1
EI develop their own evaluation methods.
trate major considerations.
RG 1.82, Revision 1 is the
[,
regulatory document.
d?
u,
- Letter of October 17, 1983.
+
~. - - _. -,
Table 3 (Continued)
E A9 Comment NRC Staff Response 8
G Credit should be given for top screen area if For those plant designs and calculated plant condi-
~
ff it is deep enough to reduce the potential for tions where this point could be unconditionally lL clogging (RG 1.82, Revision 1, Section C, Item 7).
substantiated, credit would be given.
8 The licensee should be free to determine methods Section 4, Item 14 states:
"The trash rack and of inspection and access requirements (RG 1.82, screen structure should include access openings to p
Revision 1, Section C, Item 14).
facilitate inspection of the structure and pump suction intake."
RG 1.82, Revision 1 will be used to evaluate sumps The need for backfitting will be based on plant-in operating plants.
This may require backfitting specific analyses that will reveal the need for, and the extent of backfitting that might be required.
at substantial costs.
The cost of backfit should be weighed against core melt costs.
T 5;
Appendix A to RG 1.82, Revision 1 requires obtaining Appendix A states:
"If the sump design deviates performance data if sump design deviates significantly from the boundaries noted, similar i
significantly from the guidelines provided.
performance data should be obtained for verification For operating plants, this may result in costly of adequate sump hydraulic performance."
sump testing.
l NRC estimates for man-rem costs associated The value impact analysis has been revised based on with insulation replacement are grossly cost data received during "for comment" period.
underestimated.
The value impact analysis addresses only PWRs.
The value impact analysis revision clearly addresses If the NRC has concluded that this issue only BWR and PWR concerns.
R applies to PWRs, then the document should Ei reflect this.
E The AIF comments are addressed separately; see above.
The commentor concurs with the comments Os submitted separately on this document by the AIF.
__m Table 3 (Continued)
E A
?
Comment NRC Staff Response j
S 5
~
1 Owens-Corning E1 Detailed comments addressed the wide variation Detailed comments received on insulation types; E
of insulations employed, descriptions, suggested descriptions, etc. have been used to revise E
terminology, etc.*
j e
4 Comments recommended including transport and head Data-frow NUKON" tests have been referenced and i
}
loss data for NUKON" fiberglass tests.*
major findings summarized in the revised NUREG-0897.
The commentor questioned Table B-1, Criterion 2, Transport tests on reflective metallic foils were i
that reflective metallic insulation foil debris conducted and revealed that they can be transported would not be transported at velocities less than 2.0 at low velocities (0.2 - 0.5 ft/sec).
ft/sec.*
>O The commentor questioned the concept that if there Inputs received have been used in revising NUREG-0869.
i is all reflective metallic insulation there is no l
problem.*
i Comments on recommended changes to various tables as Inputs received have been used in revising NUREG-0869.
discussed at the June 1 and 2,1983, public meeting.*
1 i
Numerous comments suggesting word changes that would Inputs received have been used in revising NUREG-0869.
j minimize singling out fibrous type insulations as the screen blockage concern without considering blockages due to reflective metallic insulation materials.*
l I
e o
Comments on recommended revision to reflect current Inputs received have been used in revising NUREG-0869.
)
3 status of insulations employed in nuclear power 3
i Pr plants.*
\\
N
?
I I
- Letter of June 23, 1983.
t
Table 3 (Continued)
E Ao a
Comment NRC Staff Response The potential for screen blockage by reflective A set of experiments to determine transport velocities
,'j metallic debris has not been adequately addressed.
(similar to those performed on fibrous insulations) 7 In particular, the water velocities required to has been completed by Alden Research Laboratory.
p to transport debris and hold it against the sump The results are summarized in NUREG-0897 and used in
=
screen have not been studied.*
The assumption that all fibrous blankets and The 7 L/D criterion is based on experimental studies pillows within 7 L/D of a break are destroyed is of representative samples of fibrous pillows exposed overly conservative.
Different designs of pillows of high pressure water jets.
These small water jet have varying resistances to destruction by water studies showed that increasing pressure (40-60 psia) jets."
results in destruction of pillow covers and release of core material.
Furthermore, blowdown experiments in tne German HDR facility showed that fiberglass insulations (even when jacketed) were destroyed within g;
6 to 12 feet of the break, and distributed throughout containment as very fine particles.
Unless conclusive experimental evidence is obtained that accurately replicates the variety of conditions that may exist in a LOCA, it is prudent to retain the conservative 7 L/D criterion.
The 7 L/D envelope is a significant reduction from the previously proposed 0.5 psia stagnation pressure destruction criterion in NUREG/CR-2791 (September 1982) and (in general) limits the zone of maximum destruction to the primary system piping and lower portions of the steam generators.
The commentor stated that estimated costs for OCF cost data are utilized in revisions to the p?
insulation installation and replacement are value/ impact analysis.
g too low.
OCF cost estimates that were provided g
are*
t$$
- Letter of July 14, 1983.
r Table 3 (Continued)
E
?
Comment NRC Staff Response 8
Cost of NUKON* = $90/ft2 for material
~
5 (as fabricated) 5.
IL Cost of reflective metallic = $100/ft2 9
for material (as fabricated)
Installation cost = $112/ft2 for labor and related support The commentor provided recommendatic.
Descriptive classifications provided for insulation
~ ' -
classification of various insulating m _'.eriais types have been combined with similar classifications stressing differences between NUKON" (an OCF obtained from Diamond Power Company for inclusion in product) and other fiberglass and mineral NUREG-0897, Revision 1 and NUREG-0869, Revision 1.
wool materials. The commentor noted the
>4 differences between NUKON" and high density fiberglass.*
The commentor identified 14 reactor plants that OCF plant information have been utilized, along with have been reinsulated with NUKON", are in the infonnation from Diamond Power Company, to develop process of installing NUKON", or may install a current picture of insulation utilization in nuclear NUK0N*.*
power plants.
The major finding is that the number of plants using or are planning to use fibrous insu-lation is larger than previously estimated.
For example, the Diamond Power list reveals that 25 of 130 operating and projected plants are utilizing fibrous insulation on primary system components, The commentor recommended inspection surveys of The recommendation for physical plant surveys (or o
?
plants to identify actual insulations employed and inspection to identify types and quantities insula-hE recommended the modification of a draft generic tions employed) is a good one.
However, the use of letter to include this requirement.*
a generic letter is to reconfirm adequate NPSH U8
- Letter of July 14, 1983.
Table 3 (Continued)
E iR 5'
Comment NRC Staff Response E$
margins, and will be based on the actual types and j"
quantities of insulation employed within a given plant
{
without imposing a need to report in detail.
((
A report on "HDR Blowdown Tests with NUKON This report has been included as Appendix F in Insulation Blankets" was submitted as a NUREG-0897 Rev. 1.
The tests demonstrated that supportive document for the capability on jacketed and unjacketed NUKON" blankets within 7 L/D NUKON" insulation to withstand the impact of will be nearly totally destroyed.
However NUKON" a high pressure steam-water blast.**
blankets enclosed in standard NUKON* stainless steel jackets withstood the blast better.
But these were insufficient number of tests to draw conclusions for similar insulations.
Power Component Systems, Inc.
2, I
4 A report on " Buoyancy, Transport and Head Loss The formula provided for fibrous debris blockage Characteristics of Nuclear Grade Insulation head loss is included in Section 5 of NUREG-0897, Blankets"was submitted as a supportive document Rev. 1.
for relative efforts in the area of fibrous insulations.***
1 2
?
o K,
t U
CD
- Letter of July 14, 1983.
- Letter of February 18, 1985
- Letter of November 8, 1984
,~,
APPENDIX B BACKGROUND AND
SUMMARY
OF MINUTES OF MEETINGS OF COMMITTEE TO REVIEW GENERIC REQUIREMENTS (CRGR)
REGARDING UNRESOLVED SAFETY ISSUE (USI) A-43 RESOLUTION (CRGR MEETINGS NOS. 26, 28, AND 66)
NUREG-0869, Revision 1 October 1985
s BACKGROUND AND
SUMMARY
OF MINUTES OF MEETINGS OF COMMITTEE TO REVIEW GENERIC REQUIREMENTS REGARDING UNRESOLVED SAFETY ISSUE A-43 RESOLUTION CRGR MEETING NOS. 26, 28, AND 66)
BACKGROUND The staff's proposed resolution of Unresolved Safety Issue (USI) A-43, " Con-tainment Emergency Sump Performance," was sent to'the Committee to Review Generic Requirements (CRGR) on October 27, 1982 and was discussed in meetings with CRGR on November 24, 1982 and December 21, 1982. The, December 21, 1982 CRGR minutes state that CRGR' agrees with the staff's findings and proposed changes to Standard Review Plan Section 6.2.2, " Containment Heat Removal Systems," and Regulatory Guide 1.82, " Sump for Eme'rgency Core Cooling and Containment Spray S) items." However, CRGR agrees only with " forward fit" implementation.
The CRGR minutes cite the Deputy Executive Director for Regional Operations and Generic Requirements (DE090GR) staff analyses. Tne CRGR questioned four key assumptions in the Office of Nuclear Reactor Regu-lation (NRR) calculations of averted public dose and stated that the DEDR0GR staff fem. that the dose was high by a factor 'of 100.
In conclusion, the CRGR recommendedithat the NRR staff review these risk-reduction calculations, re-affirm or rev,ise the proposed backfit actions, and then meet again with, CRGR.
In response to the CRGR recommendations, the staff made additional calcu-lations to estimate the frequency of large loss-of-coolant accidents.
These calculations were based on a detailed piping-and-break probability analysis and estimates of the percentage of these breaks thattcould lead to sump screen blockage.
The results of these calculations are in NUREG/CR-3394, which was published in July 1983.
These findings, along with public comments received during the for comment period for USI A-43 (May-June 1983), were used in revising NUREG-0897 and NUREG-0869.
A third meeting was held with the CRGR on July 11, 1984.
The summary minutes of Meeting No. 66 pertaining to USI A-43 are those noted as Enclosure 3 to the minutes for CRGR Neeting No. 66, which are included in this appendix.
CRGR's views are as notediin this enclosure.
After the July 11, 1984 meeting, the staff again revised the proposed resolution of USI A-43.
s Y
c i
r q
t NUREG-0869, Revision 1 B-1 October 198,5 s
l
SUMMARY
OF CRGR MEETING NO. 26 (November 24,1982)*
The CRGR met on Wednesday, November 24,1982, from 1:00 - 6:00 p.m.
S. Hanauer, NRR, presented for CRGR review the NRR recommendations to resolve USI A-43, Containment Emergency Sump Performance.
The overall safety concern embodied in USI A-43 is related to the capability of the containment emergency sump to provide an adequate water source to sustain long-term recirculation cooling following a large LOCA.
The problem can be subdivided into (a) sump hydraulic performance, '5)
LOCA generated debris effects, and (c) recirculation pump performance under post-LOCA conditions.
Each has been studied by NRR and the technical findings are reported in NUREG-0897 and associated references.
With this view, NRR proposed the following actions:
(1) Revise the NRC Standard Review Plan (SRP) Section 6.2.2, " Containment Heat Removal Systems," and Section 6.3, " Emergency Core Cooling Systems."
Issuance of the proposed revisions to the SRP is needed to correct previous sump review criteria that are not supported by current findings from full-scale sump tests and generic plant studies (i.e., ju gment of sump hydraulic acceptability principally on vortex formation.
(2) Revise Regulatory Guide (RG) 1.82 to reflect the findings in NUREG-0897, " Containment Emergency Sump Performanca," to incorporate the results of 2 years of sump testing and generic plant studies and to correct deficiencies such as the 50% screen blockage criterion.
Generic plant calculations addressing LOCA generated debris effects have shown that the 50% blockage criterlon can be excessive in some plants and nonconservative in other plants.
Continued use, without revision, of this regulatory guidance would permit the sump designer to bypass the need to assess debris blockage effects and to continue to show that a 50% blocked screen does not result in excessive head loss.
Appendix A has been included in the proposed revision to RG 1.82 to provide guidance and criteria for assessing sump hydraulic performance, LOCA-induced debris effects, and pump. performance under adverse conditions.
A combined consideration of these three aspects is necessary to determine overall sump performance and acceptability with respect to assurance that adequate pump NPSH margin will exist.
(3) Operating plants should assess the extent of debris blockage potential and, based on the outcome of plant assessments, action should be taken to modify the cump screens or to replace all fibrous insulation with encapsulated insulation.
" Verbatim copy.
t NUREG-0869, Revision 1 B-2 October 1985
y i
The Committee commended the staff for the thorough technical analysis described in NUREG-0897 and agreed with recommendations (1) and (2) above, which reduce requirements on future OL applicants.
In support of recommend-ation (3), NRR presented cost-benefit analyses which showed the benefits (using $1000 per person-rem averted) outweighed the costs of the proposed requirements in (3) for operating plants.
The Committee suggested that the benefits (reduction in core melt probability) appeared to be overstated by at least a factor of 10, and perhaps 100, and that the costs appeared to be understated.
Thus, it was not clear to the Committee that recommendation (3) couldbejustifiedonacost-benefitbasis,eventhoughitwasacknowledged to be good engineering practice to replace unencapsulated fibrous insulation with encapsulated insulation.
In respoi.ce to a question whether the staff has considered the effects of paint debris on sump performance, NRR said they had not considered ft in the context of USI A-43, but they agreed to review what consideration had been given to paint debris in previous staff reviews.
The Committee decided to discuss USI A-43 in a subsequent meeting after information on the potential effects of paint debris has been received from NRR.
i NUREG-0869, Revision 1 B-3 October 1985
SUMMARY
OF CRGR MEETING N0. 28 (December 21,1982)*
The CRGR met with respresentatives of NRR to further pursue questions regarding USI A-43 Containment Emergency Sump Performance.
The CRGR, during Meeting No. P.6, had questioned the potential for sump blockage due to paint removed from containment surfaces during a LOCA.
The question of the potential for sump blockage due to paint removal and transport to the sumps was addressed in a memorandum from H. Denton to V. Stello dated December 16, 1982.
The NRR position on the paint blockage issue was that:
(1) Analyses indicate that there is not a basis for concern as a generic safety issue; (2) The issue will be further evaluated within established NRR procedures for treating proposed new generic issues, to determine the priority for further evaluation; (3) The possible issue of paint removal therefore should not delay obtaining industry and public comment on the defined A-43 issue.
The CRGR accepted the NRR position on the paint blockage issue.
THe CRGR addressed the level of risk reduction, or benefit, to be obtained from the analyses and potential modifications proposed to be required of the several licensees that might be found to have combined insulation / sump designs that could lead to failure of long-term recirculation cooling.
The Committee (as reflected in the minutes of CRGR Meeting No. 26 November 24,1982) has agreed with the forward-fit aspects of the NRR proposed requirements.
A revised Standard Review Plan Section 6.2.2 and a revised Regulatory Guide 1.82 would incorporate changes in design criteria that would provide greater assurance of sump performance, but would be imposed only on Operating License and Construction Permit applicants filing Final or Preliminary Safety Ar.alysis Reports at some time after ths effective dates of the revised Standard Review Plan Section and the revised Regulatory Guide.
To support the proposed backfit requirements, NRR provided a generic value/
impact assessment comprised of a probabilistic risk analysis of the effects of loss of sump function and estimated costs of the backfit requirements proposed for licensees to reduce the risks of such loss.
The probabilistic risk analyses resulted in an expected value of offsite public dose (person-rems) that could be averted from the estimated six to ten plants that are expected to need modifications.
Key assumptions in this NRR analysis are:
" Verbatim copy.
NUREG-0869, Revision 1 B-4 October 1985
(1) The expected value of large LOCA (greater than 6" diameter pipe) incidence is 10-4 per reactor year.
(2) For.those plants having fibrous insulation that could potentially result insumpblockage,.itisassumedthat50%ofallLOCAsinpipinggreater than 6' diameter will result in complete failure to pump any water from any containment sump.
(3) The assumed failure of recirculation flow (from sump) is assumed to conditionally fail both reactor building spray and emergency core cooling, thereby leading to a core melt with containment failure overpressure. -No credit was given for potential beneficial opera.by tor action to prevent sump blockage by throttling the emergency core cooling system pump or to utilize alternate water sources and systems to prevent either core melt or loss of containment function.
