ML20195K400

Draft Technical Study of SFP Accidents for Decommissioning Plants

ML20195K400
Person / Time
Issue date:	06/16/1999
From:	NRC
To:
Shared Package
ML20195J463	List: ML20195J457 ML20195K400
References
NUDOCS 9906220082
Download: ML20195K400 (115)

Text

{{#Wiki_filter:"

• • 00 y

1 UNITED STATES j NUCLEAR REGULATORY COMMISSION WASH}NOToN, D.C. 20em.0001 June 16, 1999 MEMORANDUM TO: John A. Zwolinski, Director Division of Licensing Project Management Office of Nuclear Reactor Regulation FROM: Gary M. Holahan, Director N/// Division of Systems Safety and Analy s Office of Nuclear Reactor Regulation

SUBJECT:

PRELIMINARY DRAFT TECHNICAL STUDY OF SPENT FUEL POOL ACCIDENTS FOR DECOMMISSIONING PLANTS i in response to a request by the Nuclear Energy Institute at a meeting with public stakeholders on Monday, June 7,1999, the staff is releasing the attached preliminary draft of portions of the Technical Working Group's technical study of spent fuel pool accidents for decommissioning plants. The preliminary draft study includes the scope of the Technical Working moup effort, the assumptions used in the deterministic and probabilistic analyses, and basic insights. Since work is still in progress, the findings, conclusions and recommendations are not included in the released draft. It should be recognized that the preliminary draft study precedes the upcoming workshop with NEl and other public stakeholders in July of 1999, and j the independent, technical, quality review process. As such, it does not factor in comments from the public stakeholders or any other technical organizations. 9906220082 990616 PDR REVGP ERGNUMRC PDR

DRAFT DRAFT Technical Study of Spent Fuel Pool Accidents for Degppning Plants DRAFT Technical Working Group: Group Leader:Vonna Ordaz, DSSA/SPLB - Members: Goutam Bagchi, DE Christopher Boyd, RES Michael Cheok, DSSA/SPSB Edward Connell, DSSA/SPLB Tanya Eaton, DSSA/SPLB Edward Ford, DIPM/lOMB Christopher Gratton, DSSA/SPLB Ken Heck, DIPM/lOMB j Diane Jackson, DSSA/SPLB Glenn Kelly, DSSA/SPSB Larry Kopp, DSSA/SRXB l James O'Brien, DIPM/lOLB Jason Schaparow, RES Joseph Staudenmeier, DSSA/SRXB ) Edward Throm, DSSA/SPSB

DRAFT Technical Study of Spent Fuel Pool Accidents for Decommissioning Plants ggyntents EXECUTIVE S U M MARY................................................. 1 1.0 Introduction.. .................................. 5 2.0 Deterministic Evaluation of Spent Fuel Heatup............. ...7 2.1 Fire Protection Evaluation of the Phenomena of a Zirconium Fire......... 7 2.2 Spent Fuel Heatup Analyses . 10 2.3 Potential for Criticality...................................... 21 3.0 Risk Assessment of Spent Fu k at Decommissioning Plants......... 23 3.1 Beyond Design Basis Spent Fuel Pool Accident Scenarios (Intemal Events). 26 3.2 Beyond Design Basis Spent Fuel Pool Accident Scenarios (Extemal Events) 32 3.3 P.abebilistic Assessment insichts................................ 45 4.0 Supporting Information................. ............................. 48 4.1 M aintenan ce Rule........................................ 48 4.2 Quality Assurance........................................... 50 4.3 Water-basin Type Independent Spent Fuel Storage installations......... 51 4.4 Desig n Basis Events............................................ 54 5.0 Refere n ce s. ................................................... 56 6.0 Acronyms...... ........................................... 60 Appendix 1 Assumptions for ', pent Fuel Pool Probabilistic Risk Assessment............ A1-1 Appendix 2 Decommissioning Plant Spent Fuel Pool Event Trees and Fault Trees....... A2-1 Appendix 3 Assumptions for Spant Fuel Heatup Time Estimate.................... A3-1 Appendix 4 StructuralIntegrity of Spent Fuel Pool Structures Subject to Tomados and High Winds...................................A4-1 Appendix 5 Structural Integrity of Spent Fuel Pool Structures Subject to Heavy Load Drops........................................ A5-1 Appendix 6 Aircraft Crashes into Spent Fuel Pools.............................A6-1 l l

p-List of Tables Table 3.1-1 ' Loss of Spent Fuel Pool Cooling Initiating Events....................... 37 Table 3.1-2 Event. Probabilities / Frequencies Used in the SFP Risk Analysis........... 38 Table 3.1-3 Spent Fuel Pool Cooliag@RAftyTs Frequency of Fuel Uncovery (per year). 45 List of Figures Figure 2.2-1 Adiabatic Heatup Time g Time for a PWR................... 20 c o Time for a BWR................... 20 Figure 2.2-2 Adiabatic Heatup Time W. S Figure 3.0-1 Representative Spent Fuel Pool Model............................... 36 DRAFT J l 3 I

EXECUTIVE

SUMMARY

. Atter a nuclear power plant isp9 hut down and the reactor is defueled, the traditional accident sequences that domma at reactor risk are no longer applicable. The predominant source of risk remaining at permanently shutdown plants involves accidents associated with spent fuel stored in the spent fuel pool. Previous NRC-sponsored studies have evaluated unlikely severe accident scenarios that involve drainage of the spent fuel pool coolant

and shielding water that uncogr,tte speqt. fuel. Given certain combinations of spent fuel storage configurations and de:sf1A r: tee spent fuei assembiies couid heat up to a temperature at which rapid runaway oxidation of the zirconium fuel cladding (zirconium fire) might occur and result in cladding failure and a large offsite release of radioactive materials from the spent fuel. When evaluating the acceptability of license exemption requests from regulations in the areas of emergency preparedness and insurance for permanently shutdown plants, the staff has assesseg qbility of the spent fuel to a zirconium fire accident.

Since the exemptions were grant on a plant-specific basis, they have resulted in different analyses and criteria used for the bases of the exemptions. In some cases, the staff has requested heatup evaluations of the spent fuel cooled only by air, A temperature of 565 *C based on the onset of clad swelling was used as a conservative limit to ensure no radiological. release occurred. However, differences in licensee and NRC evaluations of spent fuel heatup

phenomena have resulted in questions from industry stakeholders regarding the NRC's technical and regulatory bases for evaluating specific spent fuel pool severe accident scenarios

' for decommissioning plants. To increase the efficiency and effectiveness of decommissioning requirements, the staff acknowledges that a predictable, risk-informed approach for rulemaking and for the review of exemptions during the interim period prior to the completion of rulemaking needs to be. established. This need was discussed with the Commission during a meeting on March 17,1999. On the basis of this need and the lack of a clear technical basis for evaluating severe accidents in spent fuel pools, the staff formed a technical working group to assess' the . existing technical and risk information on spent fuel pool accidents at decommissioning plants. The working group plans to develop a risk-informed technical basis to establish a predictable method for reviewing exemption requests and for followup rulemaking activities. The working group was also tasked to identify the need for any research in areas of large uncertainty. The staff considers that such an approach will contribute to safety and reduce unnecessary regulatory burden, as well as, increase public confidence. The working group performed generic calculations and assessments of the heatup of the fuel-and cladding to determine the potential for zirconium oxidation and ignition. The working group also performed a probabilistic analysis for spent fuel pool accidents at decommissioning plants. - If the decay heat of the spent fuel is sufficient to raise the temperature of the fuel and clad to the point that self-sustained zirconium oxidation would occur, then the heat of the oxidation reaction . would further increase the temperature to cause the ignition of the zirconium. A zirconium fire would raise the fuel temperature to sufficiently high temperatures that the radionuclides in the fuel could be released. The most extensive work to date was in support of Generic Safety issue (GSI) 82, "Beyond Design Basis Accidents for Spent Fuel Pools." [Ref. 3],' The working group used the studies in 1 1 i i.

l - References'1 and 2 as the starting point for its assessment in support of GSI 82. The working group noted that characteristics of a zirconium fire had not been well defined in previous SFP studies. Aliterature review fo

gd rzirconium fires bum with little flame, generate little smoke, but attain high temper gnison temperatures for zirconium were reported at temperatures between 1281 *C and 1600 *C. The heatup analyses performed in support of GSI 82 reported zirconium oxidation temperatures between 850 - 900 *C, The working group