Thus, for the class of plants above, the NRR analysis assumed the core melt frequency for this LOCA sequence is 5 x 10-5/Rx-Yr.
(4) The offsite consequence model used to predict expected values of population dose assumed an average site, a 50-mile radius, and no evacuation of population during the accident.
An analysis by the DEDR0GR staff indicated that each of the assumptions above was probably too conservative and that the NRR predicted value of averted public dose of about 65 person rems per plant per year was too high by a factor of at least 100. -If this were indeed the case, the proposed
~
implementation plan actions would not appear to be justified.
The CRGR recommended that NRR review its risk reduction analysis in light of the analysis performed by the DEDR0GR staff with the objective of developing the
. most realistic assessment of averted radiological dose.
NRR should then reaffirm or revise the proposed backfit actions, and discuss with CRGR again if they believe the cost benefit analysis justifies the proposed backfit actions.
NUREG-0869, Revision 1 B-5 October 1985
SUMMARY
OF CRGR MEETING NO.66 (July 11, 1984) to the Minutes for CRGR Meeting No. 66 CRGR REVIEW 0F THE PROPOSED RESOLUTION TO
-UNRESOLVED SAFETY 155UE A-43
- CONTAINMENT EMERGENCY SUMP PERFORMANCE"
-The NRR Division of Safety Technology representatives T. Speis, F. Schroeder, K. Kniel and A. Serkiz presented the proposed resolution for CRGR review.- The package' submitted for review was transmitted by a rhemorandum dated June 14, MS! f.om H. Denton to V. Stello, Jr; it included the following docurrents:
1.
Sumaries of USI A-43 References.
-2.
NUREG 0897, Revision 1, March 1984, describing the technical findings of the effort..
3.
- Regulatory Guide 1.82, Rev.1, May 1984, " Sump Design and Water Sources for Emergency Core Cooling."
4 Standard Review Plan Section No. 6.2.2, Revision 4 " Containment Heat Removal Systems."
5.
.NUREG 0869, Rev. 1. USI A-43 Regulatory Analysis, containing a value/ impact analyses, suninary of public coments received and action taken,- and a proposed generic letter for implementation of R.G.1.82, Rev. 1.
6.
Draft Generic Letter, subject: " Assessment of Available NPSH Margin.for.
Long Tem Cooling."'
7.
Note to A. Serkiz from 8. Shields, April 3, 1984,
Subject:
Generic Letter-on Containment Emergerty Sump Perfomance.
The NRR presenters at the CRGR meeting also provided a handout titled "USI j
A-43, Containment Emergency Sump Performance," which is attached.
CRGR was requested to recomend to the EDO approval of the following final actions:-
1.
I,ssue NUREG-0897, Revision 1 as the technical findings for resolving USI A-43; 2.
Issue Regulatory Guide (RG) 1.82 Revision 1 and SRP Section 6.2.2, Revision 4 for guidance and use in the OL and CP review cycle as part of the normal review process.
3.
Issue a generic letter to all LWR licensees and applicants pursuant to 10 C;1 50.54(f).
This letter would request a plant-specific assessment of NUREG-0869, Revision 1 B-6 October 1985
post-LOCA debris bl.ockage effects (using guidance provided in Appendix A of RG 1.82, Revision 1) on net positive suction head (NPSH) 6.argin, and would request the responders to submit the calculated NPSH margin available, and a description of any plant modifications shown to be necessary by this assessment.
The staff did not propost a generic solution to the variety of potential deficiencies that were postulated to exist among licensees.
The proposed generic backfit requirement was for licensees to complete an assessment and report to NRC using the staff specified criteria.
Further staff / licensee interaction to define and approve acceptable solutions to design / operational deficiencies would be pursued on a plant-specific basis.
Five to 20 plants were expected to be in this category.
The safety rationale for the proposed requirements was that the containment recirculation mode of long term decay heat removel following a LOCA must be assured. The proposed backfit would not provide substantial additional protection over that thought to exist, but would assure a level of safety previously thought to exist. The backfit could in most, cases be limited to analysis to verify the sump / pump performance. Staff rey pw and followup actions would be limited to those plants with identified ~ problems. CP and OL applicants would be required to demonstrate adequate design margin for NPSH during the licensing review, in accordance with the SRP and Regulat6ry Guide revisions presented in the CRGR package.
The staff proposal originally presented to the CRGR at two meetings in November and December 1982 was issued for public coment in early 1983. The current proposal includes revisions made as a result of public coment.
In addition, a more sophisticated analysis was completed-to assess the likelihood of sump blockage than was cresented in support of the initial proposal in late 1982.
The current propor,a1 was presented as based on (1) A revd.ew of expected LOCA probability as a function of pipe size, wels' type, and joint configuration.
(2) An examination of the Salem-1 plant containment layout and selection of 238 locations in pipe as those expected to represent all LOCAs significant with respect to then current staff guidelines for selection of postulated breaks in high energy pipe within containment (SRPSections6.2).
(3) A mathematical model describing the volume of insulation removed by LOCA as a function of break size and location relative to adjacent insulated pipes and vessels.
(4)
Investigations of the'liblihood of transport of stripped insulation to the sump.
Five operating plant layouts were modeled analytically to evaluate the water v4'iocity through expected pathways for transport of insulation.
NUREG-0869, Revision 1 B-7 October 1985
(5) Results of experiments with various insulation materials to detemine minimum water velocities necessary to transport typical insulation debris.
(6) Calculations of-pressure loss across typical sump screens due to insulation deposition on the screens.
i A value-impact analysis was presented based on the calculations of sump failure probability and cost data, much of which was supplied by industry during the sump failure probability, based on an SAI. report 1pssumed to be identical wit public consnent period. Core melt probability was completed in September 1982. To calculate net safety benefit from reducing this risk, the assumption i
was also made that given a core melt, the containment fails with a probability of 1.
This containment failure assumption led to offsite radioactivity
. releases commensurate with the PWR 2 release category of WASH 1400, Calculation of offsite dose comitment in person-rem was done using the CRAC j
L code and a si.te population and meteorology selected to represent an average of all. plants.
j The value-impact results proguced by the staff are suninarized on page 9 of the attachment to this enclosure. The staff pointed out that the value/ impact ratio was only moderate and that for most plants no risk reduction was anticipated.
However, the staff's position is that the deteministic licensing bases of 10 CFR 50.46 (b)(5) mandated the confirmatory analysis steps proposed in the draft generic letter. An additional concern voiced by the staff was the observed recent tendency of utilities to change insulation materials to types that are more likely to block a sump.
The following major points were made 'inTthe discussion of the staff proposal at this meeting:
1.
-The application of the initiating event probability data in Table 1, Appendix B, page B-4, NUREG 0869, was considered unclear by the CRGR. The application of the data was called into question because the specification of 238 postulated break locations seemed to be a low number of locations relative to the total length of piping in containment subject to a high pressure LOCA. The tabulated probabilities as a function of pipe size g
g were intended to represent the population of all pipe in containment i
9 i
subject to LOCA failures.
1 Further, the Comittee was concerned that the selected break locations were such that the probability of pipe displacement to give the postulated l
double ended or even single ended flow model used to predict insulation removal was much lower than the probability of pipe rupture given in the 1/ Ferrell, W.L., et al, "Probabilistic Assessment of USI A-43", Science Applications Inc., September 1982
~
Table on Page 9 of attachment is duplicated on Page 8-11 NUREG-0869, Revision 1 B-8 October 1985 7
.q i
tabulated data. The tabulated data is representative of a data base that includes all kinds of disruptive failures in pipe, not just the sudden L
explosive rupture with deflection at the break necessary to provide a full flow area blowdown..
t 2.
'The CRGR believed that there was considerable but undescribed conservatism in the model for determining, given a pipe break of full flow area, how much insulation is torn off a designated target area or volume.
Two such i
conservatisms would be:
a.
The reduction in (initiating event) probability due to the selection of pipe break direction.such that the maximum amount of insulation is targeted.
b.
The use of a hemispherical volume of radius 7 x the pipe diameter
[
oriented to contain the most target insulation, and the assumption that all insulation within that volume is stripped from the component l
,even if, as in the case of a steam generator, the "back" side of the component is not impinged directly by the break jet flow.
4 3.
The CRGR noted that there was considerable conservatism in the conclusion i
. drawn from the analyses of five plants that all insulation debris torn off components would be transported to the sump.
NUREG/CR 2791, where the i
analyses are reported, reports that in four plants insulation debris arriving at the sump would not approach amounts necessary to significantly i
degrade sump operation.
In one plant, Maine Yankee, material might adversely affect the sump based on the assumption that all meterial i
completes the trip to the sump, a scenario which seemed unlikely to the l
report's authors.
I i
4 The CRGR noted further conservatisms in considering the effects of material at the sump, even given that material. arrives at the sump in quantities necessary to cause sufficient screen blockage to unacceptably reduce the net positive suction head (NPSH) at the recirculation pump.
l Two such conservatisms were noted:
(a) The assumption that insulation debris distributes unifomly over the i
sump screens, (b) The assumption that recirculation flow in the range 6 to 10 thousand gallons per minute is the appropriate design flow requirement for the sump screens.
It was stated in the meeting that most plants have i.
flow instruments in ECCS recirculation lines, so operators could be expected to reduce pump flow to the minimum required to cool the l
reactor core after vessel refill innediately following blowdown after a large LOCA.
This flow requirement is likely to be in the range of 300 to 1000 gallons per minute, considerably less than the flows postulated in the NRR proposal.
Since sump blockage according to the NRR presentation is not expected in less than 1-2 hours, the lesser cooling flow requirements are highly likely.
NUREG-0869, Revision 1 B-9 October 1985
.~
i 5.
The CRGR noted prob,able conservatism in the postulated WASH 1400 release categories used in the analysis.
It was suggested that the release source
. terms associated with an early, energetic, above ground containment failure such as the WASH 1400 PWR-2 category would be excessively conservative.
Evidence pointed to at most a delayed ucore meltdown with eventual core melt through the containment base mat, resulting in releases no higher than that associated with a PWR 5 or 6 release.
4 6.
The CRGR overall consensus was that given the questions raised about the assumptions and/or levels of conservatism of the analyses, NRR's position that the proposals based on the analysis are of only moderate benefit versus cost importance, the probability that further clarification or
~
resolution of CRGR concerns umy suggest a lower safety benefit than does the current proposal, and the estimated applicability of the proposed 4
solution to only a few plants, all combine to argue against approval of a backfit requiring all licensees to expend resources to demonstrate that their designs are acceptable.
i-Conclusions n
4 The Committee concluded the following as a result of its review of the material-transmitted and the meeting discussion:
1.
The proposed requirements package was rated of medium importance by NRR.
CRGR review of the information presented on the safety benefit to be achieved resulted'in uncertainty about the validity of the' analyses, with
.the possibility that the proposed potential risk reductions may be far overestimated and would apply to relatively few plants.
2.
The proposed backfit requirement on operating reactors to analyze containment sump performance and report to the NRC should not be promulgated at this time ~due to the uncertainties raised in item 1 above.
l 3.
The CRGR would be willing to reconsider this proposal or a modified proposal by NRR at a regularly scheduled meeting after receipt of
-responses to the i,ssues raised at the.neeting and discussed in the meeting 7
3 sunnary.
i NUREG-0869, Revision 1 B-10 October 1985
VALUE/ IMPACT OVERVIEW PLANTS W/0 (CATEGORY A):
$45K/ PLANT ANALYSIS COST
=
AVERTED RELEASE =
0 MAN-REM /Rx PLANTS W/SOME BACKFITS (CATEGORY B):
$ 65K/ PLANT Af.hYSIS COST
=
$300K/ PLANT BACKFIT COST
=
$365K/ PLANT PLANTS W/ INSULATION REPLACEMENT (CATEGORY C):
$ 85K/ PLANT ANALYSIS COST
=
$820K/ PLANT BACKFIT COST
=
$905K/ PLANT -
AVERTED RELEASE =
650 MAN-REM /RX INDUSTRY. DISTRIBUTION:
AVERTED ASSUMED NO.
COST RELEASE V-I RATIO CATEGORY OF PLANTS
($M)
(MAN-REM)
(MAN-REM /sM)
A 90 4.1 0
0 l
B 15 5.5 9,750 1,770 C
__ji
_.!LJ 3,250 720 i
TOTALS:
110 14.1 13,000 920 NUREG-0869, Revision 1 B-11 October 1985
l i-l l
SUMMARY
OF CRGR MEETING NO. 80 (SEPTEMBER 9, 1985) t j to the Minutes of CRGR Meeting No. 80 U5I A-43 containment Emergency Sump Performance d
Dr. T. Speis and A. Serkiz of NRR presented for CRGR review a proposed final resolution to USI A-43. The CRGR was requested to recomend several actions i
pursuant to resolving USI A-43:
l i
1.
Issue the staff's technical findings in NUREG-0897, Revision 1 for use as en infomation source by applicants, licensees, and the staff in i
addressing the design and operation of containment emergency sumps and BWR RHR suction intakes.
2.
Issue Regulatory Guide 1.82, Revision 1, to include the technical findings reported in the new NUREG-0897 Revision 1.
This would provide improved regulatory staff positions as guidance for assessment of sump performance and BWR RHR suction intakes, including debris blockage effects.
3.
IssueNRCStandardReviewPlan(SRP)Section6.2.2, Revision 4,to I
incorporate the guidance provided by the revised Regulatory Guide 1.82 and the technical findings in NUREG-0897, Revision 1.
4.
Issue a generic letter to all applicants and li:ensees outlining the i
potential for safety concerns related to post-LOCA sump blockage and the l
fact that the original Regulatory Guide 1,82 (Revision 0) is inadequate in light of the more current information resulting in Revision 1 to Regulatory Guide 1.82.
f A copy of the viewgraphs used for the A-43 proposal is attached. The proposed A-43 reso' ution discussed was a revision to a previously presented proposal which was discussed at CRGR meeting nurtber 66 on July 11, 1984. The new proposal was to take actions to assure that the l
technical findings and new guidance are available to the nuclear industry, to advise the industry of the inherent safety benefit to be gained by using the new guidelines, to state the staff's position that this new guidance will be used by the NRC only in new CP reviews and certain standardized design reviews, and to recommend that the staff's technical j
findings on A-43 should be considered when insulation is replaced at operating plants.
j Revisions in the prior regulatory analyses (discussed at meeting number 66) were discussed at this meeting in support of the proposed resolution. These are highlighted briefly:
i 1.
Credit for operator action to recover ECCS flow that may have been lost is
.now given as a 50 pe* cent likelihood that the operator will detect and mitigate a loss of flow given a sump blockage event.
2.
Plants of varying containment types and having different accident I
mitigating systems were evaluated to better define the risks inherent in each design type. Expected values of offsite consequences and i
i i
NUREG-0869, Revision 1 B-12 October 1985
i value/ impact ratios were calculated for all the plant types considered and were generally such that much more than $1000/ person-rem would be required to effect safety-beneficial changes in the plants.
In addition, the conservatism inhere 1t in pipe break probabilities used in the regulatory 1
i analysis was recognized in the regulatory analysis and discussed in the CRGR meeting.
Pipe break data used in the analysis has since been superseded by advanced fracture mechanics analysis and experiments. The more recent work shows that breaks in ductile pipes larger than eight inches in diameter, the size rar.ge necessary to provide a significant sump blockage probability, may occur at frequencies that are several orders of 4
i magnitude less than the frequencies used in the A-43 analyses.
The CRGR decided to reconnend approval of the proposed A-43 resolution presented, with several specific comments on changes that should be incorporated in the various documents:
1.
The more recent work on fracture-mechanics resulting in lower estimates of pipe-break frequency should be explicitly recognized and referenced in the summary section of the regulatory analysis.
I 2.
Implementation wording in the Regulatory Guide and SRP section should be modified to clearly show that the NRC staff use of the new review material will be forward-fit only.
1 3.
The generic letter should clearly state that NRC application of the new i
guidance to an operating plant, particularly with respect to the NRC staff t
reviews of licensee 10 CFR 50.59 reviews, will be treated by the NRC staff as a plant-specific backfit action pursuant to 10 CFR 50.109.
i i
l i
l i
NUREG-0869, Revision 1 B-13 October 1985
,-,r,
.-y--n--,.
-,-,v-r y
-g---
w w
.n-wn-~,-----
--,,w~mc-w,.-mn, vnwe--,-
,,-w.,--w w-,
-w sv,n,vewe,.e-a,--
w
i REFERENCES
]
H. R. Denton to V. Stel,10, Jr., memorandum dated October 27, 1982, FCRGR Review of Proposed Revisions to SRP Section 6.2.2 and RG 1.82-and the j
Supporting Technical Information Document NUREG-0897, as related to USI j
A-43, ' Containment Emergency Sump Performance.'
~
V. Stello, Jr. to W. J. Dircks, memorandum dated December 10, 1982,
" Minutes of CRGR Meeting Number 26."
H. R. Denton to V. Stello, Jr., memorandum dated December 16, 1982,
" Potential Sump Screen Blockage Due to ' Paint Sheets' (Ref. CRGR Meeting 3
of 11/24/82 on USI A-43) "
V. Stello, Jr. to W. J. Dircks, memorandum dated January 11, 1983,
" Minutes of CRGR Meeting Number 28."
H. R. Denton to V. Stello, Jr., memorandum dated February 28, 1983,
" Response to CRGR Comments on USI A-43."
f V. Stello, Jr. to W. J. Dircks, memorandum dated July 24, 1984 " Minutes of CRGR Meeting Number 66."
J H. R. Denton to W. J. Dircks, memorandum dated August 20, 1984, " Feedback l
and Closure: CRGR Meeting Number 66 (RE: Proposed Resolution of LSI A-43)."
V. Stello, Jr. to W. J. Dircks, memorandum dated September 13,198S,
[
" Minutes of CRGR Meeting Number 80."
i i
t 1
i i
i 1
1 NUREG-0869, Revision 1 0-14 October 1985 f
+-e-4.-,...<.--,...---e.---_-...,_.~.-e--.%,--.---------------me.---.,,
APPENDIX C ESTIMATION OF PIPING FAILURE PROBABILITY l
l NUREG-0869, Revision 1 October 1985
ESTIMATION OF PIPING FAILURE PROBABILITY For USI A-43 evaluations, it was necessary to consider a large number of potential break locations over a wide range of piping sizes used for an actual reactor coolant system design.
This, in turn, necessitated assigning piping failure probabilities to each of the potential break locations and piae sizes considered.
The assigned pipe failure probabilities shown in Taale 1 were used as basic inputs for subsequent computerized calculations that weighted these values by the number of welds and the number of piping sections.
The calculations were done for particular diameter size cate-gories and for the plant piping designs utilized (see NUREG/CR-3394).
The pipe failure (rupture) probabilities shown in Table 1 were derived from the assessments presented by Dr. S. H. Bush in his October 1977 paper entitled, " Reliability of Piping in Light Water Reactors," IAEA-SM-218/11.
This paper included assessment of the validity of piping failure prob-abilities cited from various world-wide literature sources and their applicability to nuclear systems.
Bush also evaluated the safety signi-ficance of reported failures in nuclear piping systems and presented in-formation that would allow an estimate to be made of the relative probability of severance because of (1) general faults (e.g., inadequate pipi g flex-ibility) and (2) various components (e..
straight piping runs, oints tees,andelbows),dependingontheirsze,andpotentialcrackorentation.
Using those world-wide sources and failure experiences deemed relevant to nuclear piping systems, Bush reached the following conclusions:
(1) The failure probabilities for larger sizes of nuclear piping are considered to be in the range of 10 4 to 10 8 per reactor year, exclusive of intergranular stress corrosion cracking (IGSCC).
(2) Small pipe sizes of lesser safety significance have much higher failure rates.
(3) In boiling water reactors (BWRs), IGSCC can cause failure rates much higher than 10 4 per reactor year (10 2 per reactor year) in piping 4 to 10 inches (102 to 205 mm) in diameter.
(4) Suggested failure mechanisms apply in most instances exclusive of IGSCC.
(5) Catastrophic failure would appear more likely from operator error or design and construction errors (water hammer, improper handling of dynamic loads, and undetected fabrication defects) rather than conventional flaw initiation and growth or fatigue.
Utilizing the failure information that Bush deemed relevant to nuclear plants, an overall piping failure probability of 3 x 10-4/ reactor year can be derived from his paper for pipes greater than or equal to 3 inch diameter.
NUREG-0869, Revision 1 C-1 October 1985
TABLE 1 Piping failure probability estimates Pipe Failure Weld failure probability distribution size Diameter Probability Weld Weld Weld (inches) class (1/Rx-Yr) type 1 type 2 type 3 D
J P
W N
N j
n a
e 2 to <6 1
3x10-5 0.7 0.15 0.15 6 to <10 2
4x10~5 0.5 0.30 0.20 i
10 to <16 3
3x10-5 0.5 0.30 0.20 16 to 28 4
3x10-6 0.5 0.30 0.20
>28 5
3x10-6 0.5 0.30 0.20 j=3.76x10-4/Rxyr IP Weld Type 1 = fabricated and non-standard joints WeldType2=highrestraintjointsandteeswithjoints WeldType3= elbows, reducers,andstraightpipingrunswithjoints NUREG-0869, Revision 1 C-2 October 1985
I This value is consistent with pipe failure estimates assigned in WASH-1400 l-for piping in the size range of more than 2 inches to less than 6 inches in diameter.
The Bush paper can also be used to develop a failure probability i
distribution as a function of pipe diameter of about 4 x 10-4/ reactor year l
frequency, with the relative distributions (as an approxin:.te percentage value)being 6 inches to < 10 inches diameter 13%
10 inches to < 16 inches diameter 10%
16 inches to 23 inches diameter 1%
(
Furthermore, the Bush paper indicates that circumferential cracks (if they exist) would be expected to be of greater significance (by about a factor of
- 2) than axial cracks relative to the rupture probability.
High restraint, fabricatedjointswouldalsobeexpectedtomakeahighercontributiontothe overall rupture probability than would straight runs of pipe.
Therefore the abovepercentagescanbefurtherredefined,thistimeasafunction'ofpIping joints,etc.
The results of these types of considerations are reflected in the piping and joint failure multipliers shown in Table 1, which were utilized in the USI A-43 sump blockage assessments (see Appendix 0).
In his paper Bush also observed that a well planned program of periodic inspection should dramatically reduce the probability of catastrophic failure of piping, and he cites various studies that have suggested that inspection benefits could result in reducing estimated failure frequencies by as much l
as 1 to 3 orders of magnitudes.
Experimental and analytical work based on mechanistic fracture mechanics that has been done after the 1977 assessment by Bush (see, for example NUREG-1061,Vol1, August 1984)alsoindicatesthattheruptureprobabilityof 4
large-size ductile piping (unaffected by IGSCC) could also be significantly less than the assigned values of Table 1, perhaps by about several orders of magnitude.
The term " leak before break" has been coined from this later work.
The sump blockage assessment performed toward resolution of USI A-43 did not, however, give any explicit probabilistic credit for this " leak I
before break" cor. cept.
References Bush S. H., " Reliability of Pipin Atomic Energy Agency, IAEA-SM-218/g in L10ht-Water Reactors," International 11, October 1977.
l U. S. Atomic Energy Commission, WASH-1400, " Reactor Safety Study," October 1975(alsoissuedbyNRCasNUREG-75/018).
U. S. Nuclear Regulatory Commission, NUREG-1061, " Report of the USNRC Piping Review Committee Investigation and Evaluation of Stress Corrosion Cracking inPipingofBollingWaterReactorPlants,"Vol1, August 1984.
l l
NUREG-0869, Revision 1 C-3 October 1985
APPENDIX D ESTIMATION OF PWR SUMP FAILURE PROBABILITY NUREG-0869. Revision 1 October 1985
ESTIMATION OF PWR SUMP FAILURE PROBABILITY SUMP FAILURE Sump failure is defined as a loss of pressurized water reactor (PWR) sump (or boiling water reactor (BWR) suction intake) capability to provide an adequate water source and net positive suction head (NPSH) margin to the residual heat removal (RHR) and containment spray system (CSS) pumps during the period after a loss-of-coolant accident (LOCA) because of the effects of debris blockage.
Stated another way:
does the head loss across a debris-blocked screen or suction strainer exceed the NPSH margin available under zero blockage conditions?
SUMP FAILURE PROBABILITY The sump failure probability (because of debris effects) is a function of:
(1) The probability of a pipe break or weld failure, because the LOCA is the initiating event that can destroy insulation.
(2) The potential break sizes and locations within containment with respect to other piping and insulated primary system components (e.g., steam generators pressurizer, pumps,safetyinjectiontanks)becausethe expandingjetwilldestroyinsulationthatfallswithinthejet expansion envelope.
(3) The ty9es and quantities of insulation employed, because blockage effects will vary with insulation type (i.e., fibrous debris versus metallic insulation debris) and the location of such insulation.
(4) The containment layout and sump (or suction inlet) location, which can control debris transport.
(Can the sump be directly targeted by the break jet, resulting in prompt transport, or does the debris transport occur later because of recirculation flow drag?).
(5) The size of debris screens (or suction inlet strainers), because larger screens can accommodate larger quantities of debris without incurring large head losses.
(6) Post-LOCA recirculation flow and pump NPSH requirements, which determine whether a blocked screen situation will result in loss of NPSH margin and pumping capacity.
Thus, arriving at an estimated sump failure probability becomes a complex and plant-specific evaluation based on:
(1) probabilistic estimates (i.e., pipe failure probabilities), (2) plant design features, and (3) a deterministic analysis of debris generated, potential transport to the sump, and potential attendant blockage which could lead to loss of NPSH.
Such an evaluation NUREG-0869, Revision 1 0-1 October 1985
begins witl. an estimation of pipe failure probabilities (which are a function of pipe size and weld type), followed by an estimate of the volume of debris that can be generated by any break postulated (which is a function of break size, break-to-target locations and possible combinations, and break jet model); debris transport potential; and blocked screen head loss (which is a function of the quantity of debris transported, the available debris screen area, and the post-LOCA recirculation flow rate requirements).
The examples that follow are provided to illustrate such evaluations.
ESTIMATING PIPE WELD FAILURE PROBABILITIES The first step is to estimate the probability of pipe (or weld) failure to calculate the initiating event probability (i.e., LOCA probability).
The probabilities shown in Table 1 were estimated by M. Taylor (DEDR0GR staff based on his review of " Reliability of Piping in Light Water Reactors" by S. Bush (IAEA-SM-218/11, October 1977; see also Appendix C of this report.)
They represent the estimated failure probabilities for all pi)ing in a typical nuclear plant for the diameter classes shown.
For example, tie estimated pipe failure probability of any pipe in the 10- to 16-inch diameter range is 3E-5 per Rx yr, and the failure probability of a fabricated or non-standard weld in this diameter range is 1.5E-5 per Rx yr.
To estimate pipe failure probabilities as a function of pipe diameter size and the type of weld (the assumption being that failure would occur at the weld joint), the data shown in Table 1 can be used to calculate a weld failure probability (Pwk), as foHows:
(P )(N Njn+N,Wja + N,Wje) j n
Eq (1) p
=
n jn a ja + X,Wje)
+XW (X W Where:
Pwk = pr bability of a weld failure in diameter class "k" pipe, weld type weighted P) =probabilityofpipebreakinanypipeindiameterclass"j" N
= number of welds of type "n" and diameter "k" n
X
= number of welds of type "n" in diameter class "j" n
jn = probability weighting factor for type "n" welds W
N,
= number of welds of type "a" and diameter "k" X
= number of welds of type "a" in diameter class "j" a
ja = probability weighting factor for type "a" welds W
N,
= number of welds of type "e" and diameter "k" X
= number of welds of type "e" in diameter class "j" e
je probability weighting factor for type "e" welds W
0-2 October 1985 NUREG-0869, Revision 1
TABLE 1 Piping failure probability estimates Pipe Failure Weld failure probability distributir)n size Diameter probability Weld Weld Weld (inches) class (1/Rx-Yr) type 1 type 2 type 3 0
J P)
W W
W n
a e
2 to <6 1
3x10-4 0.7 0.i5 0.15 6 to <10 2
4x10-5 0.5 0.30 0.20 10 to <16 3
3x10-5 0.5 0.30 0.20 16 to 28 4
3x10-6 0.5 0.30 0.20
>28 5
3x10'8 0.5 0.30 0.20 IP) = 3.76x10-4/Rx yr Weld Type 1 = fabricated and non-standard joints WeldType2=highrestraintjointsandteeswithjoints Weld Type 3 = elbows, reducers, and straight piping runs with joints NUREG-0869, Revision 1 0-3 October 1985
_ - =.
The weld failure probabilities that were derived from the failure assumptions shown above and treated algebraically as described in Equation 1 were used to l
l estimate the LOCA probability, using the Salem 1 plant and piping layout.
NUREG/CR-3394 details those analyses.
The weld sizes and distributions derived from a typical Salem 1 plant primary cooling piping loop are shown in Table 2.
The breaks were assumed to occur at weld locations [following the criteria in Section 3.6.2 of the NRC Standard Review Plan (NUREG-0800)]. The loop analyzed contained 238 welds associated with piping that can be classified as LOCA-sensitive piping.
I Because the estimated pipe failure probabilities in Table 1 must be distributed on a per weld basis the first ctep is to apportion (or redistribute)thetotalprobabilityforadiametricsizeclasstoall existing pipe sizes.
Table 3 has been constructed to illustrate how Equation i
1 was used to develop such a distribution.
If weld type is ignored in the first step, the per weld failure probabilities distribute as a function of iae size) to the total thefractionofthenumberofwelds(ofagivenp(tlatis,j=j, class) l number of welds in a particular diametric class 1
2, 3, 4, and 5 as shown in Table 3).
The summed totals (for a particular alwayssumtothetotalprobabilityforthatdiameterclassextractedfrom Table 1.
Table 3 illustrates this process.
In addition, Table 3 compares NUREG/ype of fractional distribution with the weld type weighted values fr this t CR-3394.
The variability within a particular diameter class (j = 1 2,
4,and5)resultsfromthedistributionofweldthpesintheSalem1 plant 3pIpinglayoutanalyzed(seeagainTable2).
NUREG/C 3394 analyses were based on both weld type and pipe segment probability distributions, l
i l
The initiating event probability (P,) attributable to all 237 welds was calculated to be 3.7E-4/Rx yr and la in agreement with the IP) value shown in Table 1.
Thus this type of single loop analysis is applicable to all loops.
Had the distribution methodology discusseo above been a) plied to four loo)s, the same initiating break probability would have been catained, because tie overall probability cannot exceed the summation noted in Table 1.
Because satisfactory operation of the sump is essential when breaks occur in the primary coolant system LOCA-sensitive pressure boundary, the overall probability (P ) discussed above should be reduced (for purposes of l
g l
estimating sump blockage probabilities) to account only for such breaks in l
the primary coolant system.
In many PWRs (such as Salem 1) such piping is located within the crane wall region and that break probability is designated P.
Therefore, elimination of secondary system piping weld locations and l
j l
piping outside the crane wall resulted in a calculated Pg of 1.84E-4/Rx yr for Salem 1.
These break locations (associated with P ) were further j
analyzed for debris generation as discussed below.
l NUREG-0869, Revision 1 0-4 October 1985
TABLE 2 Break size distribution for Salem plant analysis System weld distribution No. of welds W p)ipe System No. of diameter (inches designation System welds 1
Hot leg 8
8 9 34 2
Cold leg 6
6 9 36.3 3
Cross over 6
6 9 32.3 4
Safetyinjection/coldleg 41 15 9 10, 4 9 11 11 9 6, 11 9 2 5
Safety injection / hot leg 33 696,2792 6
Chemical volume and control 13 5 9 16, 8 9 14 7
Feedwater 95 22 9 4, 7 9 3, 66 9 2 8
Main steam 20 19 9 30, 1 9 32 9
Pressurizer 16 16 9 14 SUBT0TAL WELOS
= W for loop analyzed
-16 for pressurizer
-13 for chemical volume and control
-33forsafetyinjection/hotleg 175 x 4 loops
= 704 Pressurizer loop welds
= 16 2 chemical volume and control loop welds (13)
= 26 2safetyinjectionandhotlegpenetrationswelds(33) = 66 TOTAL. WELDS
=812 for 4 loops Diameter No. of (inches)
Welds 34 8
hot leg 36.3 6
cold leg
- > 40 welds for 16 in to 34 in, pipe 32.3 6
crossover.