. assumes that the latter temperature range is for the onset of oxidation that precedes runaway oxidation orignition. The wod ~ t p.c oted that the previous SFP studies provided good insights into the phenomena c ' Rh pxidation. The studies identified that the initiation of a zirconium fire was highly dependent on decay power and fuel storage configuration. In reviewing the studies, the working group identified that operating reactor spent fuel management practices may affect the calculated decay time in the spent fuel heatup analysis necessary to preclude a zirconium fire. Some of these practices include the increase in fuel bumup, which leads to higher decay power, gpl storage racking, which could reduce heat removal. The working group's preliminary results indicate that on a generic basis, the decay time required to maintain cladding temperature below the self-sustaining zirconium oxidation temperature using air cooling only may be longer than the generic studies performed for operating reactors. Decay time is generally defined as the length of time elapsed since reactor shutdown for the most recently discharged fuel. However, previous plant-specific emergency preparedness exemptions are unaffected because they were approved using analyses that reflected the actual conditions at the particular plant. The working group has identified two potential deterministic criteria for assessino the potential for a zirconium fire, which are described below. One potential criterion for allowing the reduction of existing regulation with respect to emergency preparedness (and possibly other regulatory areas) is the determination that the decay heat is suf6ciently low that air cooling is adequate to maintain the clad temperature below the point of self-sustained zirconium oxidation. The working group's preliminary estimates, using generic, near-bounding thermal-hydraulic spent fuel heatup assumptions, indicate that 3 to 5 years of decay time may be needed to reach a point at which air cooling of the fuel is sufficient to prevent self-sustained zirconium oxidation. The working group notes that a plant-specific analysis, using actual parameters such as decay heat and spent fuel pool configuration, should yield shorter time estimates..This type of analysis has been relied upon for several previous exemption requests using a criteria based on a maximum fuel temperature of 565 *C, which is the temperature for the onset of clad swelling. However, the working group recognized that the criteria to demonstrate that clad swelling could not occur was conservative. The working group reviewed the temperature criteria used in the spent fuel analysis, and the preliminary results j indicate that a maximum allowable temperature of 800 *C may be acceptable if certain boundary conditions are met. These conditions would include criteria such as demonstrating that the maximum temperature including uncertainties remained below the temperature limit, including a model for higher temperature effects, and demonstrating that a release of the radionuclides in the gap between the clad and the fuel was not a concem. The temperature limit was based on the lowest temperature for the onset of self-sustaining zirconium oxidation identified by the GSI,82 studies. The second potential criterion for allowing the reduction of existing regulation with respect to emergency preparedness is the determination such that sufficient time is available after the fuel 1 2

4 'is uncovered that mitigative actions, and if necessary, offsite protective measures could be taken for the public without preplanning. The working group performed generic, bounding calculations to correlate the d =--- 'iii6 fuel since final shutdown to the heatup time of the ~ fuel from uncovery to runawad bxidation. The calculations were conservatively based on the heatup time for the hottest rod to heat up from 30 to 900 *C assuming adiabatic conditions (no heat is lost). The working group recognizes that the conditions are not realistic because some heat removal would occur. However, it would encompass additional events such _as a piece of flat material fallirgrga.b6ilding wall or roof on top of fuel assemblies. This type ~ f calculation was used to supyo % piivious plant-specific emergency preparedness o exemptions. The workiag group's preliminary generic results indicate that at 2 years of decay time for a boiling-water reactor, and 2.5 years for a pressurized-water reactor, about 10 hours will be available for mitigative actions and offsite protective measures before onset of runaway zirconium oxidation. The staff previously determined that 10 hours was sufficient time to take mitigative actions and, if necog,gprotective measures without preplanning in its evaluation of a plant-specific emergency preparedness exemption. However, further evaluation is needed to determine whether a generic time to take protective actions without preplanning is appropriate for all plants (or a group of plants) or whether a site-specific time should be determined for each plant. The staff would need to coordinate with the Federal Emergency Management Agency in this evaluation, in addition, a more realistic calculation using - plant-specific parameters or not using adiabatic conditions could yield shorter estimates of the decay time needed to achieve a given heat up time. In the study, the working group identified that decay heat from Branch Technical Position (BTP) ' ASB 9-2 [Ref. 4] should not be used as the basis of decay heat inputs for fuel heatup calculations in spent fuel pools. BTP ASB 9-2 is only valid for times less than 10,000,000 seconds (115 days). The draft ANS-5.1/N18.6 - 1973 standard [Ref. 5], which is'the bases for BTP ASB g-2 may be used to calculate decay heat if the applicable uncertainties are included. However, BTP ASB 9-2 is non-conservative in the approximate time range of 1 to 4 years after shutdown compared to best estimate ORIGEN computer code calculations. The working group reviewed.the potential for criticality and the preliminary results indicate that - criticality is not a concem for decommissioning plants. Several scenarios were evaluated for PWRs and BWRs and were not considered credible. The only exception that could be identified was if Boraflex was significantly degraded and a raw water reflood of the spent fuel pool occurred assuming that the plant credits Boraflex for subcriticality margin. ' The working group also performed a preliminary probabilistic analysis of the initiating events that could lead to fuel uncovery at decommissioning plants. The analysis considered a wide range ^ of initiating events. An important assumption in this type of analysis is the amount of redundancy and diversity of spent fuel pool. heat removal systems, spent fuel pool makeup systems, and their support systems. On the basis of information gathered by the working group on site visits to four decommissioning plants,~ the system configurations analyzed had significantly reduced levels of redundancy and diversity in these areas relative to operating plants.. In the analysis, it was assumed the spent fuel had 1 year of decay time and the spent fuel pool support systems were representative of plants that had been shut down for 2 or more years. The working group notes that no decommissioning plant today matches the conditions assumed in this risk assessment. Also, the conditions assumed in the working group's risk 4 3 p

e - j - p 4 assessment do not apply to operating plants because of differences in areas such as support systems and personnel availability.' The analysis calculated the frequency of fuel uncovery. Frequency of fuel uncovery ispy lont to the frequency of a zirconium fire. A zirconium fire may result only from fuel arte duling a limited period after final shutdown. The working group's preliminary results indicate that a seismic event may not be the largest contributor to risk; there are several other credible initiators for decommissioning plants including: intamal fires; loss.of coolant inventories; cask drop; loss of offsite power; tomado missiles; loss of pool - cooling; and aircraft impact. gq 'a,of site visits, current decommissioning plant configurations, and probabilisw s Mescribed herein, the working group made a ~ preliminary estimate that the frequency of fuel uncovery is about 1E-5 per year. No single initiating event is dominant for uncovering the fuel and the frequency of many of the initiating events is heavily influenced by human error, o In support of the risk assessrggidiRing group reviewed the implementation of the-maintenance rule and quality assurance (QA) programs for decommissioning plants and the design of water-basin independent Spent Fuel Storage Installations (ISFSis). Implementation of the maintenance rule [Ref. 6] has been essentially the same as for operating facilities; however, the population of systems, structures, and components (SSCs) within the scope of the rule is. reducede For Appendix B QA programs, licensees generally retain their approved Appendix B QA programs [Ref. 7] through the decommissioning phase. Changes to the QA program commitments are submitted for NRC approval through the 50.54(a)(3) [Ref. 8] change control process. In addition, the working group reviewed the regulatory guidance for the design of a water-basin ISFSI, which is licensed under 10 CFR Part 72 [Ref. 9]. The concept of defense in depth is maintained in the design of the water-basin ISFSI oy mt(E g.means to protect the clad i as the fission product barrier and to alert personnel to the event that could cause a breach in the clad.'. Many systems and controls were identified to maintain defense in depth to protect against the release of radioactivity. Since the Commission meeting on March 17,1999, the staff has held three public meetings with stakeholders to discuss the mission of the working group, the outline of the team's effort, the l preliminary results as described herein,~ and the plans for the stakeholders* further involvement.

On June'7,1999, the working group presented the preliminary results to the Nuclear Energy Institute and the other stakeholders for discussion purposes only and stated that it would be premature to apply any findings to the regulatory process at this time. This preliminary draft of portions of the study, which include the scope of the study, the assumptions used in the deterministic and probabilistic analyses, and basic insights is being released to facilitate a public workshop on the study to be conducted in July of 1999. After the workshop, the working group plans to involve outside technical organizations with the report for an independent, technical, quality review to further refine the working group's assumptions and results. The review process is expected to be completed in December 1999. The working group will consider the information shared in the workshop and the resultc from the independent, technical, quality review process before completing the final report of the risk-informed technical basis for reducing existing regulatory requirements with respect to emergency preparedness,' insurance, and possibly other requirements at decommissioning plants in March 2000. Depending on the outcome of the workshop in July 1999, the staff may shorten the schedule for completing the final report.