(P = 3E-6) h fg I
Main steam 16 5-29 welds for 10 in, to 16 in. pipe 14 24 (P = 3E-5) 10 15 1 19 Selds for 6 in, to 10 in pipe 8
4l (P =4E-5) 0 6
17 J 22
>150 welds for 2 in, to 6 in, pipe l
4 3
7 (P = 3E-4) 0 2
104 Source:
NUREG/CR-3394, Vol 2, Table A.1-1.
NUREG-0669, Revision 1 0-5 October 1985
F TABLE 3 Probability distributions as a function of diameter
- i l
(1) for j = 1, Pj = 3E-4 Wn Wa We k = 6", Nk6
- 17 l-0.7 0.15 0.15
= 4", Nk4 = 22
= 3", Nk3
- I
= 2",'Nk2 = 104 l
Nk= 150weldsforj=1 w/o Weld Specification NUREG/CR-3394 Values P =2" = (3E-4)(104/150) = 2.07E-4 2.07E-4 l
k P=3"=(3E-4)(7/150)
= 0.14E-4 0.13E-4 k
P=4"=(3E-4)(22/150) = 0.44E-4 0.41E-4 j
k Pb6" = (3E-4)(17/150) = 0.34E-4 0.38E-4 P _g
= 2.99E-4 2.99E-4 I
l j
)
(2) for j = 2: k = 8" and 10", Nk8 4 and Nkl0
- 1D
[
w/o Weld Specification NUREG/CR-3394 Values t
P =8"
= (4E-5)(4/19)
= 0.84E-5 0.71E-5 k
P=10"=(4E-5)(15/19) = 3.16E-5 3.29E-5 k
P -2
= 4.00E-5 4.00E-5 j
(3) for j = 3: k = 14" and 16", Nk14 = 24 and Nk16
- 0 w/o Weld Specification NUREG/CR-3394 Values P =14" = (3E-5)(24/29) = 2.48E-5 2.47E-5 k
P=16"=(3E-5)(5/29)
= 0.52E-5 0.53E-5 k
P =3
= 3.00E-5 3.00E-5 j
(4) for j = 4: There was no piping in this diameter range for Salem 1 (5) for j = 5: k = 28", Nk32 7, Nk34 = 8, Nk36 = 6 w/o Weld Specification NUREG/CR-3394 Values P =30" = (3E-6)(19/40) = 1.42E-6 0.93E-6 k
P ='i2" = (3E-6)(7/40)
= 0.53E-6 0.88E-6 k
P =34" = (3E-6)(8/40)
= 0.60E-6 0.72E-6 i
k P =36" = (3E-6)(6/40)
= 0.45E-6 0.47E-6 k
P =5
= 3.00E-6 3.00E-6 j
1
(
- Based on 238 welds / loop, as in Salem 1.
~
l i
NUREG-0869, Revision 1 0-6 October 1985
i l
)
DEBRIS GENERATION r
Estimating debris generation is a function of break size, jet expansion model, and the break versus target locations.
For the analyses repo'rted in NUREG/CR-3394lnfluencethatareshowninFigure1,becausethedecompression a hemispherical jet expansion region model was ' selected with the zones of pressure field for a high pressure, subcooled jet can be approximated with a hemispherical model.
The energy levels (stagnation pressure level) within this expanding jet are also a function of distance from the break (or length to diameter, L/0).
Calculational models that have been correlated with experiments show that at L/D = 3, the jet stagnation pressure is very nearly the same as the stagnation pressure within the jet emanating from the break location.
For a PWR, this means pressures on the order of 2200 psi and extreme insulation destruction would take place, particularly for fibrous insulation. At L/D = 7, the PWR subcooled break jet stagnation pressure has been calculated to reduce to 20 to 40 psi.
Although this is still a very high velocity field, experiments have shown (see NUREG/CR-3170) that the fiberglass covering shreds rather than totally destructing at these reduced pressures.
For analysis purposes three L/0 ratios of 3 5, and 7 were selected to assessdebrisgenerationeffectsparametrically.
The lower bound (L/D = 3) represents the highest jet intensity, because the exaanding jet dynamic pressure at that distance is nearly that at the brea( jet exit plane and no conservatism exists for survival.
The outer bound (L/0 = 7) represents an axialdistancewherejetstagnationpressureshavedecreasedto20to40 psia, and at this distance the assumption that total destruction takes place does carry some conservatism.
(See NUREG-0897, Revision 1 for further discussion on the selection of these L/D ratios to represent the different zones of insulation destruction.)
In a,iltion, blowdown tests in the HDR facility have demonstrated the highly d structive capabilities of LOCA jets.
The weight of evidence is strongly against conceptualizing a high pressure pip'an e
break as a simple water jet model.
Such a break can be better termed as explosion."
These flow rates investigated parametrically represent emergency core cooling system flow rates given in Final Safety Analysis Resorts (FSARs) submitted by l
applicants for operating licenses.
The range of de)ris screen areas evaluated is rearosentative of PWR sump designs. Because there is no standard sump design or ECCS flow requirement, this range was used to scope the range of sump blockage probabilities for PWRs; it is discussed in detail in NUREG/CR-3394.
The break location (that is, weld locations)ing the plant insulation versus target combinations for Salem I were systematically evaluated utiliz distribution. (The insulation was approximately half reflective metallic and half encapsulated fibrous insulation.) Table 4 shows the level of detail employed to evaluate all possible break-to-target combinations.
The debris volumes associated with 14-inch pipe breaks are used for illustration, because these medium size breaks contribute significantly to the calculated overall sump blockage probability.
NUREG-0869, Revision 1 0-7 October 1985
EA oS
'(-
Crane war A
g O
\\
\\
T
\\sG
\\
1 i
]/
fs [ uD=3 1
r o
l l
/
/
\\
l l
P Q HL rs V
g i
j j w
y '-
Where RPV - Reactor Pressure Vessel UD=3 8
I SG - Steam Generator UD=5 RCP - Reactor Coolant Pump UD=7 Q (
\\
HL - Hot Leg UD=7 UD - Distance t s Target tu Up=5 Divided by Break Diameter (D) e h
CO - Crossover 5
-s G$
Figure 1 Zones of influence utilized for debris generation estimates in NUREG/CR-3394
)
TABLE 4 Illustration of break vs. target combinations and debris volumes calculated est&R f40047.. TARGti lh$ut.., INSut. eud, Isolut,. _,fA44G L,famegt g ggggg,,,,,, ggggg $ ggg, ggggg, sansegn ogassi. LtNGt06 tMICE.
ITPE A47. v0Luut Glass.
Lametse IMICa.
type esp. yoguang ft.
.IhC MES
- . FT *e3 IN.
ff.
IIICteEl gw
=
, pt ee
y.----=f---_=_.=--_-- ;.- - -.t
,,j
,,,,=*----==--,- -- - -
et 94.ee t.Se 3.Se
$4 3
0.7 to.se.
3.37 3.Se gg 3
g,3 tt3.et...t4 33 3.64 s( __a 349 t
. et.._ le.ee._, 3.se..
3.ee
,_st 3
3,t..,_]gde 3,ST 3._9e H
3
).4
~*
173.00 te.33 3.54 SE 3
33e.t p
a*
,,,..... 3. u -' 5.~u-~ st 3
G W;h G7 3'.h n
3 '~~' ' 3. 3 - ' ' * - - ~~ m'
$f3.00 to.33 3.be le 3
33e.t g
et 14.00 0.76 3.ee St 3
?.S to.se 0.94 3.04 at 3
to.e L-173.ee
-- f oie ?
3.lg u
a agt,g y
0" 3,e4 u
is...
t.ie 3..e 3
t..
__..t..
to.re _ 3.Se g,
Ig..
3e.co 7.se st t.3 T ~tei it;e4"~~~'~6.~iii 5.ee M
3 e.t g
~
in--
.. e4 - e.63 r.e4 er s
ti.e
= %
"' tea '~~ ~ t e.ee "
- e. t e ~" $. se ~~'~~u~
5 e.t v
~'
toe
04.00 e.oe '~
3.00'~~~ ~ $4 3
to.I
~ * *
~'
^
"see'"**te.ee 6.33'"~~f.ee ~ ~ u 3
to.4 tes t o.'oe 4.ee 3.ee.
si 3
s.e te.u v. 3'8 3.se sa 3
e3
~
set se.oo t.se 3.co st 3
s.e sta.ee to.et 3.se sa 3
3et.e t
Ite_
s.ee late e.oo u
33 a.i i
3M 3.te _ 1.48 8.H _ sa 3a e
t.es _ 2.33
.t.ga ens at g.:
t l
l L
Source: Table B.5-3, NUREG/CR-3394 Volume 2 i
l NUREG-0869, Revision 1 D-9 October 1985
i L
y-t N
N UNACCEPTABLE DEBRIS GENERATION Unacceptable debris generation is defined as that volume of debris that, if transported to the sump screen, would result in a blockage head loss greater than NPSH requirements.
Thus, the debris volume calculated through use of the break-to-target zones (or L/0 ratios) shown in Figure 1 must be compared to the volume that could generate unacceptable screen blockage head losses that exceed the NPSH margins.
This effect was analyzed parametrically and is reported in detail in NUREG/CR-3394.
Experiments have shown that.the head loss for forced flow through fragmented fibrous insulations deposited on a debris screen can be expressed as an i
exponential relationship (NUREG/CR-2982, Rev. 1) and that such head losses are highly dependent on material type.
The following empirical relationships for head loss (HL) have been obtained:*
HL = 1653 (Q/A)1.84 (V/A)1.54 for high density fiberglass HL = 68.3 (Q/A)1.79 (V/A)1.07 for NUK0N
}
HL = 123 (Q/A)1.51 (V/A)1.36 for mineral wooY-In addition, these experiments have shown that shredded fibrous insulation materials distribute uniformly over debris screens.
The data for NUK0N were submitted by the Owens Corning Fiberglass Corporation during the For' Comment period for USI A-43.
The analyses reported in NUREG/CR-3394 have parametrically investigated the following plant design and operational variables:
Variable Range investigated Pipe' break-to-target distance 3 to 7 pipe diameters Sump flow rate 6,000 to 10,000 gpm Debris screen area 50 to 200 ft2 1
y M
- Q/A is the approach velocity as calculated from volumetric flow (ft3/sec) divided by debris screen area (ftz) and the equivalent debris thickness (V/A).
V/A comes from the transported debris volume (V) divided by the debris screen area (A).
t NUREG-0869, Revision 1 D-10 October 1985 h
i The analyses reported in NUREG/CR-3394 (as illustrated in Table 4) considered each singular break-target combination to determine debris generation volumes.
Each possible combination was considered on a singular weld / target basis for determining if that particular break probability should be retained for estimating overall sump blockage probabilities.
The criteria for maintaining a particular break were based on a calculated head loss of 1, 2, or 5 feet of water because these evaluations were performed parametrically.
Table 5 shows the probability distribution as a function of pipe diameter for the following:
sump screen area of 50 ft2; recirculation flow rate of.8000 gpm; and an allowable head loss of 1 foot of water.
The estimated sump blockage probabilities are 1.8E-5, 3.3E-5, and 4.5E-5/rx yr for L/D = 3, 5, and 7.
The calculations reported in NUREG/CR-3394 considered insulation with debris associated both with piping and targeted compartments.
They are admittedly conservative because shadowing effects attributable to piping supports and other structures are not included. To illustrate the sensitivity of such an assumption, a series of calculations have been performed to assess the significance of insulation only on failed piping, and as a function of diameter size, thereby ignoring other targets.
These results are shown in Table 6.
These estimated insulation volumes and associated head losses show that smaller diameter primary coolant system piping (6-to 10-inches), for the flow and debris screen area assumptions shown in Table 6, will not generate sufficient quantities of fibrous debris to result in head losses exceeding 1 to 2 feet of water, thus they support the conclusions that can be derived from Table 5.
These findings would be most applicable to plants having small NPSH margins and small screen areas.
The analyses reported in NUREG/CR-3394 were run with the high density fiberglass head loss correlation, and it was assumed that the insulation (for L/D = 3, 5, and 7) was totally destroyed, was transported to, and was deposited on the sump screen.
The assumption of total transport is an imbedded conservatism for plants that have recirculation flow velocities less than 0.2 ft/sec within the containment and for those plants where intervening structures would inhibit transport.
On the other hand, there are PWRs where the primary ccolant system piping and the sump location are not isolated by intervening structures.
SUMP FAILURE PROBABILITY Wht # the calculational methods described above were applied to the plant parameters investigated they produced the range of estimated sump failure probabilitie; shown in Table 7.
(More detailed tables are provided in NUREG/CR-3394.)
For high flow rates (10,000 gpm), small debris screen area (50 ft2),
L/0 = 7, and low allowed head loss (1 ft of water), the calculated sump failure probability was 5.4E-5/Rx yr.
For values of 6000 gpm, 200 ft2, L/D = 3, and an allowed head loss of 2 ft. of water, a NUREG-0869, Revision 1 0-11 October 1985
i TABLE 5 Pipe break and sump blockage probabilities ~vs. pipe diameter I
l
..?AOLE 0.9.t-30.
SuomeARY OF PROSASILITIt$ 01 AssETER SAll$ WlfH WELD TYPE WEIG41 tis 0.... -
SCREEN AREA
- 4. FEET OF H2O
.DIA -. PO P!
BLOCMAGE FREO.
BLOCMAGE. FREg_.. SLOCuant FREG-... BLOCRAGE PROS _ 8LOCRAGE PROS-.-. SLOCKAGE.Pana j
II.
L DVER 0 e 3 L OVER D e S L OVER 0 e 7 L OVER D e 3 L OVER D e S L OVER 0 e 7 PO SFW P1 SFW PO BFW PI Sf W PO OFW PI SFW P0 M PI ON PO M PI BM PO BK PI M
.---.+,s.
.esm -ares e.se-es r-
,ee i,..emma nenne-e.--se.=m r--ese.e--.e e-na.e=
3.'3.07t-04 7.001-09 0.9 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
- 3. 0.0 0.0 0.0 - _ 0.0 0.0 0.0 4.0 0.0 0.8 a.a 4.0 0.0 0.0 4t.0
- 3. 8.304-0S 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 C. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1
.4. 4.10E-00 0.0
. 0.0 0.0 0.0 0.E S.O 0.0 0.0 a.a 0.0 0.a 0.0-0.0 S. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 C. 3.0$t-OS 3.098-00 0.0 0.0 2.74E-St 2.74E-Ot 3.7tt-St 3.7tt-08 0.0 0.0 t.;E-05 9.lt-09 1.4t-0$ t.4E-05
- i. 7. tit-00 7.918-08 0.0.. 0.0 0.0 0.0 9.50EeSt_2.50E-01.0.0
^0 0.0 n.0 9.SE-08.1.sg,ag th 3.20t=0S 3.20E-06 9.3S4-09 9.3S1-01 2.70E-St 2.70E-08 2.70E-04 2.70E-09 4.4L-00 4.4E-06 0.St-00 0.0E-00 0.0E-00 0.0E-OS ts. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
.8 % 3.478-06 2.306-OS 3.33E-S t.3.SStue1 A.33Ee01.3.03En01 7,00&ee t.S.70&=0 6. 0.2naa m.0Ee0&_1.tts08.0.0tm04 4.7E-45 3.0E=OS-
- 92. S.34E-04 3.70E-00 9.00E+00 1.00E*00 9.00E+00 1.00E*00 1.00E*00 1.00t+00 S.3L-00 3.7E-00 S.3E-00 3.7E-00 S.3E-OS 3.7E-00 ll. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 30 0.0 - 0.0 0.0 -
0.0 S.O a.a n,g a,a n_a_
n_a n."
9.0 a.O.- - a.Q
- 33. 0.0 0.0 0.0 0.0 0.0 e.0 0.0 0.0 0.0 P g0.0 0.0 0.0 0.0 0.0
- 34. 0.0 0.0 0.0 0.0 0.0
%.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
- 08. 0.0.
- 0. 0....
0.O r 0.0 0.0 a.0 a.0 a.a 8.a 0.a n.O 0.0 0.0 0.a a 30. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 I 0.0 0.0 0.0 0
- 30. 0.318-07 0.0 1.00E+00 0.0 1.00E+00 0.0 1.00t*00 0.0
- 8.3E-07 0.0 0.3t-07 0.0 0.3t-07 0.0 ta8. 33. 0.70E-07 0.2 t E-07 S.004-0 4 4. 4 7 8 0 8. 5.00E-01 4. 8 76=SS. ASSE*0 8.4. l?t=01-.S.2L*0La.0&=07 5.2 E-07 2.Ste47.5.2E-07 2.SE*97
%s 34. 7.24E-07 7.24t-07 S.00E-01 S.00E-O g S.00E-O f S.00E-01 S.00E-01 S.C05-0 6 3.0L-07 3.SE-07 3.SE-07 2.SE-07 3.SE-07 3.SE-07
- 30. 4.00E-07 4.00E-07 S.66E-01 S.50E-01 0.87t-09 S.S?E-01 7.78E-01 7.70E-Ol 2.6L-07 2.St-07 3.t t-07 3.tt-07 3.0E-07 3.0E-07
-30. 0.0 0.0 --
0.0 c.
0.0 0.0 m,a g,a e,g g,a n.g a,g 0.0 a.O 0.a
- 40. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
- 83. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
. 44. 0. 0.-
0.0
- 0. 0 - -- 0. 0 0.0
- 0. 0 -- -0. "
4.0 0.8 9.a 0."
4.a 0.m n,g
- 40. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0
- 40. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 C.G 0.0 0.0 0.0 0.0 0.0 l
00
- 0. 0 -. 0. 0 - ~.- 0.0
-.0.0 0.a m.S a.0 0.4 S.O a.a a.6 Aa a a
^^
- 53. 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 70f 3.738-04 8.70E-04 0.0 0.0 0.0 0.0 0.0 0.0 2.0L-OS'O.SE-09 3.0E-05 3.3E-05 S.0E-05 4.St-08 j
N0utteCLATURE flA -1008784fl008 Evtlef OI AastTER-INCHES PO -OVERALL PR00ASILif f OF INti. EVENT OCCURRENCE
- CI--* PROS.- IIIII. EVENT-PR IGI.-STST-4N-CN A805 WAt i SLOCAAGE-SRE0=R AT IO -0WhMIS-GAW68 NG-SLOGM S OLOCRAGE P900-PR00ASILiff 0F Ue8 ACCEPTABLE TO TOTAL EVLMTS(WEIGHTED SASIS)
ECCS SuuP SLOCKAGE L OVER D
- TARGET DISTANCE TO BREAM OIAast1ER RATIO
--4PW--OLOCuAGE FREGUENCY WELO TVPt. WEIGH 7IB8G-44Sil.- -GFL--OLOCMAGE-# REQ,-StGasENS-4&e8 Glee-WElG84 Tit 84-SAS&S i
PO FOR OvtRALL PROS ANO PI FOR print IN CW PROS PO FOR OVERALL PROS AND PI FOR PRII4 IN CW PROS i
WI -0Locua0E PROS PO FOR OVERALL Pt FOR PRIII IN CW 1
l l
I t
I l
l Source:
NUREG/CR-3394, which contains the complete calculated data sets NUREG-0869, Revision 1 D-12 October 1985
TABLE 6 Headloss as a function of pipe size, insulation type and debris volume PIPE INSULATION VOLUME SCREEN RECIRCUL.
HEAD (1)
HEAD (2)
DIAMETER THICKNESS L/D DESTROYED AREA FLOW LOSS LOSS j
INCHES INCHES Cu Ft Sq Ft 8pm Ft Water Ft Water 36.30 3.50 3.08 27.58 50.00 10000.00 149.45 8.50 36.30 3.50 5.00 45.97 50.00 19808.08 328.21 14.69 36.30 3.50 7.00 64.35 58.00 19800.00 551.04 21.06 34.00 3.50 3.00 24.34 50.00 1E200.00 123.29 7.44 34.08 3.50 5.00 40.57 50.00 19095.00 278.74 12.85 34.00 3.58 7.00 56.79 58.80 19800.00 454.56 18.42 32.30 3.50 3.00 22.57 50.00 19800.00 106.07 6.70 32.38 3.50 5.00 36.79 50.00 10000.00 232.93 11.58 32.38 3.50 7.88 51.51 50.00 10000.00 391.00 16.59 16.00 3.88 3.00 4.97 50.88 10000.08 18.69 1.36 16.00 3.00 5.00 8.29 50.00 10000.00 23.47 2.35 16.00 3.00 7.00 11.61 50.00 19800.00 39.41 3.37 14.08 3.05 3.00 3.89 50.88 10000.00 7.33 1.05 14.00 3.00 5.00 6.49 50.OO 18088.00 16.18 1.81 14.88 3.00 7.00 9.09 50.00 10000.00 27.84 2.59 10.00 3.00 3.00 2.13 58.80 10000.00 2.89
.55 10.08 3.00 5.00 3.55 55.08 10000.00 6.35
.95 10.00 3.00 7.90 4.96 50.00 10008.80 10.65 1.36 8.00 3.08 3.00 1.44 50.00 10000.00 1.58
.36 8.00 3.00 5.00 2.40 50.00 10000.88 3.48
.62 8.00 3.00 7.80 3.36 50.00 10000.00 5.84
.89 6.00 3.00 3.00
.88 50.00 10000.08
.75
.21 6.00 3.00 5.99 1.47 58.88 10000.05 1.64
.37 6.00 3.00 7.00 2.06 50.00 19888.00 2.75
.53 6.00 1.50 3.00
.37 50.00 18080.50
.19
.08 6.00 1.59 5.00
.61 50.00 19800.08
.43
.14 6.00 1.58 7.05
.86 50.00 19080.00
.72 e21 l
l 2.00 1.50 3.00
.06 50.00 1e000.00
.01
.81 l
2.00 1.55 5.c5
.10 59.50 10000.Os
.02
.02 l
2.08 1.50 7.00
.13 50.00 10000.00
.04
.33 (1) High Density Fiberglass, H=1653((Q/A)^1.84*(V/A)^1.54 (2) Low Density Fiberglass, H=68.3((Q/A)^1.79*(V/A)^1.07 NUREG-0869, Revision 1 0-13 October 1985
=.-
TABLE 7 Summary of the probability of sump failure RECIRCULATION SCREEN ALLOWED EST'D FAILURE PROBABILITY FLOW RATE AREA HEAD LOSS (BASED ON: WELD TYPES)
(GPM)
(SQ FT) (FT WATER) L/D=3 L/D=5 L/D=7
.6000 50 1.O 1.1e-5 1.7e-5 2.Be-5 6000 75 1.O 5.7e-6 1.2e-5 1.Se-5 i
6000 100 1.O 3.1e-6 6.7e-6 1.Be-5 E
6000 200 1.O 2.9e-6 5.9e-6 6.9e-6 6000 50 2.0 9.9e-6 1.2e-5 2.4e-5 6000 75 2.O 3.1e-6 9.2e-6 1.Go-5 6000 100 2.O 3.1e-6 5.9e-6 1.4e-5 6000 200 2.O 2.9e-6 5.Be-6 6.9e-6 6000 50 5.O 4.7e-6 1.2e-5 1.Be-5 6000 75 5.O 3.1e-6 5.9e-6 1.Se-5 6000 100 5.0 3.1e-6.
5.9e-6 7.7e-6 6000 200 5.O O
5.3e-6 6.9e-6 8000 50 1.O 1.6e-5 3.3e-5 4.5e-5 BOOO 75 1.O 9.le-6 1.2e-5 2.4e-5 9000 100 1.O 3.9e-6 1.2e-5 1.Se-5 BOOO 200 1.0 2.9e-6 5.9e-6 7.7e-6 B000 50 2.0 1.ie-5 1.7e-5 2.7e-5 8000 75 2.0 4.7e-6 1.2e-5 1.Be-5 B000 100 2.0 3.1e-6 6.7e-6 1.6e-5 0000 200 2.0 2.9e-6 5.9e-6 6.'9e-6 l
8000 50 5.O 7.4e-6 1.2e-5 1.9e-5 8000 75 5.0 3.1e-6 6.7e-6 1.7e-5 l
8000 100 5.O 3.le-6 5.9e-6 1.3e-5 l
8000 200 5.O 2.9e-6 5.7e-6 6.9e-6 l
10000 50 1.O 1.6e-5 4.1e-5 5.4e-5 10000 75 1.O 9.9e-6 1.2e-5 2.4e-5 10000 100 1.O 4.Be-6 1.2e-5 1.Be-5 10000 200 1.O 3.le-6 5.9e-6 9.4e-6 10000 50 2.O 1.1e-5 1.Be-5 4.3e-5 10000 75 2.O 7.4e-6 1.2e-5 1.So-5 10000 100 2.O 3.1e-6 9.2e-6 1.Be-5 10000 200 2.O 2.9e-6 5.9e-6 6.9e-6 10000 50 5.0 9.9e-6 1.2e-5 2.4e-5 10000 75 5.O 3.le-6 9.2e-6 1.Be-5 10000 100 5.O 3.le-6 5.9e-6 1.4e-5 10000 200 5.0 2.9e-6 5.se-6 6.9e-6 NUREG-0869, Revision 1 D-14 October 1985
4 blockage probability of 2.9E-6/Rx yr is calculated.
Because insulation change out is'an ongoing plant activity, the staff does not specifically know the types and quantities of insulation employed.
Therefore, calculation of a
. generic value is not possible.
References Besh, S., " Reliability of Piping in Light Water Reactors," International Atomic Energy Agency, IAEA-SM-218/11, October 1977.
U. S. Nuclear Regulatory Commission, NUREG-0800, " Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants," July 1981.
--, NUREG/CR-2982, Revision 1, " Buoyancy, Transport, and Head Loss of Fibrous Reactor Insulation," D. N. Brocard, Alden Research Laboratory, July 1983 (also Sandia National Laboratory, SAND-82-7205).
--, NUREG/CR-3170, "The Susceptibility of Fibrous Insulation Pillows to Debris Formation Under Exposure to Energetic Jet Flows," W. W. Durgin, and J. Noreika, Alden Research Laboratory, January 1983 (also Sandia National Laboratory, SAND-83-7008).
--, NUREG/CR-3394, "Probabilistic Assessment of Recirculation Sump Blockage Due to Loss of Coolant Accident," J. J. Wysocki, Burns and Roe Inc.,
July 1983 (also Sandia National Laboratory, SAND-83-7116).
1 i
I NUREG-0869, Revision 1 D-15 October 1985
i APPENDIX E CONSEQUENCES OF LOSS OF RECIRCULATION CAPABILITY l
NUREG-0869, Revision 1 October 1985
CONSEQUENCES OF LOSS OF RECIRCULATION CAPABILITY Risk analyses were utilized to assess the public health consequences assoc-iated with a loss of recirculation flow capability in the period after a loss-of-coolant accident (post-LOCA period) as a result of the effects of debris blockage.
Blockage of a sump in pressurized water reactors (PWRs) and blockage of residual heat removal (RHR) suction inlets in a boiling water reactor (BWR) have similar consequences since loss of the redundant recircula-tion systems can result in core uncovery, which will lead to core melt.
In addition, containment sprays will fail if a loss of suction takes place and containment overpressurization can occur.
Loss of containment structural integrity leads to high public doses.
To obtain initial estimates of public exposure, the staff referred to probabilistic risk assessments (PRAs) of five plants:
Surry, Calvert Cliffs, and Crystal River-3, which are PWRs, and Peach Bottom and Grand Gulf, which are BWRs.
The accident sequence of interest for the PWRs is designated AHF; a large break LOCA (A) followed t4y failure of recircula-tion to the core (H) and to the containment sprays (F).
For the BWRs, analogous sequences designated AHI, AI, and AGHI were used as the basis for staff evaluations.
All five PRAs are based on the reactor safety study (RSS) methodology (NUREG-75/014).
Consequently the conditional probabilities for various containment failure modes were approximately the same for all three PWRs.
A synopsis of conditional probabilities of source terms is shown in Table 1.
Table 1 shows that 70% to 80% of the AHF sequences lead to release category PWR-6 (basemat meltthrough), and 20% to 30% lead to PWR-2 (early overpressure failure).
There are also small conditional probabilities of PWR-1 (steam explosions) and PWR-4 (failure to isolate containment).
TABLE 1 Conditional probabilities of release categories for large break LOCA with loss of ECCS recirculation, as estimated in plant-specific PRAs for the five reference plants, using RSS/RSSMAP methodology Plant release category probability Plant PWR-1 PWR-2 PWR-4 PWR-6 Surry 0.01 0.207 0.78 Crystal River 0.01 0.2 0.007 0.8 Calvert Cliffs 0.01 0.3 0.007 0.7 BWR-1 BWR-2 BWR-3 BWR-4 Peach Bottom 0.008 0.165 0.824 0.003 Grand Gulf 0.01 0.79 0.2 NUREG-0869, Revision 1 E-1 October 1985
For the two BWR plants (using RSSMAP methodology), about 1% of the events are expected to result in EWR-1 releases (steam explosions), and about 80% would lead to BWR-3 (radiatior, release to the reactor building).
For Peach Bottom, there is a 16.5% probability of BWR-2 (direct release to the environment) and less than 1% chance of BWR-4 (failure to isolate the drywell).
For Grand Gulf, there is no chance of a BWR-2, but a 20% probability of containment isolation failure reported in the RSSMAP study.
The offsite doses associated with each release category were evaluated with the CRAC code, under the following assumptions:
a uniform population distri-bution of 340 people per square mile, meteorological conditions typical of the Byron site, no evacuation, and integration of conditional consequences to a 50-mile ~ radius from the plant.
The resulting estimates of public exposure are shown in Table 2.
It should be noted that the conditional consequences are not monotonically decreasing with release category, as would be expected considering'the release category definitions. This is most likely a result of neglecting the dose beyond 50 miles, the higher release energy for BWR-1 deposits of fission products at greater distances.
TABLE 2 Calculated conditional public radiation exposures from each of the release categories for the A-43 reference site conditions using RSS methodology Calculated Calculated Release consequences Release consequences category (person-rem) category (person-rem)
PWR-1 5.4E+6 BWR-1 5.4E+6 PWR-2 4.8E+6 BWR-2 7.1E+6 PWR-3 5.4E+6 BWR-3 5.1E+6 PWR-4 2.7E+6 BWR-4 6.1E+5 PWR-5 1.0E+6 PWR-6 1.5E+5 PWR-7 2.3E+4 The conditional consequences of a core melt due to sump failure are obtained by weighting the calculated public exposure for each release category (Table 2) with its associated conditional probability (Table 1), and then summing over release categories.
The resulting conditional consequences for the five i
reactors are listed in Table 3.
The risk, in person-rem per reactor year, is the product of core melt frequency and conditional consequences.
NUREG-0869, Revision 1 E-2 October 1985
. TABLE 3.
~ Preliminary estieate of conditional con' sequences for loss of. recirculationu(based on WASH-1400 methods of assessing severe accident risks, using 1RSS/RSSMAP methodology)
Conditional consequences-Plant (person-rem)
Surry.
1.2E+6
- Crystal-River-1.1E+6 Calvert Cliffs 1.6E+6 Peach Bottom 5.4E+6 Grand Gulf 4.2E+6 1
LSince the publication of the WASH-1400 study in 1975, a grea't deal of 1
_ progress-has bee.n made:in two areas relating to the calculation of severe t
accident consequences:.(1) containment response characteristics and (2) radiological release fraci. ions. When the preliminary version of the A-43 resolution package was presented to the Committee for Review of Generic
- Requirements.(CRGR), members.of the Committee _ inquired as to whether recent.
J
-research results related to containment response would substantively alter z
7 the consequen_ce evaluation presented above.
The staff has performed such an
~
' assessment, and has examined the possible impact of the revised methodology l
for estimating radiological source terms.