4

r O j 1 1.0 Introduction f As the number of a decommi ' br ts increase, the ability to address regulatory issues generically has become more t.~ After a nuclear power plant is permanently shut down and the reactor is defueled, the traditional accident sequences that dominate operating reactor risk are no longer applicablei The predominant source of risk remaining at permanently shutdown plants involves accidents associated with spent fuel stored in the spent fuel pool. Previous NRC-sponsored stuMM5dluated unlikeiy severe accident scenarios that involve drainage of the spent fuel pool coolant and shielding water that uncover the spent fuel, i: Given certain combinations of spent fuel. storage configurations and decay times, the spent fuel assemblies could heat up to a temperature at which rapid runaway oxidation of the zirconium fuel cladding (zirconium fire) might occur and result in cladding failure and a large offsite release of radioactive materials from tg g When evaluating the acceptability of decommissioning licensee exemption requests from regulations in the areas of emergency preparedness and insurance for permanently shutdown plants, the staff has assessed the t susceptibility of the spent fuel to a zirconiu~m fire accident.' j Since the exemptions were granted on a plant-specific basis, they have resulted in different analyses and criteria used for the basis of the exemptions. In some cases, the staff has requested heatup evaluations of the spent fuel cooled only by air. A temperature of 565 *C based on the onset of clad swelling was used as a conservative limit to ensure no radiological release. This criterion was used because of national laboratory studies that had identified the 4 potential concem for a significant offsite radiological release from a zirconium fire which may 1 occur when all water is lost from the spent fuel pool. Zirconium is used in an alloy to make the hollow rods, called cladding, that hold the fuel pellets. The material is also used in boiling water reactors as a channel box around each fuel assembly, if the decay heat of the spent fuelis sufficient to raise the temperature of the fuel and clad to the point that self-sustained zirconium { oxidation would occur, then the heat of the oxidation reaction would further increase the temperature to cause the ignition of the zirconium. However, differences in licensee and NRC evaluations of spent fuel heatup phenomena have resulted in questions from industry stakeholders regarding the NRC's technical and regulatory bases for evaluating specific spent fuel pool severe accident scenarios on decommissioning _ plants. The staff acknowledged that a predictable, risk-informed approach for requesting and granting exemptions is needed. This need was discussed with the Commission during a meeting on March 17,1999. As a result, the staff formed a technical working group to develop a risk-informed technical basis to establish a predictable method for reviewing exemption request and for followup rulemaking activities. The working group was also tasked to identify the need for any research in areas of large uncertainty,. The staff considers that such an approach will contribute to siafety and reduce unnecessary regulatory burden, as well as, increase public confidence. g ~ The working group performed a generic deterministic calculations and an assessment of the . p' revious spent fuel pool studies. Section 2.0 provides a review of the available technical information for deterministic evaluations of the heatup of spent fuel when only air cooling is available; The working group's assessment includes the characteristics of a zirconium fire, the ' spent fuel failure criteria, existing spent fuel heatup analyses, available generic code evaluations, heatup calculation uncertainties and sensitivities, critical decay times for reaching 5' U

{, b sufficient air cooling, an estimated heatup time of uncovered spent fuel and the potential for L vitk,ety as a result of a zirconium fire event. The working group also perfos inary probabilistic snalysis of the initiating events that could lead to fuel uncovery and the consequences of a zirconium fire at decommissioning . plants..The analysis considered a wide range of initiating events and calculated the frequency of fuel uncovery. Risk is generally defined in terms of core damage frequency or large early release frequency from a rea r v n An assessment of spent fuel pool risk with no mly long decay times introduced new analytical issues for ' associated operating reactor the staff. Risk to the public depends on the event probability and consequences. For a plant permanently shut down, the probability of some events (such as a seismic event) do not change However, due to changes in equipment and personnel, other events may become more probable in a decommissioning plant. The consequences and, therefore, the risk to the public may be reduced though prov gigative actions. The working group has considered the initiating event probabilities a controls on certain risk parameters that could affect the safe storage of the spent fuel. Section 3 provides the preliminary results of the risk assessment, which evaluated intomal initiating events (e.g., intamal initiator frequencies, event trees and fault . trees, equipment failure conditional probabilities and human errors), and extemal events (e.g., seismic, tomado, and aircraft). The availability and reliability of systems, structures and components at decommissioning plants could affect the probability and consequences of events. As supporting information, the working _ group considered the effects of the Maintenance Rule [Ref. 6] and Quality Assurance [Ref. 7] programs at decommissioning plants, which are documented in Sections 4.1 and 4.2, respectively.L Further, the working group reviewed the design considerations and systems at . water-basin type independent spent fuel storage installations (ISFSis) [Ref. 9], as documented in Section 4.3. Since a zirconium fire is considered as an event beyond the design bases of a nuclear power plant, the working group provided information in Section 4.4 that includes events that are within the design bases of nuclear power plants. T 6

n 70 Deterministic Evaluation'of Spent Fuel Heatup. 2 The most severe accident postulated for SFPs is the complete draining of water from the pool. if , sufficient heat exists in the spent fuel, the clad will heat up, swell, and burst. The breach in the' clad could result in the release of radioactive gases present in the gap between the fuel and the clad, called "a gap release." If the Qo heat up, the temperature of the zirconium clad will reach the point of rapid oxid . IFnis reaction of zirconium and air is exothermic. The energy released from the reaction combined with the decay energy can cause the reaction to become self-sustaining and lead to the ignition of the zirconium, or a " zirconium fire." The increase in heat from the oxidation reaction could also raise the temperature in adjacent fuel assemblies and cause the propagation of the oxidation reaction.~ The fuel temperatures reached after the onset of rapid oxidation are sufficient to release fission products . trapped in the fuel. At lower tempera ly.nlad rupture and gap release may occur. At decommissioning plants, fire protection equipment may be the only water source available to provide a large flow rate for the mitigation of a zirconium fire. Therefore, the working group reviewed the phenomena of a zirconium fire in Sectbn 2.1 from a fire protection perspective. In - Section 2.2, the working group reviewed the charactorization of a spent fuel heatup from a fuel analysis perspective. Existing studies regarding zirconium oxidation are included with respect to the applicability to spent fuel pools at Joning plants. Calculations performed by the working group augment the insights rio Safety issue (GSI) 82, "Beyond Design Basis - Accidents in Spent Fuel Pools," studies to provide an updated, deterministic ostimate of generic shutdown decay times. Section 2.3 addresses the potential for criticality for a zirconium fire ~ ' event. J 2.1 _ Fire. Protection Evaluation of the Phenomena of a Zirconium Fire 9 Fire protection requirements for permanently shutdown plants (those covered under 10 CFR' 50.82), are specified in 10 CFR 50.48(f), which requires that the licensee maintain a fire protection program to address the potential for a fire-induced release of radioactive materials. Draft guidance on the fire protection program for permanently shutdown plants is provided by Draft Regulatory Guide 1069, " Fire Protection Program for Nuclear Power Plants During Decommissioning and Permanent Shutdown." The draft guide contains specific guidance on the level of fire protection to be provided for structures, systems and componenta (SSCs) that are necessary to provide protection of the spent fuel. 2.1.1 Comparison of Reported Zirconium MetalIgnition Temperatures For loss _of coolant events that are not the result of an extemal fire (e.g., seismic or human error), the working group conducted preliminary research on the combustion characteristics of zirconium and zirconium alloys. The working group reviewed ignition temperatures for zirconium j - and similar metals reported and found that the ignition temperature of combustible metals that j -. undergo heterogeneous combustion (such as zirconium) is highly dependent upon fuel j -geometry (e.g., particle size). The ignition temperature is defined as the surface temperature' just prior to rapid temperature runaway and is brought about by an exothermic oxidation reaction between the solid metal and its gaseous environment. Throughout this section, the point at which the cladding undergoes a rapid, self-sustaining oxidation in air, is referred to as the 7 i Ls.

= 1 '), : O7 . ignition temperature. For example', Mellor [Ref. 3] reports that bulk zirconium could not be . ignited at temperatures below 1300 *C.' The National Fire Protection Association (NFPA) Fire . Protection Handbook [Ref. 4] reports an ignition temperature of 1400 *C in oxygen, and Cooper ( [Ref. 5] reports that ignition tests lwith single 8-mm long sechons of unirradiated Zirc4 tubing does not ignite at temperatures below 1600 "C. The oxidation temperature of zirconium - reported by Sandia National Laboraic LMBofs.1 and 2)is 850 *C to 950 *C. This

oxidation temperature range is lower IDr tition temperatures for zirconium for aH other combustion sources that the working group reviewed - The working group believes that the :

oxidation temperature reported by SNL was the onset of oxidation, that would lead to ignition, but not the temperature at which rapid, runaway oxidation or ignition occurs as reported by the referenced combustion manuals. gu .to occur in air exists. However, the existing combustion literature indicates that 'y n p: onium cladding is considered probable when The potential for a self-igniting bulk z' u wr some combination of the following conditions are available: (1) an oxide-free surface exposed to ' a high oxygen concentration in a high pressure environment, (2) the presence of fine particles of zirconium, and (3) fuel assembly loading configuration and decay heat load. Some of the studies that were cited in this report were conducted in pure oxygen environments and produced high ignition temperatures. 2.1.2 ' Mechanisms, Causes, and COMkicIof a Zirconium Fire incidents involving zirconium fires and explosions [Ref. 7] have frequently involved materials of relatively high surface-to-volume ratios such as powders, machine tumings, grinding residues, and sponges. The particle size is typically below 10pm; however, the exact relationship on . article size, size distribution, temperature, pressure, humidity, and surface characteristics is not p well defined. - Furthermore, the senskivity of ignition temperature to particle size, mass, or - additions of impurities to zirconium, such as lead or manganese, increases its likelihood for auto-ignition. Markstein [Ref. 8] reported that ignition of metals is always preceded by slow oxidation, which ~ occurs on the metal surface or on and within a protective oxide layer. Once rapid oxidation begins, the self-sustaining reaction proceeds quickly. The ignition properties of metals have been linked to their low-temperature oxidation properties. However, ignition temperatures are not well defined for those metals that form a protective oxide layer, such as zirconium. . After ignition, the combustion process may continua to take place on the metal surface or on and within a usually molten oxide layer covering the metal. Altematively, the reaction may occur in the surrounding vapor, which is characterized by a high buming rate and the presence of a luminous reaction zone that extends some distance from the metal surface. Surface buming has been observed if the oxide is more volatile than the metal. Vapor-phase buming occurs only if the metalis more volatile than the oxide, but may be suppressed by the formation of a protective oxide layer. The heat of formation of zirconium-oxide reported by Sandia of 262 . Kcal/ mole is in agreement with the value of 1097 kJ/ mot reported by Glassman [Ref. 9]. The adiabatic combustion temperature reported by Glassman for stochiometric combustion of zirconium-oxide is 4300 K at 1 atmosphere (atm). + 8 =_