I
' Reassessment of Containment Response Th'e containment response assumed in estimating the consequences of sump failure in the reference PWRs ma from the RSS/RSSMAP studies there are (1)y be characterized as follows:
about a-20% conditional probability of_a serious radiological release to the environment (PWR-2) and-(2) about an-80%
i probability of a benign release (PWR-6).'
If the estimates of.-societal risk are to be significantly reduced from those given above, the'(conditional)
- probability of a serious release would have to be significantly lower than 20%. To' state with confidence that the release rate is that low would-4 require a great-deal of. confidence in the systems that prevent containment
. failure _and reduce the resulting source term.
i l
There.isLone containment-type which meets this criterion, the PWR large dry L
containment with safety grade fan coolers (a large majority of large dry L
containments safety grade.have fan coolers).
The staff has a great deal of p
confidence that overpressure failure of these containments can be prevented either by operation of_.the fan coolers or by connecting the sprays to an
~
- alternative source of water..The probability that the fans will fail and that spray-hookup to an alternative water source will not be made is judged to the i
small.
Large dry containments are not. susceptible to abrupt failure due to
. hydrogen burns and steam spikes, and the probability of steam explosions or
- failure to isolate containment is small.
5 f.
s
- NUREG-0869,- Revision 1 E-3 October 1985
For subatmospheric containments, prevention of eventual overpressure failure depends on connecting the sprays to an alternative source of water.
Sub-atmospheric containments do not have adequately sized heating, ventilation, and air conditioning (HVAC) systems to meet the energy addition to containment associated with a large-break LOCA.
Noncondensible gas production due to core-concrete reactions could overpressurize containment in a day or more. Hydrogen burns associated with invessel hydrogen production are.not expected to cause the containment to fail, but late hydrogen burns with the containment at an elevated partial pressure of steam and/or noncondensible gases may threaten containment integrity.
However, if inplant capability to detect flow degradation exists and corrective operational procedures have been developed, there is time to prevent eventual overpressure failure by operator action.
In an ice condenser containment, powered hydrogen ignitors would protect against a deflagration threat.
However, because there would be water in the reactor cavity, containment failure due to steam production would occur at the time of vessel failure or within a few hours thereafter, depending on the amount of ice remaining at the time of core melt.
The assumption is made in this analysis that refueling plugs have been removed in accordance with Technical Specifications and hence a wet cavity would exist at the time of vessel failure.
The staff's base case assessment of BWR consequences is equivalent to a 100%
probability of severe radiological releases for Peach Bottom and an 80%
probability for Grand Gulf.
Current understanding of containment response indicates that lower probabilities are possible.
The Browns Ferry BWR Mark I containment has a specific design feature that allows the operator in the control room to connect the spray suction line to the condensate storage tank.
However, it would have to be verified that all other Mark I containments have the same-fature before generic credit for a radiological release reduction if this feature is taken.
It would also have to be verified that procedures are in place to ensure that in the event of failure of recirculation sprays, the proper realignments and the necessary interlocks overridden within the available time.
BWR Mark II and III designs do not have the option of taking spray suction from the cor.densate storage tank.
It is possible to align spray suction to the fuel ' storage pool, but the staff is not aware that the necessary procedures are in place for such plants.
Furthermore, a source of makeup water to the fuel storage pool would have to be found.
The physical plant layout for Mark III design has the advantage that, for the most likely mode of failure, the fission products will bubble through a subcooled suppression pool, yielding a radionuclide release significantly lower than BWR-1. -The resulting consequences (in terms of person-rem) would be reduced substantially.
In addition, no credit is being given for scrubbing effects beyond that assumed in prior BWR analyses.
A similar mechanism for release reduction for Mark I and II containments would result if the operator relieves pressure by venting the containment through existing wetwell penetrations to the atmosphere.
NUREG-0869, Revision 1 E-4 October 1985
Current emergency procedure guidelines (EPGs) for BWRs describe containment venting.
Forthcoming BWR owners group EPGs will discuss wetwell venting in greater depth, and it appears that controlled venting will be adopted by all BWR owners.
For each of the containment types, however, there are credible mechanisms that would allow fission products to bypass the suppression pool under certain circumstances.
In the Mark I, drywell failure can occur either at the time of core melt as a result of overpressurization, or several hours later when penetration seals fail as a result of high temperatures.
In some Mark II designs the molten core would be retained on the diaphram floor and contain-ment overpressure failure would occur in the drywell.
Hence a direct path for some radionuclides could occur.
In the Mark III design, direct leakage from the drywell to the wetwell can result either from existing measured leakage paths or from failure to isolate the drywell.
Reassessment of Radiological Releases A wealth of experimental and theoretical research on radiological releases in severe accidents has been performed in the past decade.
The NRC Office of Research (RES) has developed a revised source term methodology based on that researcn and has documented the results of sample accident sequences for several reference plants (BMI-2104)..A committee of the American Physical Society has reviewed the methods and results, and committee members have con-cluded, among other things, that there is considerable uncertainty in some aspects of the methodology.
They further stated that the radiological release estimates for some sequences could be higher than predicted by WASH-1400 (NUREG-75/014).
When the new results and their associated uncertainties are integrated into the-analysis of specific sequences for specific plants, the calculated releases are generally lower than previously predicted, but the uncertainties are large.
For the purpose of this assessment, the staff confines its consideration to those cases in which the new source term methods unambiguously predict lower releases of radionuclides.
For sequences in which containment failure is predicted to occur long after the core has melted, the new methods predict considerable decreases in the suspended aerosol concentrations in containment, because of enhanced agglomeration and gravitional settling.
During the early and intermediate time period (less than 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />) after core melt, there are competing mechanisms that would tend to somewhat offset this reduction: most notably, enhanced releases of refractory fission products during core-concrete interactions in some reactor cavity designs, and the possible revaporization of fission products deposited in the primary system.
The latter mechanism is not a substantive consideration for A-43, because there is very little primary system retention of fission products in large-break LOCA sequences.
NUREG-0869, Revision 1 E-5 October 1985
For large dry containments with fan coolers, containment failure would occur more than a day after core melt, if at all.
The revised source term methodology would predict significantly lower radiological releases and offsite consequences than the staff assumed in this base case and prior A-43 analyses.
Other large dry containments and subatmospherics fall into two categories:
those in which the reactor cavity would be full of water following sump failure, and those in which it would be dry.
In the wet cavity case, the core-concrete interaction would be suppressed, and the source term would be greatly reduced when containment fails as a result of steam overpressurization (estimated to occur at about 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> after core melt).
For dry cavity designs, enhanced production of refractory metals during core-concrete interaction could occur, but the predicted containment failure time--if the containment fails--is much later because of the absence of steam production.
Consequently, the staff also expects a greatly reduced source term for dry cavity cases.
Because the enhanced aerosol removal does not affect noble gases, organic iodine, or gaseous fission products, the staff would limit the predicted reduction in offsite person-rem within 50 miles to about a factor of 10 for large dry and subatmospheric PWR containments.
Because containment failure in an ice condenser is estimated to occur early after core melt, the staff cannot be confident that the radiological releases are any lower than predicted in the base case.
For BWRs, the containment failure times are generally much less than 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br />, and the enhanced aerosol settling may not be as significant as estimated for PWRs.
Furthermore, for early containment failure, the releases of refractory metals can be significantly higher than previously assumed in the Mark I containment.
Consequently, the staff does not expect significant reductions in the predicted fission product releases for recirculation failure in BWRs as a result of the revised NRC source term methods.
Summary On the basis of its reassessment of containment performance and radiological releases, the staff has developed revised estimates of the offsite consequences for each reactor type.
Because of the very approximate nature of the review, the results are quoted with only order-of-magnitude accuracy. That is, the assessment determines whether the average consequences are about the same as a severe release (BWR-3 or PWR-3), an order of magnitude less (x 0.1), or two orders of magnitude less (x 0.01). The results for PWRs are shown in Table 4 and for BWRs in Table 5.
NUREG-0869, Revision 1 E-6 October 1985
TABLE 4 Approximate consequence reduction factors
- for PWR containments based on reassessment of containment response and source terms +
Without spray With spray Containment type recovery recovery _
Large. dry designs with fan 0.01 coolers Large dry designs without fan 0.1 0.01 coolers and subatmospherics Ice condensers 1
Because of-the approximate nature of the revised consequence estimates, they are quoted with only order-of-magnitude accuracy.
Revised consequence values can be obtained by multiplying the
+
reduction factors with the consequence estimates associated with PWR-3 (5.4 x 106 person-rem).
TABLE 5 Approximate consequence reduction factors
- for BWR containments, based on reassessment of containment response and source terms +
Without With wetwell wetwell Containment venting venting Mark I 1
0.1 Mark II 1
0.1 Mark III 0.1 0.1 Because of the approximate nature of the revised consequence estimates, they are quoted with only order-of-magnitude accuracy.
Revised consequence values can be obtained by
+
multiplying the reduction factors with the consequence estimates associated with BWR-3 (5.1 x 106 person-rem)
NUREG-0869, Revision 1 E-7 October 1985
Large dry containments with safety grade fan coolers are not estimated to fail.
Furthermore, in the event of a failure, radionuclide releases would be greatly reduced because of the long time to failure.
For other_ large dry and subatmospheric containments, failure could occur within
-a day after core melt, but the reduction in source terms due to the enhanced gravitational settling.would lead to an order of magnitude (x 0.1) reduction in consequences.
A further order-of-magnitude reduction (x 0.01) would result if containment failure were prevented by connecting the sprays to an alternative water source.
Ice condensers are expected to fail at the time of vessel failure, or a few hours thereafter, depending on the amount of ice remaining at the time of vessel failure. While source terms could be lower than predicted in WASH-1400, this circumstance would not, lead to an order-of-magnitude reduction in the pre-dicted person-rem.
Furthermore, recovery of spray operation before containment failure could not be ensured, because containment failure may be rapid.
Con-sequently, the staff concludes that the consequences of sump failure in this type of plant could be more severe than the estimates for other PWRs.
BWR Mark I and Mark II containments are expected to fail within a few hours after core melt, and the radionuclide releases are expected to be on the same order as for BWR-3 (Table 3).
Failure could be later and releases could be lower for some Mark II plants, but because of the variability of the Mark II designs and the uncertainty of the phenomenology, the staff cannot draw this general conclusion.
In both the Mark I and Mark II, successful venting of the wetwell prior to containment failure could lead to substantial source term retention in the suppression pool and a dramatic reduction in con-ditional consequences.
A similar reduction would result if the drywell sprays could be connected to an alternative source of water.
The staff has limited its consequence reduction factor to one order-of-magnitude, (x 0.1) because there is some possibility that the suppression pool may be bypassed because of early containment failure or leakage from the drywell.
The Mark III design fails in such a way that the fission products are channeled through the pool before release to the environment, with or without wetwell venting. As with the Mark I and Mark II plants, the reduction factor is limited to 0.1 because of the possibility of pool bypass.
The revised estimates of offsite consequences are shown in Table 6.
They are obtained by multiplying the reduction factors in Table 4 for PWRs and Table 5 for BWRs by the consequences associated with the PWR-3 and BWR-3 releases, respectively.
1 NUREG-0869, Revision 1 E-8 October 1985
TABLE 6 Revised estimates of offsite consequences based on current understanding of containment response and radiological source terms Conditional consequences (person-rem)
No spray recovery With spray i
PWR containment type recovery Large dry (safety grade fan coolers) 5 x 104 Other large dry and subatmospheric 5 x 105 5 x 104 Ice condenser 5 x 106 Conditional Consequences (person-rem)
No venting er With venting or BWR containment type spray recovery spray recovery Mark I 5 x 106 5 x 105 Mark II 5 x 106 5 x 105 Mark III 5 x 105 5 x 105 e
Estimation of Offsite Releases Utilization of the estimated consequences for PWR ice condenser plants shown in Table 6, without consideration of plant-specific sump design features and recirculation flow requirements, could lead to the conclusion that consequences associated with a failed sump are 10% higher.
Table 7 provides an overview of ice condenser plants; the significant factors are:
(1) All ice condenser plants utilize reflective metallic insulation on the primary coolant piping and major components (steam generators, pressur-izer, reactor coolant pumps, etc.); thus debris blockage concerns assoc-iated with transport of fibrous insulation debris are not present (at least not in significant amounts).
(2) The majority of these plants have lower recirculation flow rates and larger debris screen areas.
The net effect is that approach velocities are less than 0.2 ft/sec; therefore debris transport is not likely to occur.
(3) Net positive suction head (NPSH) margins are high, significantly higher for most plants than the 1 to 5 feet of available NPSH margins utilized in the-sump failure probabilities discussed in Appendix D.
NUREG-0869, Revision 1 E-9 October 1985
These factors all lead to the conclusion that sump failure probabilities developed for PWRs with a mix of fibrous and reflective metallic insulation and low NPSH margins should be reduced for applications to ice condenser plants.
Interpolation within the values shown in Table 7 of Appendix D can be used to derive a sump failure probability (for ice condensers) in the range of 3 to 9 x 10-6/Rx yr. This value was used for estimating the consequences discussed below.
TABLE 7 Overview of ice condenser plant design an operational features RHR Debris NPSH Approach flow screen margin Insulation velocity Plant (gpm)
(Ft )
(ft water) used (ft/sec) 2 Catawba 1&2 6000 135 7.9 (RHR)* Reflective 0.10 Metallic D.C. Cook 1&2 6000 90 21.9 (RHR)
Reflective 0.15 Metallic McGuire 1&2 6800 120 6.9 (CSS)* Reflective 0.13 16.9 (RHR)
Metallic Sequoyah 1&2 9500 43 2.8 (RHR)
Reflective 0.49 Metallic Watts Bar 1&2. 8000 43 11.5 (RHR)
Reflective 0.41 8000 164**
Metallic 0.11
- RHR: residual heat removal; CSS: containment spray system.
- Trash rack area; this structure would intercept large size insulation debris before transport to the sump debris screen structure could occur.
Utilization of the consequence values in Table 6 and estimated blockage probabilities (from Appendix D) can be used to estimate averted risks.
Table 8 contains such estimates and the blockage probabilities that were developed.. These values were used to calculate the value-impact ratios discussed in Section 4.4 of the main body of NUREG-0869, Revision 1.
i i
k NUREG-0869, Revision 1 E-10 October 1985
TABLE 8 Overview of consequences associated with sump blockage 2
l Estimated Estimated Assumed Estimated 4
conditional blockage core melt risk averted **
consequences probability conditional (AR Containment type (person-rem)
(1/Rx yr) probabil'ity*
person-rem /Rx) 3 to 50x10-8 0.5 PWR dry w/SGFCs+
^~
PWR ice condenser 5x108 1 to 9x10-8 0.5 40 to 560 PWR dry w/o SGFCs and subatmospheric 5x105 3 to 50x10-8 0.5 19 to 313 PWR dry w/o SGFCs 5x104 3 to 50x10-8 0.5 2 to 31 and subatmospheric, w/ spray recovery Mark I and II 5x108 4 to 20x10-8 0.5 250 to 1250 Mark III 5x105 4 to 20x10-8 0.5 25 to 125 Mark I and II 5x105 4 to 20x10-8 0.5 25 to 125 w/ venting or spray recovery
- The assumption is made that 50% of the time that blockage occurs, core melt would occur.
This assignment of a conditional core melt probability is realistic in view of potential operator dection and mitigating actions which could be taken.
- An outstanding reactor life span of 25 years has been assumed.
+ SGFC: safety grade fan cooler References U.S. Nuclear Regulatory Commission, NUREG-75/014, " Reactor Safety Study," 1975 (formerly WASH-1400).
BMI-2104, "Radionuclide Release Under Specific LWR Accident Conditions", by Gieseke, Cybulskis, Denning, Kuhlman and Lee, Battelle Columbus Laboratories,
[
Columbus, Ohio, July 1983.
NUREG-0869, Revision 1 E-11 October 1985
APPENDIX F CONTAINMENT SURVIVABILITY NUREG-0869, Revision 1 October 1985
CONTAINMENT SURVIVABILITY The several different containment design concepts currently in use for the many operating plants can be grouped as follows:
(1) The dry containment structures for pressurized water reactors (PWRs) must absorb all loads from accident conditions.
The results are characterized by large containment volumes (i.e. 2 x 106 ft3) and high design pressures (i.e. 60 psig) with considerable margin beyond the design point.
(2) Containment structures which incorporate a means of passive steam con-densation have taken advantage of this approach to design smaller and lower pressure containment buildings.
Boiling water reactor (BWR) con-tainments utilize a water pool for condensing steam and PWR ice condenser plants utilize an annular ice bed for condensing steam.
Containment Structural Capabilities Containment failure modes and attendant releases were analyzed in WASH-1400 and related to a major loss of containment capability through: (1) steam explosion induced failures (a mode); (2) hydrogen burn induced failures (p mode); (3) over-pressurization of the containment building resulting from steam generation (molten core interacting with water) and noncondensible gases (molten core interacting with basemat) (y mode); and (4) basemat penetration (c mode).
These WASH-1400 studies also assumed.a loss of containment without assessing the significance of any design margin available in the different containment structures as currently designed.
Risk assessments performed in recent years indicate that risk from nuclear power plants is dominated by severe core melt accidents.
Typical coatainment loading pressures and temperatures associated with whole core melt scenarios are on the order of approximately 100 psia and approximately 300 F.
More recently, mecha-nistic models for containment failure (see NUREG/CR-3653) have also been included in these assessments of severe accident scenarios.
[
These severe accident and containment structural capability studies provide the following insights:
L (1) Because of structural design margins, containments have inherent capabilities beyond their design basis. This provides a capability to contain or mitigate a wide spectrum of severe accidents.
(2) Best estimate analyses of containment performance indicate that containments can retain structural integrity at pressures as high as 2.2 to 2.5 times the design pressure.
Extensive yielding" is the term used in NUREG/CR-3653 to describe loss-of-structural integrity for these best estimate analyses.
NUREG-0869. Revision 1 F-1 October 1985
-. ~ -
(3) Although it is possible that leaks through penetrations could occur before loss of structural integrity, the risks from such leaks would be considerably smaller than from the gross containment failures assumed in WASH-1400 studies.
Containment Heat Removal Systems Makeup water flow is needed to absorb the energy released into the containment atmosphere or to the pools as a result of the long-term decay heat and to keep the core covered.
Although this flow is small compared to the design-basis accident flow, if recirculation flow to the vessel is redeced, boil-off will occur. The steam released into the containment will raise the pressure and temperature. This energy must be removed from the-containment so the containment pressure is maintained below its failure pressure.
The containment heat removal systems (CHRS) are sized to remove the energy released into the containment due to decay heat.
Depending on the containment type, any one or combination of the following systems might be used as the CHRS:
suppression pool cooling systems, containment atmosphere spray systems, and air cooling systems.
BWR plants (Mark I, II'and III containments) consider the pressure suppression
. pool as the short-term heat sink for both the blowdown energy and the decay heat.
In the long term, the pool cooling system is the most important CHRS in limiting the maximum pool temperature (and pressure for Mark III).
Reducing the flow to the suppression pool water cooling system because of debris breakage would result in increasing the pool temperature.
If the reduction of the flow rate is small (some few percent), the temperature increase would be small because of heat ex-changer capacity margins.
However, a major reduction (a factor of 5 to 10) in flow rate would very soon lead to a temperature increase in the pool beyond design limits.
This increase would result because of the strong connection between mass flow rate through the heat exchanger and overall heat transfer coefficient of the. heat exchanger, together with the reduced flow rate.
Containment atmosphere spray systems are used in all types of containments for
-both fission product and energy removal from the atmosphere in accident situations.
The relative importance of the spray system is dependent on the containment type.
In ice condenser containments the spray system is the only means to remove the decay heat released as steam into the containment atmosphere after the ice beds have melted.
A reduction of the containment spray flow rate would immediately lead to higher containment pressure and temperature in addition to a reduction in the' atmospheric fission product removal effectiveness.
This change would result because of several factors:
(1) smaller total amount of water flow, (2) bigger droplet sizes because of smaller pressure drop over the spray nozzles, (3) shorter stay time in the containment atmosphere, (4) smaller totai heat transfer area of the droplets, (5) loss.of thermodynamic equilibrium between containment atmosphere and droplets, and (6) smaller coverage by spray in the containment.
It is expected that a reduction of containment spray flow rate would cause a significant reduction of spray cooling efficiency.
NUREG-0869. Revision 1 F-2 October 1985
[
For Mark I and II designs, the spray systems are not required to mitigate the course of the transient.
Mark III containments do however rely on sprays, primarily for the short-term mitigation.
As a result, loss of spray system effects can be considered to be of secondary importance for BWR desi'gns.
Dry containments usually employ both sprays and fan coolers.
Each system is. sized for 100% capacity.
Therefore, the coolers could perform the same heat removal function without the sprays being operational.
Safety grade air coolers, typical for dry containment designs, are not affected by a reduction of residual heat removal (RHR) system flow rate and therefore would not be affected by loss of this system.
However, the availability of containment air coolers, which are not safety grade and are not designed to operate during post-LOCA accident situations, cannot be independently relied upon.
Because of the elevated pressure and steam-air mixture density in the containment, such fcns would be expected to trip because of overload.
Another factor would be, of course, the very questionable survivability of the related electrical components not qualified for the post-accident environment.
Expected Containment Design Response Due to Loss of Recirculation Water Sources The influence of reduced RHR-system flow rate upon the overall containment behavior is a plant-specific matter, depending on the containment design (Mark I, II, III, ice condenser, or dry containment).
Mark I and II containments are both small and quite similar in their pressure responses following an accident.
The peak pressure is reached very early, and
~ after one hour the pressure is already low compared to the design pressure.
However, the pool temperature is high and still increasing, and it would probably be-the first limiting design parameter associated with reducing the RHR system flow rate that would be exceeded.
This elevated water temperature to the RHR pumps could jeopardize their operation.
Thus, a total loss of the RHR system, or a major reduction of the flow rate, would result in a slow overpressurization and potential failure of the containment.
This failure could be expected in less than 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br />, if the containment is not vented as recommended by the 4
emergency procedure guidelines (EPG).
Basedoncurrentengineeringjudgment, a rupture of the containment can be expected to occur when pressure reaches 2 to 3 times the design pressure.
l In Mark III containment, the pressure and pool temperature are still high and increasing after 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />.
A major reduction of RHR system flow could mean that both the containment pressure and pool temperature would exceed the design values and would result in the same consequences discussed above for Mark I and II plants.
l
.Without venting, containment failure could be expected during the first 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following loss of recirculation flow capability.
The ice of an ice condenser containment will melt in about 1 to 2 hours2.314815e-5 days <br />5.555556e-4 hours <br />3.306878e-6 weeks <br />7.61e-7 months <br />.
After the ice is melted, the containment spray system is needed to keep the pressure below the design limit.
A reduction in the RHR system flow rate could also NUREG-0869. Revision 1 F-3 October 1985
cause the containment design pressure to be exceeded in this type plant.
Reduction of the RHR system flow rate would occur in the same time period when the ice would be totally melted and'would result in rapid overpressurization and early containment failure. This could be expected in about 3 hours3.472222e-5 days <br />8.333333e-4 hours <br />4.960317e-6 weeks <br />1.1415e-6 months <br />.
The peak pressure in a dry containment (normal atmospheric or subatmospheric) is reached early in the transient, and, after 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, the pressure is expected to be very low.
The dry containments also have safety grade air coolers, which could maintain containment pressure and temperature below the design values and, therefore, a reduction of RHR system flow will not adversely affect the dry containment.
Thus, containment failure for large dry containments with safety-grade fan coolers (SGFCs) would not be expected.
Severe Accident Study Insights For large dry PWR containments with safety grade fan coolers, the staff can con-clude that the containment response assumed previously in the A-43 consequence 2
analysis is conservative because the probability of containment failure as a result of overpressure has been overestimated.
Reaching a similar conclusion for BWR Mark I containments would be contingent on the licensee or applicant demonstrating that there is a high likelihood of the operator successfully exercising the option to align the containment sprays to take suction from the condensate storage tank.
The staff is not aware of a generic design option for PWR subatmospheric and ice condenser designs and BWR Mark II and Mark III plants that allows the operator to readily switch spray suction to an alternative water source.
Although alternate i
water sources exist, and in some cases the piping appears to be in place, it remains to be demonstrated that the proper connection can be made, valves aligned, and safety system interlocks overridden within a reasonable time under accident conditions. Without the sprays, the probability of overpressure (or overtempera-
- ture) failure for these containments could be higher than previously assumed. The PWR subatmospheric containment design has an advantage over the ice condenser, Mark II and Mark III containments because its predicted failure time is much later than the failure time for the others. This additional time allows more time for operators to recover the sprays.
The BWR Mark III containment stands out because its principal failure mode leads to significant scrubbing of fission products in the suppression pool.
The reduction in conditional consequences that would result is offset somewhat by the staff's conclusion that the probability of early failure for Mark III plants is higher than prior estimates.
A similar mechanism for source term reductions for Mark I and Mark II containments would result if the operator can successfully relieve pressure by opening a wetwell vent. Without a plant-specific analysis, the staff cannot say with confidence that the containment response estimates developed for resolution of USI A-43 are conservative for BWR containments (see i
also Gieske, 1984).
NUREG-0869. Revision 1 F-4 October 1985
Conclusions In summary, the following situations appear to exist:
(1) PWR dry containments (with safety grade fan coolers) are likely to survive a core melt situation, even with a loss of the containment emergency sump.
Even without safety grade fan coolers, large dry containments will not overpressurize for 6 to 12 hours1.388889e-4 days <br />0.00333 hours <br />1.984127e-5 weeks <br />4.566e-6 months <br /> following a LOCA, leaving time for the operator to take corrective action to provide an alternate water source for containment spray (provided a loss-of-sump condition is detected and acted upon).
The availability of safety grade fan coolers to minimize containment pressure, thereby enhancing containment survival, is a significant factor.
(2) Mark I containments have a high containment overpressure survivability if alternate containment spray suction can be provided should loss of RHR suction occur.
However, Mark I containments are most susceptible to containment failures that lead to release consequences comparable to WASH-1400 (NUREG-75/014) releases.
(3) Controlled venting of any of the BWR containment designs will preserve structural integrity for all BWR containment designs, if such venting is implemented correctly.
(4) Subatmospheric PWR containments (without safety grade fan coolers) and PWR ice condenser plants are most susceptible to loss of integrity if the containment emergency sump is lost and core sprays cannot be recovered through use of alternate water sources.
References Gieske, J. A., et al., "Radionuclide Release Under Specific LWR Accident Conditions," BM-2104 (Volumes 1-7) Battelle Memorial Institute, July 1984.
U.S. Nuclear Regulatory Commission, NUREG-75/014, " Reactor Safety Study,"
(originall published in October 1975 as U.S. Atomic Energy Commission report WASH-1400.
....., NUREG/CR-3653, " Final Report, Containment Analysis Techniques, a State-of-the-Art Summary," March 1984.
NUREG-0869. Revision 1 F-5 October 1985
APPENDIX G ESTIMATION OF COSTS TO REPLACE INSULATION NUREG-0869, Revision 1 October 1985
ESTIMATION OF COSTS TO REPLACE INSULATION The monetary costs of insulation replacement are dependent on plant design, material, and preplanning efforts.
Table 1 shows the insulation replacement costs and estimated exposures that were used in the For Comment version of the USI A-43 value-impact analysis (NUREG-0869), which was published in April 1983.
Cost information received from the industry (Atomic Industrial Forum) and insulation vendors (0 wens-Corning Fiberglas and Diamond Power Company) is summarized in Table 2.
This information was used to develop the reinsulation costs shown in Table 3.
The variability in estimates from industry sources is large and further illustrates the variances in plants and installation experience.
The composite average estimates of $153/ft2 (Table 3) for insulation replacement can be compared to the earlier USI A-43 estimate of
$75/ft2 Table 3 also shows an estimated insulation replacement cost range of $92/ft2 to $244/ft2, The most severe monetary impact would result from a decision to replace all of the problem insulation in a given plant.
The PWR sump failure probabiTity study showed (NUREG/CR-3394) that insulation on the primary system piping and the lower one-third of the steam generator are the principal sources of debris that leads to unacceptable sump blockage conditions. Appendix D of this report discusses the significance of insulation on the primary system
> 10-inch diameter or greater), steam generators, reactor coolant piping (iid pressurizer that could lead to unacceptable debris screen pumps, a blockage.
Table 4 illustrates the variability of fibrous debris as a function of pipe size and was derived from the Salem plant analysis (NUREG/CR-3394).
Plant insulation variability (as installed) is illustrated for the Salem 1 and Maine Yankee plants as shown in Table 5.
Approximately 2400 ft3 (covering an area of 8200 ft2) of potentially troublesome insulations are identified. On the basis of these two illustrative plants, it is estimated that only 4400 ft2 to 5235 ft2 of insulation might have to be replaced rather than replacing all plant insulation.
In the cost impacts developed in Section 2 of this report, these amounts are used as the upper bounds of the amount of insulation that would have to be replaced for determining insulation replacement costs for a typical PWR.
NUREG-0869, Revision 1 G-1 October 1985
TABLE 1 Estimates of costs to replace insulation and associated exposures illustrating plant variability Unencapsulated Cost est* Cost est** Cost est+ Estimated ++
insulation 1
2 3
exposure Plant (ftz)
($x103)
($x103)
($x103)
(person-rem)
Salem 1 13,200 281 238 660 99 Maine Yankee 6,700 142 121 335 47 Gir.na 1,000 21 18 50 8
Millstone 2 1,300 28 23 65 10
- These costs are derived from Surry 1 and 2 steam generator removal and reinstallation data, and discussions with onsite staff. ~ A per-unit rate of 0.85 hr/ft2 for replacing insulation was derived, and labor costs of
$25.00/hr were assumed.
- Telephone estimates from New England Insulation Company (Maine Yankee has employed this firm) were:
$3/ft2 to remove insulation, $11/ft2 to fabricate new panels, and $3 to $5/ft2 to install the new panel.
+ Telephone estimates of $35 to $50/ft2 for MIRROR
- insulation fabrication and instal'iation were obtained from Diamond Power, which supplies such insulation.
The value of $50/ft2 was used.
++ Exposure data were derived from Surry 1 and 2 data.
Discussions with the Surry. site staff indicate that a 50 person-rem exposure level for insulation replacement is realistic if the job is adequately preplanned.
An equivalent dose of 7 x 10 3 person-rem /ft2 of insulation to be replaced can be derived.
NUREG-0869, Revision 1 G-2 October 1985
TABLE 2 Insulation fabrication and installation cost estimates received during the A-43 (NUREG-0897) for comment period Commentor Comment / Estimate Atomic Industrial Forum Productivity:
Industry experience shows an average rate of 2.0 to 2.5 hr/ft2 for installation.
Labor:
Industry labor costs are $30 to
$45/hr.
Total:
The averaged cost of $550,000 per plant is off by at least a factor of 2; material costs are not included.
Diamond Power Company MIRROR *:
The cost of material supplied at (manufacturer of MIRROR
- Cooper was approximately $~0/ft2,
. insulation)
Productivity:
Installation rates averaged 1.24 hr/ft2 Labor:
The estimate of $25/hr used previously seems reasonable.
Owens Corning Fiberglass NUK0N*:
Material cost is estimated to Company (manufacturer of be $90/ft2, NUK0N* insulation)
Reflective:
An assumption of $100/ft2 would be very reasonable.
Labor:
Current labor costs of $40 to
$50/hr are common.
Productivity:
Installation rates of 7 to 12 ft2/hr are typical.
Total:
This leads to a cost of $112/ft2 for labor and $100/ft2 for material, for a total cost of $212/ft2 to remove existing insulation and replace it with reflective insulation.
NUREG-0869, Revision l' G-3 October 1985
y
,I'.