1 Markstein's research on the combustion of metals shows that metal ignition is strongly dependent on total gas pressure as well as oxygen partial pressure. Above critical oxygen concentrations, this research revealed that the manner in which zirconium wires bumed, while suspended vertically, depended largely on the wire diameter. In a related experiment, Markstein recorded zirconium rod ignition temperatures that were measured in oxygen at various pressures and found that exposure of r surface to oxygen, combined with the details of the rod's rupture characteristics pia ant role for ignition of zirconium. For zirconium metals, the propagation rate of the combustion zone is largely dependent upon oxygen percentage at 1 atmosphere, in a pure oxygen environment at 0.1 atmosphere, the

propagation rate was 1 cm/sec; at 1.0 atmosphere (such as in a SFP building) the propagation rate increased to 3 cm/sec. In addition, the rate of buming on the surface of molten-oxide metal mixture is limited by the diffusion of o n !?L gh the surrounding atmosphere.

The' zirconium cladding is continuously submerged under water and exposed to cyclical heating and cooling and radiation exposure before it is removed from the reactor vessel and stored in the SFP. The relationship between these varying exposure conditions and the combustion properties of zirconium, have not been addressed in any of the literature that could be located by the working group at this time. The NFPA handbook [Ref.14] reportMMIrconium bums with very iittie fiame and generates little smoke but attains high temperatures. Eyewitness accounts of actual fire experience with N-Reactor irradiated fuel cladding hulls at the DOE Hanford Reservation from 1967 through 1968 reported that buming cladding had the appearance of buming charcoal, with the cladding surface glowing red with an occasional small flame. Zirconium dusts, fines, or powders are pyrophoric and bum rapidly with an intense white glow [Ref.15]. At the request of industry representatives, on Thursday, May 13,1999, the working group . contacted two employees of Teledyne Wah Chang, a zirconium manufacturer, by telephone to gain more information on zirconium fires. As a result of this call, some references were sent to the working group that discussed zirconium cladding. The results of both the call and the references enabled the working group to determine that the information presented by Teledyne Wah Chang was consistent with the working group findings previously mentioned in this report. 2.1.3 Mitigating Actions At decommissioning plants, fire protection equipment may be the only water source available to provide a large flow rate for the mitigation of a zirconium fire. The working group has reviewed available information published by SNL [Ref.1, 2], Department of Energy (GOE) [Ref.11], NFPA [Ref. 4,14), and the United Kingdom Atomic Energy Authority (UKAEA) [Ref.13]. Assuming that an initiating event causes the draindown of the spent fuel pool, and following the event, the fire protection water supply is operable, SNL recommends the use of fire hoses to provide an emergency water spray to maintain fuel cooling on a temporary basis until the pool leak can be repaired. SNL notes that the required spray rates would be approximately 100 gallons per . minute (gpm) and that the exposure to an individual located 50 feet from the pool would be approximately 200 Rem per hour. The working group does not believe the mitigative actions proposed by SNL are the optimum actions to take because of personnel exposure. DOE 9 i l l 1

m

[Ref.11] states that fires exposing massive pieces'of zirconium can be extinguished with water based on limited tests conducted by industrial Risk Insurers. If a large water supply.is available, the working group would recommend consideration of connecting a portable monitor nozzle, which is a common piece of fire-fighting equipment, that plant personnel can set up at a much greater distance from the spent fuel pool than a manual hose line,'which does not require

] continuous manning, thus reducing tt t la personnel radiation exposure. A portable 1 monitor nozzle can discharge 'over SC Se ndent on the water supply) with an effective range of more than 100 feet. ' Many licensed nuclear power plants have portable monitor nozzles available either on site or from the local municipal fire department. For fires involving zirconium powder or fines, all investigators [Ref.12 and 13] suggest'using Class D extinguishing agents such as sand and graphite or inert gases, such as argon. If provided in sufficient quantities, delivoqi tqt'ge.hase of the fire, this method of extinguishment may also be effective on bulk zirconit nr. M As an altomative to the use of unborated water, the working group considered the use of high expansion foam. A high expansion foam generator can also be operated with minimum

manning once placed in operation. One significant advantage of high expansion foam over plain water is the low water requirement. High expansion foam has an expansion volume (foam-to-

._wster) ratio in the range of 200:1 to 1 1.Another advantage is that high expansion foam will be lost relatively slowly through the le i ifFs;ent fuel pool. This method may completely - submerge the fuel assemblies in foam, extinguishing a potential zirconium fire and limiting the potential for airbome release. Many nuclear power plants have high expansion foam equipment available either on-site or from the local municipal fire department. The effectiveness of high expansion foam on bulk zirconium fires is not known at this time. 2.2 Spent Fuel Heatup Analyses. Spent fuel heatup analyses model the decay power and configuration of the fuel to characterize the thermal hydraulic phenomena that will occur in the spent fuel pool and the building following a postulated loss of water accident. This subsection reviews the existing studies on spent fuel ' heatup and zirconium oxidation, the temperature crwi,a used in the analyses, and how it applies to decommissioning plants. The working group also provides generic estimates involving deterministic spent fuel heatup calculations for two criteria that have been used for . previous plant-specific exemptions. 2.2.1 Spent Fuel Failure Criteria The working group determined that the most severe damage would be caused by rapid, runaway zirconium oxidation. This would lead to significant fission product release even after the gap activity has become insignificant. The onset of rapid oxidation may occur as low as 800 *C [Ref. 4]. Runaway oxidation can raise clad and fuel temperatures to approximately 2000 *C, which corresponds to the melting temperature of zirconium. The release of fission ' products trapped in the fuel can occur at fuel temperatures of approximately 1400-1500 *C, Runaway oxidation starting in a high powered channel could also propagate through radiative and convective heat transfer to lower power assemblies because of the large heat of reaction in zirconium oxidation. 10 i

Several different fuel failure criteria have been used in NRC-sponsored spent fuel pool accident [ (studes. Benjamin, est. al. [Ref.1] used the onset of runaway clad oxidation as the fuel failure criterion in NUREG/CR-064g. This criterion was criticized because clad rupture can occur at a much lower temperature causing a gap release. The consequences of gap release can be . significant if the radioactive iodine has not yet decayed to insignificant amounts. SHARP (Spent Fuel Heatup Analytical Response Pr r ) ations'[Ref. 2] used the onset of clad swelling as a criterion. The onset of clad swe to gap; release occurs at approximately 565 *C, 1 which corresponds to the temperature for 10 hour creep rupture [Ref. 3]. A' cladding temperature of 570 *C is used as a thermallimit under accident conditions for licensing of spent fuel dry storage' casks. The next temperature thresho!d that may be of concem in spent fuel pool accidents is the melting temperature of the aluminum I natituent of some types of the spent fuel storage rack's boron poison plates. Its at approximately 640 *C. No evidence was found that boron carbide will dissolve in the aluminum forming an eutectic mixture at a temperature below the molting point of aluminum, but an exhaustive literature search was not j performed Eutectic formation is important for in-vessel severe accidents. If it is possible for the molten material to leak from its stainless case, melting and relocation of the aluminum in the ~ boron carbide-aluminum composite may cause flow blockages that increase hydraulic j t resistance. No realistic evaluation ofg Lrelocation of aluminum or aluminum / boron 1 carbide eutectic has been performed.LT er1tial for criticality is discussed in Section 2.3 of ' this study. Another concem is the structural integrity of the fuel racks at high temperatures. Steel and N zirconium form an eutectic mixture at approximately g35 *C. Several eutectic mixtures known from severe accident research [Ref. 5] may be important in SFP accidents. Steel and boron ~ I carbide form a eutectic mixture at approximately 1150 *C. Also, the steel racks may not be able l to maintain ~structuralintegrity because of the sustained loads at high temperature. Loss of rack ' integrity may affect the propagation of a zirconium fire. if the gap radioactive inventory is significant, then the cladding temperature must be kept below i 565 *C, if the consequences of aluminum / boron carbide relocation and possible criticality are - acceptable, then 800 *C may be a reasonable, deterministic acceptance temperature if uncertainties are less than the margin to 800 *C, and the effects of the higher temperatures on - the material are modeled. Otherwise, the temperature must be lower than the aluminum or ~ eutectic melting point. - 2.2.21 Evaluation of Existing Spent Fuel Heatup Analyses in the ig80's, severe accidents in operating reactor spent fuel pools were evaluated to assess ' the significance of the results of some laboratory studies on the possibility of self-sustaining