TABLEh
-Insulation replacement. cost estimates developed from public comments received Commentor Cost: Estimate Atomic Industrial Forum Reinsulation labor costs: $60 to $112/ft2 g
Insulation fabrication cost: $75/ft2 i
Combined total: $135 to $185/ft2
\\[
Average combined total: $160/ft2 Diamond Power Company '
Installation labor costs: $31/ft2 Insulation fabrication costs: $30 to 50/ft2 Subtotal: $61 to $81/ft2
,e As umed removal costs: $31/ft2 s
- s/
Combined total: $92 to $112/ft2 Average combined total: $102/ft2 4
3 j
Owens-Corning Fiberglass Reinsulation labor costs: $60 to $144/ft2 Company l!j-S Insulatio,n fabrication cost: $90 to $100/ft2
\\
Combined total: $150~to $244/ft2 s
t c
Averaged combined total: $197/ft2 Composite of all three estimates Averaged total: ($160 + $102 + $197)/3
= $153/ft2
~*
4
.k k
g NUREG'0869, Revision 1 G-4 October 1985 s.
,\\
.l TABLE 4 Maximum LOCA generated debris summarized by break size Pipe dia-Total fibrous Total all meter (in.)
debris _(ft3) type (ft3) 2 1
l' 6
2 22
~8 2
3 10 4
31 14 227 227 16 270 270 32 144 295 34 315 726 36-118 402 Notes:
(1) These values correspond to break locations in the primary system within the crane wall and represent the largest quantity of debris generated by a single break of a given pipe diameter.
(2) The insulation types and distribution within containment are those used in Salem 1.
All insulation within 7 L/Os of a break location is assumed to be destroyed and released as fragmented debris.
Source:
NUREG/CR-3394 NUREG-0869, Revision 1 G-5 October 1985
TABLE 5 Typical volumes of primary system insulation employed
- Salem Maine Yankee Component Volume Type of Volume Type of (f t3) insulation (ft3) insulation Steam generator 1284 reflective metallic /
1144 calcium silicate /
fibrous fibrous Hot leg
-160 reflective metallic 149 fibrous Cold leg 140 reflective metallic 156 fibrous Crossover 60 reflective metallic 279 fibrous Pressurizer 160 reflective metallic 302 calcium silicate /
fibrous Pressurizer surge 129 reflective metallic 57 calcium silicate /
line fibrous Reactor coolant 570 reflective metallic 149 calcium silicate /
fibrous Bypass N/A N/A 88 fibrous TOTAL **
2503 2324 SUBT0TAL***
1284 1527 (excluding reflec-(4402 ft2)
(5234 ft2) tive metallic-and calcium silicate) i
- This table is based on information provided by the operators in 1981.
Plant changes since 1981 have made the data less accurate for these specific reactors.
However, as representative data for reactors in general, the table is still valid.
- This volume includes all of the insulation that could be hit by a water _ jet from a LOCA pipe break (in pipes 210-inch diameter). -If the volume were restricted to insulation within L/D = 7 of a break, it might be signif-icantly smaller.
- For conservatism, Salem's steam generator is assumed to be covered entirely with fibrous insulation; 50% of the insulation on Maine Yankee's steam generator, pressurizer, and reactor coolant pump is assumed to be fibrous.
l NUREG-0869, Revision 1 G-6 October 1985
Using these revised insulation replacement costs and the revised estimates of the amount of insulation that must be replaced results in the cost estimates work sheet shown in Teble 6.
A cost range of $300,000 to
$1,300,000 (average cost = $750,000) is arrived at and is used in the value-impact calculations contained in Section 2.2 of this report.
Other estimated costs associated with plant evaluations and potential backfits are shown in Figure 1.
These estimated costs were derived as follows:
(1) Utilization of information provided in RG 1.82, Revision 1 and NUREG-0897, Revision 1 provides a rapid means for estimating sump air ingestion and debris potential.
An evaluation impact of $10,000 (see 1 in Figure 1) is estimated for plants having high post-LOCA water levels (which prevent air ingestion) and low containment recirculation flow velocities (i.e., 1 0.15 ft/sec) (which precludes debris transport).
(2) Should air ingestion pose a problem, the use of vortex suppressors has been shown experimentally to reduce air ingestion to zero; the relevant information is provided in RG 1.82, Revision 1 and NUREG-0897, Revision 1.
The design, fabrication, and installation of a vortex suppressor (consisting of commercially available floor grating materials, either installed horizontally or formed into a cage) are estimated to cost $35,000 to $50,000 (see 2 in Figure 1).
(3) Debris blockage problems can be assessed in two steps.
The initial step--based on limiting calculations as described in RG 1.82, Appendix A and NUREG-0897, Revision 1--is estimated to cost $15,000.
Should a second step--a detailed debris generation, transport, and screen blockage analysis--be required, the cost would be higher.
Plant-specific analyses in USI A-43 studies were on the order of $35,000 to
$50,000 per plant analyzed.
Thus, this cost impact is estimated to range from $25,000 to $65,000 (see 3 in Figure 1).
NUREG-0869, Revision 1 G-7 October 1985
TABLE 6 Estimated insulation replacement costs Fibrous Amount Estimated cost
- Plant insulation requiring
($ thousands) employed (ft2) replacement (ft2)
High Avg Low Salem 2 13,200 4,400**
880 675 440 Maine Yankee 6,700 5,235**
1050 815 525 Millstone 2 1,300 1,300 (assumed) 260 200 130 AVERAGED COST 730 565 365 AVERAGED COST (w/o Millstone 2) 965 745 485 Estimated cost variance ($ thousands)
Case 1 (averaged cost of three plants)
= 475
'940 m 730(+30%)
- 565
> 365 y
475
_365(-35%)
- - 240 Case 2 (averaged cost without Millstone 2) 1255 4 965(+30%)
- - 630 a
965 745
-485 4
630 4 485(-35%)
= 315 Case 3 (rounded values from Case 2, used for value-impact ratio calculations) 1300 1000 650 L
n a
j 1000 =
750 x 500 y
y p
u i
650 500 300
- Utilizes cost range shown in Table 3.
l l
- See Table 5.
NUREG-0869, Revision 1 G-8 October 1985 I.
m.
i v8 PWR Vortex Suppresscr-Required to
=@
r Reduce Air Ingestion E.
E Sump Hydraulics
/
Air Ingestion <2%
8 Sump Design
' Check per RG 1.82,.Rev. 1 Assessment,
=_@
per RG 1.82, Initial Debris Screen Blockage Rev. 1
% Assessment Effects Minimal per RG-1.82, Rev. I h 1
(per limit analyses)
Detailed Debris Acceptable
,P
'y Generation and,
NPSHImpact-+-@
Blockage Analysis Needs Estimated Costs Minimal G
$10,000 r Retrofits,---*@
$35,000 - $50,000 or Action
$25,000 - $65,000
$250,000 - $350,000 hardware costs
$300,000 - $1,300,000 (for replacement of large Replace-quantities of insulation)
= Problem R
Insulation C
Hi G$
Figure 1 Actions required to determine sump design acceptability 4
)
i
1 1
Should the debris ass'ssment calculations show a need for' plant' (4) e f
modifications, consideration logically should be given to alternatives that would be'less costly than replacing large quantities of insulation.
Because preservation.of the net positive suction head margin ~is the key
?
criterion, two ways the plant-could be modified are 4
Increasing debris screen area to_ reduce the impact of a loss of ji pump suction head caused by blockage.
This could be done by1 enlarging the sump or intake. screen. Along the same lines, use of screens upstream of the currently installed screen would have a y
two-fold benefit:
it would intercept undesirable debris at some-
. distance from the sump location and it would reduce the impact of-L head loss because of the reduced a'pproach velocities associated with the enlarged upstream screen area.
Such a hardware backfit is estimated to cost $250,000 to $350,000 (see 4 in Figure 1).
Re-examining the recirculation flow rates required for the long-term recirculation mode, possibly reducing the. currently established flow rates (which are' set by immediate post-LOCA flow -
requirements), and submitting the re-analysis of long-term recirculation ~needs. -This option can be considered an-analytical backfit, and the cost of such an analysis is estimated to range from
$25,000 to $65,000.
(5) 'The estimated. costs for replacement of insulation are $300,000 to
$1,300,000 (see 5 in Figure 1); these represent the major cost impact and are discussed at the beginning of.this section.
- References U. S.~ Nuclear Regulatory _ Commission, NUREG/CR-3394, "Probabilistic Assessment-of Recirculation Sump Blockage Due to loss-of-Coolant Accidents," July 1983.
E
' NUREG-0869, Revision 1 G-10 October 1985 W
--W r
?
e+ew K
ei-r-au-qu
?
-wt- * -
---Sw-
-*e-----
sur YP-ev-*'
a
--vw-w wri.
e r-eWLv
- a+r.-
APPENDIX H DRAFT GENERIC LETTER 4
NUREG-0869, Revision 1 October 1985
DRAFT TO ALL LICENSEES OF OPERATING REACTORS, APPLICANTS FOR OPERATING LICENSEES, AND HOLDERS OF CONSTRUCTION PERMITS.
Gentlemen:
SUBJECT:
P0TENTIAL FOR LOSS OF POST-LOCA RECIRCULATION CAPABILITY DUE TO INSULATION DEBRIS BLOCKAGE (Generic Letter 85-
)
This letter is to inform you about a generic safety concern regarding LOCA -
generated debris that could block PWR containment emergency sump screens or BWR.RHR suction strainers, thus resulting in a loss of recirculation or containment spray pump net positive suction head (NPSH) margin.
The potential ~ exists for a primary coolant pipe' break to damage thermal-insulation on the piping as well as that on nearby components.
Insulation debris could be transported to water sources used for long-term post-LOCA recirculation and containment sprays (i.e., PWR containment emergency sumps and BWR suction intakes in the suppression pools) and deposited on debris screens or suction strainers.
This could reduce the NPSH margin below that required for recirculation pumps to maintain long-term cooling.
This concern has been addressed as part of the efforts-undertaken to resolve USI A-43, " Containment Emergency Sump Performance." The staff's technical findings contain the following main points.
Plant insulation surveys, development of methods for estimating debris generation and transport, debris transport experiments, and information provided as public comments on the findings have shown that debris blockage effects are dependent on the types and quantities of insulation employed, the primary system layout within containment, post-LOCA recirculation patterns and velocities, and the post-LOCA recirculation flow rates.
It was concluded that a single generic solution is not possible, but rather that debris blockage effects are governed by plant specific design features and post-loca recirculation l
flow requirement.
The current 50% screen blockage assumption identified in Regulatory Guide (RG) 1.82, " Sumps for Emergency Core Cooling and Containment Spray Systems,'.' should be replaced with a more comprehensive requirement to assess debris effects on a plant-specific basis.
The 50% screen blockage assumption does not require a plant-specific evaluation of the debris-blockage potential and usually will result in a non-conservative analysis for screen blockage effects.
The staff has revised Regulatory Guide (RG) 1.82, Revision 0, " Sumps for Emergency Core Cooling and Containment Spray Systems" and the Standard
' Review Plan Section 6.2.2, " Containment Heat Removal Systems" based on the NUREG-0869, Revision 1 H-1 October 1985
A DRAFT above technical findings.
However, the staff's regulatory analysis (NUREG-0869, Revision 1, "USI A-43 Regulatory Analysis") evaluated (1) containment designs and their survivability should loss of recirculation occur, (2) alternate means to remove decay heat, (3) release consequences (which were based on pipe break probabilities which did not incorporate insights gained from recent pipe fracture mechanics analyses), and (4) cost estimates for backfits considered (i.e., reinsulating).
This regulatory analysis did not support a generic backfit action and resulted in the decision that this revised regulatory guidance will not be applied to any plant now licensed to operate or that is under construction. The revised guidance will be used on Construction Permit Applications, Preliminary Design Approval (PDA) applications, and applications for licenses to manufacture that are docketed after six (6) months following issuance of RG 1.82, ized Revision 1, and Final Design Approval (FDA) applications, for standard designs which are intended for referencing in future Construction Permit Applications, that have not received approval at six (6) months following issuance of the RG 1.82, Revision 1.
Although the staff has concluded that no new requirements need be imposed on licensees and construction permit holders as a result of our concluding analyses dealing with the resolution of USI A-43, we do recommend that RG 1.82, Revision 1 be used as guidance for the conduct of 10 CFR 50.59 reviews dealing with the changeout and/or modification of thermal insulation installed on primary coolant system piping and components.
RG 1.82, Revision 1 provides guidance for estimating potential debris blockage effects.
If, as a result of NRC staff review of licensee actions associated with the changeout or modification of thermal insulation, the staff decides that Standard Review Plan Section 6.2.2, Revision 4 and/or RG 1.82, Revision 1 should be (or should have been) applied to the rework by the licensee, and the staff seeks to impose these criteria, then the NRC will treat such an action as a plant specific backfit pursuant to 10 CFR 50.109. It is expected 2
that those plants with small debris screen areas (less than 100 ft ), high ECCS recirculation pumping requirements (greater than 8000 gpm), and small NPSH margins (less than 1 to 2 ft of water) would benefit the most from this type of assessment in the event of a future insulation change.
RG 1.82, Revision 0 with its 50% blockage criteria does not adequately address this issue and is inconsistent with the technical findings developed for the resolution of USI A-43.
This information letter along with enclosed copies of NUREG-0897, Revision 1, RG 1.82, Revision 1 and SRP Section 6.2.2, Revision 4 should be directed to the appropriate groups within your organization who are responsible for conducting 10 CFR 50.59 reviews.
NUREG-0869, Revision 1 H-2 October 1985 L
1 DRAFT No written response or specific action is required by this letter.-
Therefore, no clearance from the Office of Management and Budget is required.
If you have any questions on this matter, please contact your project manager.
Hugh L. Thompson, Jr., Director Division of Licensing
Enclosure:
' NUREG-0897, Revision 1 RG 1.82, Revision 1 SRP Section 6.2.2, Revision 4 4
5 J
o NUREG-0869, Revision 1 H-3 October 1985
u NuC m.a m u m oa,Co - ss.oN
. K.,oR T NUM. -- r,oC. e., ve, N...,,
g,oR==
'"o',"2J*,
BIBLIOGRAPHIC DATA SHEET NUREG-0869 Revision 1 stE tNsTRUCTsONgN THE REVER$$
- 3. TITLE AN. SUS Til 3 LE AVE BLANK USI A-43 gulatory Analysis
[oATE RE ORT COM,LETEo MO M
YEAR Octo r 1985 Auf oR,,,
A. W. Serkiz I
- om an'oa' 55um p,0N1,.
i VEAR 0Aober 1985 T.,ERFORMINe ORGANi2ATION NAME A ' MAsLING AooRE55,,nes me la Co.,
S PR@CT/T ASK/ WORK UNtT NUM8ER se Division of Safety echnology
/
Office of Nuclear Re tor Regulation e ~ oa caANT Nu-a U. S. Nuclear Regulat Commission Washington, DC 20555
)
.. eN,eRi e ore.N., AT,eN NAM..No MA.L,Ne,
R.,,.., Cs,
,,. Tv.,e,R.eR, Division of Safety Techno~
Office of Nuclear Reactor ulation Final Regulatory Analysis U. S. Nuclear Regulatory Co
'ssion
.,iR,oo Cov R eo
,,-~s e.,.s, Washington, DC 20555 12 $U.,LEMENT AR Y NOTES
,3 AS$ T R AC T (200 woras er 'ess, This report consists of:
(1) the regu tory.alysis for Unresolved Safety Issue (USI)A-43,"ContainmentEmergencySump erf ance"; (2) the proposed resolution; (3) a summary of public comments receive action taken; (4) the Committee to Review Generic Requirements (CRGR) minutes elated to this USI; and (5) appendices that summarize assumptions, calculational hods, consequence analyses, and cost estimates used in this regulatory analysi.
,. ooCuME NT AN Alvi.....E vooRo. of cR, Tors i.Av,A g r.
USI A-43 Regulatory Analysis for USI'A 3 Unlimited
'6 SECURITY CL A85i,1 CATION IThen ce,e,
.oENTi.im,o,EN ENoto riaus Unclassified iTh<s reeertl Unclassified i,
vu.. R o,, Ac..
t$,ReCE
UNITED STATE 3
"'"C'^**"'^*
C'UCLEAQ RE2ULATORY COMMISSION
"YM" "
WASHINGTON, D.C. 201i65 WASMDC PflIGA4T Iso G 87 OFFICIAL BUSINESS PENALTY FOR PRIVATE USE. 4300 I
e l
i l
l 4
N t
_-_-- _...., _ _ - _ - - _