zirconium oxidation and fire propagation between assemblies in an air-cooled environment, and also the increase in the use of high density spent fuel storage racks. This issue was identified as.GSI 82.. SNL and Brookhaven National Laboratory (BNL) used the SFUEL and SFUEL1W i

computer codes to calculate spent fuel heatup. While decommissioning plants were not ' addressed in the study, many of the insights gained from these studies are applicable to decommissioning plants.' 11 4-

b' More recently, BNL developed a new computer code, SHARP, that was intended to provide a 1 simpisfied analysis method to model plant-specific spent fuel configurations for spent fuel heatup ~ . calculations at decommissioning plants. Some of this work was built on the assumption used by SNL and BNL studies in support of GSI 82. --2.2.2.1 SFUEL er es ased Analyse Si B Extensive work on the phenomena of zirconium oxidation in air for a spent fuel pool configuration was performed by SNL and BNL in support of GSI 82. SNL investigated the bestup of spent fuel, the potential for self-sustaining zirconium oxidation, and the propagation to L adjacent assemblies [Ref.1,6).' SNL used SFUEL and SFUEL1W computer codes to analyze the thermal-hydraulic phenomena, assuming complete drainage of the SFP water. In NUREG/CR-4982 [Ref. 4), BNL exter dp t,Sql studies on the phenomenology of zirconium- - air oxidation and its propagation in spsh Es mblies. The SFUEL series of codes include all modes of heat transfer, including radiation. However, radiation heat transfer may have been undomstimated due to the assumed fuel bundle arrangement. In NUREG/CR-0649 [Ref.1), SNL concluded that decay heat and configuration are important parameters. SNL found that key configuration variables are the baseplate hole size, downcomer width, and the availability of open spacp airJ)ow. They also found that building ventilation is an important configuration variable. I The draft SNL report [Ref. 6) investigated the potential for oxidation propagation to adjacent assemblies.' If decay heat is sufficient to raise the clad temperature to within a few hundred degrees of oxidation, then the radiative heat from an adjacent assembly tnat did oxidize could raise _ its temperature to the oxidation level.- The report also discusses small-scale experiments involving clad temperatures greater than 1000 *C, SNL hypothesized that molten zirconium material would slump or relocate towards the bottom of the racks and consequently would not be involved in the oxidation reaction.- NUREG/CR-4982 did'not allow oxidation to occur at temperatures higher than 2100 *C to account for the zirconium melting and relocation.- .Otherwise, temperatures reached as high as 3500 *C. It was felt that not cutting off the' oxidation overstated the propagation of a zirconium fire because of the fourth power temperature dependence of the radiation heat flux. The SFUEL series of codes did not model melting and relocation of materials. In NUREG/CR-4982, BNL reviewed the SFUEL code and compared it to the SNL small-scale ' experiments and concluded that SFUEL was a valuable tool for assessing the likelihood of self-sustaining clad oxidation for a variety of spent fuel configurations in a drained pool. SNL i reported the following critical decay times in NUREG/CR-0649 based on having no runaway oxidation. 700 days' PWR,' 8 kW/MTU decay power per assembly, high density rack,10.25" pitch, 5" orifice,1 inch from storage wall 280 days PWR, same as above but for 1 foot from storage wall a. 12

180 days BWR,14 kWMTU decay power per assembly, cylindrical baskets, 8.5" pitch,1.5" onfice unknown BWR, high density rack, SFUEL1W code was limited to computation of BWR low density racks. ~ High density racks with a 5-inch orificMMst representative of current storage practices. l A critical decay time for high density BWR racks was not provided due to code limitations. Low - density and cylindrical storage rack configurations are no longer representative of spent fuel 1 storage. All currently operating and recently shutdown plants have some high density racks in { their spent fuel pools. For an assembly in a high density PWR rack, with a 5-inch orifice, a j decay power below 6 kWMTU did not result in zirconium oxidation. All of these estimates were based on perfect ventilation (i.e., unli ' ant-temperature air) and bumup rates of 33 GWDMTU. Currently, some itted to bum up to 62 GWDMTU and some BWRs up to 60 GWDMTU. For fuel bumup of 60 GWDMTU, the working group estimates the decay time for a bundle to reach 6 kWMTU will increase from 2 yeam to approximately 3 years. Therefore, the working group expects the difference between critical decay times for PWRs and BWRs to decrease and that the decay time would now be longer than the SNL estirrte for high density PWR racks. The SNL calculations also do not appear to have included grid spacer loss coefficients which can have a signific ff Laisce the resistance of the grid spacers is greater than the resistance of a 5 inch orifice. I r1o mixing between the rising air leaving the fuel racks and the relatively cooler air moving down into the pool. Including the grid spacer resistance and accounting for mixing will result in the critical decay power to be less than 6 kWMTU The SNL calculations may have' understated the effective radiation heat transfer heat sink due to the assumed fuel geometry in the calculations. A more realistic fuel I configuration pattern in the SFP would give a better estimate of the radiation heat sink and raise the critical decay power needed for significant oxidation. VAP,e 16 studies 1rrsupport of GSF82 provide useful insights to air-cooled spent fuel assemblies, it is the opinion of the working group that they do not provide an adequate basis for exemptions. The studies were not meant to establish exemption criteria and lack sufficient information for all the parameters that could affect the decay time. Additionally, the reports are based on bumup values at that time. Since bumup values have increased, the results may not. be directly applicable to today's spent fuel. The preliminary results and the characteristics described in the studies may assist in assessing issues for decommissioning plants. However, the calculated decay time values do not represent current plant operational and storage practices. 2.2.2.2 SHARP Based Analyses in NUREG/CR-6451 (Ref. 7], BNL investigated spent fuel heatup that could lead to a zirconium fire at permanently shutdown plants. BNL developed a new computer code, SHARP, to calculate critical decay times to preclude zirconium oxidation for spent fuel. The code was intended to study thermal hydraulic characteristics and to calculate spent fel heatup to temperatures up to approximately 600 *C, SHARP is limited to low tempeutures since it lacks models for radiation heat transfer, zirconium oxidation, and materials melting and relocating. 13 i l 1

SHARP also lacks modeling for grid spacer losses and neglects mixing between the rising hot air and the falling cooler air in the SFP. BNL reported the following generic critical decay times using the SHARP code. 17 months PWR, high density rack,60 GWD/MTU bumup; 10.4" pitch; 5" orifice 7 months BWR, high de r GWD/MTU bumup; 6.25" pitch; 4" orifice The decay times are based on a maximum cladding temperature of 565 *C. The parameters listed with the critical decay times are generally representative of operating practices. Current fuel bumups, however, have increased to values higher than those used by BNL and perfect ventilation was assumed, which could lead to an underestimation of the critical decay times. The SHARP code was not significantKtppgh na ked, validated or verifiexf by the staff. The critical decay times are much shortertiidlitNo se calculated in NUREG/CR-0649 and NUREG/CR-4982, particularly when the lower cladding temperature used for fuel failure and the higher decay heats used in the earlier analyses are taken into account. This appears to be driven in part by the fact that the decay heat at a given bumup in the SHARP calculations is significantly lower than what is used in the SFUEL calculations. The working group has identified several areas that require code modifications, which will increase the calculated critical decay times. The staff has determin%f tt'ergde will be used as a scoping tool by the staff. It is not adequate for use as a technical b Sf licensees without further code modifications and verification. NUREG/CR-6541 was intended to be an assessment to steer rulemaking activities. The report was neither intended nor was it structured to provide a basis for - exemptions for decommissioning plants. The working group does not rely on this study for heatup analyses information due to the code that the decay time conclusions were based on. 2.2.3 Heatup Calculation Uncertainties and Sensitivities The phenomenology needed to model spent fuel heatup is dependent on the chosen cladding temperature success criteria and the assumed accident scenario. Many assumptions and modeling deficiencies exist in the current calculations. The working group reviewed the models to assess the impact of the modeling assumptions. Some of these uncertainties for the SFUEL series codes are discussed further in NUREG/CR-4982. For cases of flow mixing, decay heat, bundle flow resistance, and other severe accident phenomena, additional information is provided below. Calculations performed to date assume that the building, fuel, and rack geometry remain intact. This would not be a valid assumption if a seismic event or a cask drop damaged some of the fuel racks or the building. Rack integrity may not be a good assumption after the onset of significant zirconium oxidation due to fuel failure criteria issues discussed in Section 2.2.1. The building may also be hot enough to ignite other materials. Assuming that the racks remain intact is the most optimistic assumption that can be made about the rack geometry. Any damage to the racks will reduce the ability to cool the fuel. All three SFUEL, SFUEL1W, and SHARP codes assumed perfect ventilation or chimney effect conditions, that is an unlimited amount of fresh, ambient-temperature air is available. This assumption would be valid if the building failed early in the event or if large portions of the walls ' and ceilings were open. If the building does not fail, the spent fuel building ventilation flow rate 14

i would dictate the air flow available. Mixing between the rising hot air and the descending cooler air was not included in the codes . The spent fuel building ventilation flow rate is important in determining the overall building energy balance. - Air flow through the building is an important heat removal mechanism. Most of the air would recirculate in the buildin(a hLac drawn under the racks would be higher than ambient temperature and, therefore, le6 7ernoval would occur. Airflow also provides a source.of oxygen for zirconium oxidation. Sensitivity studies have shown that heatup rates increase with decreasing ventilation flow, but that very low ventilation rates limit the rate of - oxidation. Other oxidation reactions (fires) that occur in the building will also deplete available oxygen in the building. GSI 82 studies concluded that perfect ventilation was worse than no ventilation because the oxidation reaction became oxygen starved with no ventilation. Intermediate ventilation rate results a4 tjocu mented and could be worse than the perfect ventilstion case. .>. M A key fuel heat removal mechanism is buoyancy-driven natural circulation. The calculated air flow and peak temperatures are very sensitive to flow resistances in the storage racks, fuel bundles and the downcomer. The downcomer flow resistance is determined by the spacing between the fuel racks and the wall of the SFP. The storage rack resistance is determined by the orifice size at the bottom entra f ndle. Smaller inlet orifices have higher flow resistance. As shown by SFUEL a - culations, changes in the rack-wall spacing j and the orifice size over the range of designs can shift critical decay times by more than a year. ') The fuel bundle flow resistance is determined by the rod spacing, the grid spacers, intermediate flow mixers and the upper and lower tie plates. SFUEL and SHARP calculations apparently have neglected the losses from the grid spacers, intermediate flow mixers and the tie plates. These flow resistances will be higher than those from the rack inlet orifice. Therefore, inclusion i cf N M-2k-nid Tiw m-- ice may extend the entical decay time by a year or more. 1 NUREG/CR-4982 concluded that the largest source of uncertainty was due to the natural f circulation flow rates. The extent of flow mixing will determine the air temperatures at the downcomer and bundle inlet. The inlet air temperature and mass flow rates are important in determining the peak cladding temperature. The SFUEL and SHARP calculations assume a well-mixed building air space. 4 '_The downcomer inlet temperature is set equal to the building temperature. This assumption neglects the mixing that occurs between the hot air rising from the bundles and the cooler air descending down the spent fuel pool wall. Computational fluid dynamics calculations performed by the NRC's Office of Research (RES) using the FLUENT code indicate that perfect mixing is not a good assumption. The mixing that occurs between the cool air flowing down into the pool and the hot air flowing up out of the fuel bundles can significantly increase peak cladding temperatures. Even using different turbulent mixing models can affect the peak temperatures by approximately 100 *C. The calculations indicate that fully 3-dimensional calculations may be i needed to accurately predict the mixing because unrealistic flow topologies in 2-dimensional l approximations may overstate the mixing. i Radiation heat transfer is important in zirconium oxidation calculations. Radiation heat transfer can affect both the onset of a zirconium fire and the propagation of a fire. Both the SFP loading pattem and the geometry of the fuel racks can strongly affect the radiation heat transfer between adjacent bundles. Simple gray body calculations show that at clad temperatures of 800 *C, a 15

temperature difference of 100 *C between adjacent bundles would cause the radiation heat flux to exceed the critical decay power of 6 kW/MTU. Therefore, the temperature difference that could be maintained by adjacent bundles is highly constrained by the low decay heat levels. SFUEL calculations performed by SNL and BNL included radiation heat transfer, but the radiaten heat transfer was under predicted since the spent fuel placement is two-dimensional and the hottest elements are in the ngc tepool with cooler elements placed progressively toward the pool walls. Heat transfer het flotter and cooler assemblies has the potential to be significantly higher if the fuel bundles were intermixed in a realistic loading pattem. At temperatures below 800 *C, the' heat source is dominated by the decay heat. SNL and BNL found that, for high density PWR racks, that 6 kW/MTU was the critical decay heat level for a zirconium fire to occur in configurations resembling current fuel storage practices. At the fuel bumups used in the calculations, thisgi ~ cscpy heat level was reached after two years. However, reliable decay heat calcula$ol' het available to the staff. As a result, decay heat calculations from NUREG/CR-5625 wil! be used as the basis of a regulatory guide for calculating fuel assembly decay heat inputs for dry cask storage analyses. These decay heat calculations are consistent with the decay heat used in SFUEL calculations. Extrapolation of the decay heat calculations from NUREG/CR-5625 [Ref. 8] to current bumups indicate that approximately 3 years will be needed to reach a decay heat of 6 kW/MTU. . Several licensees have proposed usih MEnb Technical Position ASB 9-2 decay heat model for use in SFP heatup calculations. Using ASB 9-2 decay heat with a *k factor" of 0.1 . produces non-conservative decay heat values in the range of 1 to 4 years after shutdown. ASB 9-2 explicitly states that it is good for time less than 10,000,000 seconds (115 days). The basis of ASB 9-2 is the 1973 ANS draft decay heat standard. The standard gives k factors to use that are beyond 10,000,000 seconds. The working group has found that a k factor of 0.2 will produce cpnservative decay heat values compared to ORIGEN calculations for the range of 1 to 4 years after shutdown.- At temperatures below the onset of self-sustaining oxidation, the heat source is dominated by ' the decay heat of the fuel. When zirconium reaches temperatures where air oxidation is significant, the heat source is dominated by oxidation. The energy of the reaction is 262 kcal' per mole of zirconium. In air, the oxidation rate and the energy of the reaction is higher than zirconium-steam oxidation. Much less data exists for zirconium-air oxidation than for zirconium steam oxidation. A large amount of data exists for zirconium-steam oxidation because of the - large amount of research performed under the emergency core cooling system (ECCS) research program [Ref. 9).' If all of the zirconium in a full 17x17 PWR fuel bundle fully oxidizes in air over the period of an hour, the average power from the oxidation is 0.3 M$N. The critical decay heat as determined with SFUEL is approximately 2.7 kW for the bundle. The oxidation power source would amount to approximately 60 MWif the whole core was buming. A ' 20,000 cubic feet per minute (CFM) air flow rate is needed to support that reaction rate based on 100-percent oxygen utilization. The SFUEL oxidation rate was modeled using several { parabolic rate equations based on available data. SFUEL had limited verification against SNL experiments that studied the potential of zirconium fire propagationc BNL found that although they could not find a basis for rejecting the oxidation rate model used in the SFUEL, ] uncertainties in oxidation of zirconium in air could change the critical decay heat by up to j ~ 25-percent. It was found that the onset of runaway zirconium oxidation could occur at temperatures as low as 800 *C. Different alloys of zirconium had oxidation rates that vary by as L 16

r L 1: much as a factor of four. Apparently, it was found that oxidation in air was worse than oxidation . in pure oxygen. ' This suggests that the nitrogen concentration can have a significant impact on the oxidation rate. Since the relative concentration of oxygen and nibogen varies as oxygen is [ consumed, this causes additional uncertainty in the oxidation rate. The oxidation was cut off at 2100 *C in the BNL calculations in support of GSI 82. This was done to simulate zirconium clad l relocation when the melting point of rgegiqrrjms reached. If the oxidation was not cut off, temperatures could be as high as 35G0 UPit wa a determined that the propagation to adjacent t - bundles was over predicted if no cutoff temperature is used due to the fourth power dependence of temperature on the radiation heat fluxes. The combustion literature discussed in Section 2.1 also'shows that there is a large range in the temperature for z"conium ignition in air. Evidence cited from the literature states that bulk zirconium can not ignite at temperatuq an 1300-1600 *C. it is known from the extensive ECCS and severe accidentd p rograms that zirconium-steam runaway oxidation occurs,at temperatures below 1300 *C. Since oxidation in air occurs more rapidly than oxidation in steam, temperatures in this range are not credible for the onset of runaway oxidation in air. Correlations listed in Reference 10 give ignition temperatures for small zirconium samples in the range of runaway oxidation computed by the SFUEL series codes when the geometry factors calculated from zirconium cladding are input into the correlations. l Reference 11 appears to be the only pf t yat is applicable to zirconium oxidation in sustained heating of fuel rods, in the 4. A test, secGons of zirconium tubing were oxidized at temperatures of 700 *C,800 *C and 900 *C for 1 hour. The average oxidation rate tripled for each 100 *C increase in temperature. This is consistent with the change in oxidation rates predicted by the parabolic rate equations examined in NUREG/CR 4982. The zirconium combustion literature reviewed for ignition temperature did not discount or provide attemate l oxidation rates that should be used in the SFUEL calculations. l l As documented earlier, current operating practices bum fuel to higher levels than used in the j-evaluations. The BNL and SNL studies in support of GSI 82 represented operating practices of l the 1980's with bumup levels around 33 GWD/MTU. In NUREG/CR-6451, BNL used bumup levels of 40 and 60 GWD/MTU for BWRs and PWRs, respectively. While these values are closer to current operating practices, they still underestimate bumup values. The decay heat used in the SHARP analyses is significantly lower than that used in the SFUEL analyses. Given that bumup is an important parameter for determining the critical decay time, this is a significant. change. The increase in bumup level will increase the critical decay time needed to ensure that ' air cooling is sufficient to maintain the zirconium cladding below the oxidation temperature. The BNL and SNL studies in support of GSI 82 represented storage practices of the 1980's when plants were starting to convert to high density storage racks.' The studies did not address high density BWR racks, and the high density PWR racks in the reports were not as dense as the designs used by many plants today. The higher density racking currently used will decrease l' the air flow available for heat removal. Therefore, lower decay heat values are needed to ensure that air cooling is sufficient to maintain the zirconium clad below the oxidation temperature. I 17- 't a

2.2N. Critical Decay Times to Reach Sufficient Air Cooling Critical decay time is defined as the length of time after shutdown when the most recently ~ discharged fuel temperature will not exceed the chosen fuel failure criteria when cooled by air _ only. Earlier calculations using the SFUEL have determined a critical specific decay heat of ~ .6 kWMTU is needed for the onset of 3Mronium oxidation. This 6 kWMTU estimate

calculated using SFUEL in a high de 500 rack configuration is reasonable and is based on the best calculations to date. L However, this estimate is based on perfect ventilation conditions in the building.

For high bumup PWR and BWR fuel, the working group estimates that it will take approximately 3 years te each the critical decay heat level cited in NUREG/CR-4982.- Better modeling of flow mixing and accounting for the grid s apclia. plate flow resistance could reduce the critical Me Esp time beyond 3 years, but this may be decay power level and increase the ' counterbalanced by increased radiation heat transfer from realistic fuel bundle loading. Other assumptions such as imperfect ventilation could extend the critical decay time for the onset of a zirconium fire by 1 to 2 years. Accounting for imperfect ventilation, the working group estimates that generically it will take approximately 3 to 5 years to reach the critical decay time for current plant conditions._ A plant-specific calculation using actual plant decay heat and spent fuel configuration should yield shorter tim < 3.RA FT Previous EP exemptions using this analysis method for critical decay times were based on plant-specific calculations. The working group's generic results do not affect the previous plant-specific exemption determinations. 2.2.5 - Estimated Heatup Time of Uncovered Spent Fuel The staff recognizes that the decay time necessary to ensure that air cooling was adequate to remain below the temperature of self-sustaining zirconium oxidation is a conservative criteria for the reduction in emergency preparedness. Using the fact that the decay heat of the fuelis . reducing with time, creds may be given, if quantified, for the increasing length of time for the accident to progress _afterall water is lost from the SFP. The working group sought to quantify the decay time since final t,hutdown such that the heatup time of the fuel after uncovery was adequate for effective protective measures without preplanning. The fuel heatup time was used in two previous plant-specific exemptions. While the staff has not determined that this criterion is appropriate for generic use, it does - provide a measure of time that may be bounding for all heatup scenarios, including scenarios such as having a flat material fall on top of an assembly. The heatup time calculated by this method assists the staff in determining the amount of time available for onsite mitigative actions and, if necessary, offsite protective measures.. The staff would need to work with the Federal Emergency Management Agency to evaluate whether a generic time for all plants (or a group of plants)'or if a site-specific time should be determined for each plant. The heatup time of the fuel depends on the amount of decay heat in the fuel and the amount of heat removal available for the fuel. The amount of decay heat is highly dependent on the bumup._ The amount of heat removalis dependent on several variables that may be difficult to confirm on a plant-specific basis. As such, the working group assumed no heat removal was .18 J

available. The working group calculated the time required to heat a single fuel rod from 30 *C to 900 *C without considering any heat losses.at various times following permanent shutdown. This is very conservative because some cooling, through conduction, convection or radiation, would occur and increase the time to heatup. Figures 2.2-1 and 2.2-2 show the increase in fuel heatupilme that is required to heat a fuel pin adiabatically from 30 *C to 900 *C as the decay time following final shutdown increasg of the conservative nature of the calculation, ' more than the calculated time would be if the scenario were to occur. Additionally, a more-realistic or plant-specdc calculation should yield a shorter decay time. This type of calculation was used to support two previous plant-specific emergency preparedness exemptions. A 10-hour and a 14-hour period were determined to be sufficient time after the fuel is uncovered such that mitigative acuu.. dad offsite protective measures could be taken without preplanning. p 2.epemptions were based on plant-specific conditions, and therefore the workingh geheric results do not affect these previous exemption determinations. For the calculations, the working group used a decay heat'per assembly and divided it equally - among the pins. It assumed a 9X9 assembly for the PWRs and a 17x17 assembly for the BWRs. The design values used for the calculation are in Appendix 3. Decay heats were computed using an extrapolation of tt y er tables in NUREG/CR-5625 [Ref. 8]. The decay heat in NUREG/CR-5625 is ba EN calculations. The tables in NUREG/CR-5625 for the decay heat extends to bumups of 50 GWDMTU for PWRs, and 45 GWDMTU for BWRs. The working group recognizes that the decay heat is only valid for values up to the maximum values in the tables, but the functional dependence of the decay power with respect to bumup for values in the table indicate that extrapolation may provide a reasonable estimate of the decay heat for bumup values beyond the limits of the tables. The BWR decay heat was calculated using a specific power of 26.178 MWMTU. The PWR decay heat was calculated using a specific power of 37.482 MWMTU. Both the PWR and BWR decay heats were calculated for a burnup of 60 GWDMTU and include an uncertainty factor of 6 percent. 19 i .y i.

PWR Adiabatic Heatup 20 .18 Ei16 14 $12 ki0 ~ $8 $6 l 4 200 400 600 800 10001200140016001800 Shutdown Time (days) Figure 2.2-1 Adiabatic Heatup Time vs. Shutdown Time for a PWR BWR Adiabatic Heatup 22 $ 16 S 14 h 12 & 10 / Y8 / I 6 / 4 200 400 600 800 100012001400 16001800 Shutdown Time (days) Figure 2.2-2 Adiabatic Heatup Time vs. Shutdown Time for a BWR 20

2.3 Potential for Criticality 2.3.1 Evaluation cf the Potential for Criticality ) There are several scenarios in which the potential for criticality from spent fuel pool accidents exists: q 1 DRAFT 1. A compression of the stored assemblies could result in a more optimum geometry (closer spacing) ar.d thus creating the potential for criticdity. This could situation could exist for BWR pools which contain no soluble boron. For PWR pools with soluble boron, i this would not be a concern because of the additional reactivity holddown of the boron. However, for this scenario to occur, it must be assumed that the storage racks are i somehow removed from the I ow f4 the closer spacing of the fuel assemblies. r The working group does not i mis to be feasible. 2. If the stored assemblies are separated by neutron absorber plates (e.g., Boral or Boraflex), loss of these plates could result in a potential for criticality for BWR pools. For PWR pools, the soluble boron would be sufficient to maintain subcriticality. The absorber plates are generally enclosed by a stainless steel cover plate and the tolerances within the cover plad o and to prevent any appreciable fragmentation f and movement of the encloe1_qlr material. The totalloss of the cover plate is not l considered to be feasible. However, Boraflex has been found to degrade in spent fuel pools due to gamma radiation and exposure to the wet pool environment. Because of this, the NRC issued Generic Letter 96-04 to all holders of operating licenses, conceming Boraflex , degradation in spent fuel storage racks. Each addressee that uses Boraflex was requested to assess the capability of the Boraflex to maintain a 5% mubcriticality margin and to submit to the NRC proposed actions to monitor or confirm that this 5% margin'can be maintained for the lifetime of the storage racks. Many licens,ees have subsequently replaced the Boraflex racks in their pools or have reanalyzed the criticality aspects of their pools assuming no reactivity credit for Boraflex. Licensees that rely on Boraflex for reactivity holddown will continue to monitor for possible Boraflex degradation. 3. If sufficient fuel rod cladding damage occurred such that many of the fuel pellets became loose in the pool, the pellets might be rearranged into a configuration that could lead to a potential for criticality.' If the storage rack remained intact, that would tend to keep the pellets in separate regions and there would probably be no potential for criticality. However, if there is large-scale cladding damage as well as damage to structural parts of the pool, this could allow a large fraction of the pellets to reassemble into a minimum leakage configuration in one region causing a potential for criticality. Since any structural damage would cause the storage rack material as well as any absorber material to be admixed with the fuel pellets, the working group does not consider the assembly of a critical mass of fuel pellets alone to be a credible event. 21

p ~ Evaluation of the Potential for Criticality From Personnel Actions in Response to an 2.3.2 Accident 2 Without moderation, fuel at current enrichment limits (no greater than 5 wt%.U-235) cannot achieve criticality no matt 6r what the configuration. If it is assumed that the pool water is lost, a reflooding of the storage racks with u _wyter or fire-fghting foam may occur due to personnel actions.: However, both 15%R storage racks are designed such that they remain subcriticalif moderated by unborated water in the normal configuration. The~ phenomenon of a peak in reactivity due to low-density (optimum) moderation (fire-fighting foam) ~ is not of concem in spent fuel pools since the presence of relatively weak absorber material such as stainless steel plates or angle brackets is sufficient to preclude neutronic coupling between assemblies. Therefore, pefsonnel actions involving refilling a drained spent fuel pool would not create the potential for a 2.3.3 Potential for Criticality Summary i ' Most scenarios that could result in inadvertent criticality of stored spent fuci assemblies are not ~ . considered to be credible for the reasons given above. The only feasible scenario appears to be due to long-term Boraflex degradation. However, per NRC Generic Letter 96-04, Boraflex ' monitoring will continue for decommi 'cajqg ~plaats which use Bordlex in their storage racks, thereby avoiding any potential for crit Ti y(A lt should be mentioned that most of the fuel stored in spent fuel storage racks has been removed from reactor cores because it has lost sufficient reactivity to be useful in producing a self-sustaining chain reaction and in producing power. In addition, the reactivity of stored assemblies has continuously decreased even further during long-term storage due primarily to Pu-241 decay and Am-241 growth. Therefore, it would appear difficult to achieve c:iticality with stored spent fuel assemblies. I J i 22 1 L-

s 3.0 Risk Assessment of Spent Fuel Pool Risk at Decommissioning Plants in its probabilistic risk assessment (PRA) Policy Statement, the Commission stated that PRAs and associated analyses should be _used in regulatory matters, where practical within the - bounds of the state-of-the-art, to reduce unnecessary conservatism associated with current I regulatory requirements and staff pra ' s.12.a'pk-informed regulation the staff uses insights derived from PRAs in combination wii finistic system and engineering analyses to focus licensee and regulatory attention on issues commensurate with their importance to safety. Therefore, the level of risk associated with the spent fuel pools is one factor considered in decision making regarding exemptica requests for emergency preparedness, insurance, and safeguards requirements at decommissioning plants. Risk insights are used to complement the staff's traditional deterministic decision making process, and are not used in place of deterministic evaluations or_requirem< MtAFT The working group's preliminary deterministic evaluations (see Section 2) indicate that zirconium ) cladding fires could not be ruled out for spent fuel that had been transferred from reactors up to five years previously, based on a simplified, conservative analysis. To assess the risk in the time from shutdown to one year, the working group performed a simplified preliminary PRA, which modeled intemal and extemal initiating events to assess the potential risk associated with spent fuel pools at decommissioning i r plants. The preliminary results suggest that there may be non-negligible risk assa pent fuel pools at decommissioning plants i during the period analyzed in the risk assessment (one month after removal of last fuel from the reactor to one year after removal of the last fuel.) 4. The working group modeled a decommissioning plant's spent fuel pool cooling system based on the sled-mounted systems that are used at many current decommissioning plants. Information - 1 about existing decommissioning plants was gathe_ red by decommissioning project managers 1 and during recent visits to four sites covering all four major nuclear steam supply system I vendors (General Electric, Westinghouse, Babcock & Wilcox, and Combustion Engineering). In addition, the working group used several previously published deterministic and probabilistic evaluations of potential risks at both operating and decommissioning plants to assist in this study. PRAs are systematic methods of examining a complicated system (e.g. I ruclear power plant) to identify strengths and weaknesses in the design or estimate the level oi risk associated with operation of the plant or system.- The traditional event tree, fault tree method was used when developing the risk assessment. The working group chose one representative spent fuel pool configuration (See Figure 3.0-1) for the evaluation except for seismic events, where the PWR and BWR spent fuel pool designs (i.e., the difference'in location of the pools in PWRs and ? BWRs) were specifically considered. The working group looked at three cases as boundary conditions for its risk evaluation. Case 1 assumes that the hypothetical spent fuel pool and its support systems are available and have equipment (including instrumentation and power ) ' Event trees (ETs) are logic diagrams that analyze a sequence of events by modeling the outcome of success or failure of these events. Fault trees (FTs) are logic diagrams that model how a particular event can fail (e.g., a pump can fail to perform its function if it loses electric power, its wcter supply is cut off, the motor breaks, or it is turned off). 23 I J

y sources) available similar to that found by the working group in its visits to four decommissioning tf l li d to plants. In Case 1, transfer of the last fuel from the reactor to the spen ue poo s assume . have occurred one year previously. Case 2 is the same as Case 1, except that the transfer of the last fuel is assumed to have occurred one month previously. This is a bounding case in that it is not expected that a utility would disable all the support systems normally used to provide T spent fuelpool cooling within one mo a st fuel transfer. In fact,'it may not be possible to remove the heat load in the pool one month L for the sled-mounted systems current i after the last transfer. However, Case 2 does provide a bounding value to help determine the impact that higher decay heat loads (and therefore less recovery time) have on the frequency of fuel uncovery Case 2 is bounding in another way in that the frequency of fuel uncovery for Case 2 is highly dependent on the decaying heat load, while the analysis of Case 2 assumes that the head load stays constant (i.e., heat load at one month after transfer from the reactor) for the entire yar. DRAFT For Case 3, it is assumed that the spent fuel pool and its support systems are configured in a manner that is slightly better than the minimum allowed by current NRC regulations, and the working group believes that no prudent utility would so configure its spent fuel pool system. ' However, none of the assumptions in the " minimal state" are precluded by current NRC ' regulations. Case 3 assumes that the last fuel from the reactor was transferred one yes.r ago. The working group evaluated Case 3M IR ptarmine if there were a need for additional U MF i regulation in this area. Note that the time available for operator action for each of the three cases is based on working l group calculations. In addition to rapid draining events, the working group considered pool heatup after a loss of pool cooling followed by bulk pool boiling as a possible way to uncover the l fuel. It takes a very long time to uncover the fuel if inventory is lost in this manner due to the large amount of water in a spent fuel pool, the large specific heat of water, and the large latent heat of vaporization for water. Simple calculations for a Millstone 1-sized spent fuel pool and conservative decay heat assumptions show that it will take longer than one day to heat the water to the boiling temperature and take almost five hours to boil off every foot of water at six months after shutdown.' Therefore, it would take about five days to reach the point of fuel uncovery if the only mechanism to lose water is through heatup and boiloff. At one year after shutdown it would take longer than five days to uncover the fuel in this manner. Appendix 1 discusses the assumptions in Cases 1,2, and 3. Table 3.1-3 summarizes the preliminary calculations of frequency of fuel uncovery for each of these cases and for all initiators analyzed, in order to perform a risk-informed evaluation of spent fuel pool accidents-(beyond design bases), various references (INEL-96/0334; NUREG-1275, Vol.12; NUREG-1353; Draft for Comment NUREG " Postulated Accidents for Permanently Shutdown Reactors"; Draft " Risk Analysis for Spent Fuel Pool Cooling at Susquehanna Electric Power Station"; NUREG/CR-6451; NUREG/CR-4982; NUREG/CR-5281; and NUREG/CR-5176) were reviewed dealing with spent fuel pools and decommissioning plants to develop a list of potential initiators that could result in a loss of spent fuel pool cooling. Table 3.1-1 lists the intemal and extemal 24 l-H u

r 1 8 initating events found to be potentially important by qualitative screening processes in the above references. Each of the spent fuel pool accdont sequences can end up in one of two ways, known as end statas. The first end state is for sequences that do not result in significant offsite doses (i.e., the I fuelin the spent fuel pool does not re nature where runaway oxidation of the - zirconium clad will occur.) The secor e is where the fuel pool has been drained of water (as the result of loss of inventory and subsequent loss of cooling capability, or directly l from loss of spent fuel pool cooling capability) and the fuel has heated up to the point that there l is runaway zirconium cladding oxidation (i.e., ignition) that is called a " zirconium fire." From the l perspective of offsite consequences, the working group only concemed itself with the zirconium fire end state, because there has to be an energetic source (e.g., a large high temperature fire) l to transport the fission products offsit< r_to.have potentially significant offsite - i consequences. 'l The working group visited four decommissioning nuclear power plants to gather information on ' the as-operated, as-modified spent fuel pools, their cooling systems, and other support systems. The working group investigated the mitigation systems available as well as potential recovery actions by certified fuel handlers at the facilities. The working group also considered the l L availability of ' ;ms, instrumentation nrm== staffing, electric power sources, maintenance i personnel, fit =, protection systems, mOMIurces, redundancy, diversity, and ecuipment l location with respect to the spent fuel pool. ' Section 3.1 describes beyond design bases intemal event accident sequences that the working i group's preliminary estimates show lead to significant offsite releases from spent fuel pools (i.e., a sequence of equipment failures or operator errors that could lead to a zirconium cladding fire >j L and release of radionuclides offsite). Section 3.2 discusses beyond design bases external event } accdont sequences. Section 3.3 provides the working group's insights from this preliminary risk ,] evaluation. j 1 Appendix 1 lists the major assumptions made in the risk assessment. Appendix 2 details the event trees and fault trees used in the risk assessment. Appendix 4 discusses the integrity of ] spent fuel pools during tomadoes and high winds. Appendix 5 discusses the robustness of

spent fuel pool integrity given heavy load drops. Appendix 6 discusses the potential effect from and frequency of aircraft crashes into spent fuel pools.

The risk from sabotage is not normally evaluated in a PRA, in part because such acts are not i random events nor are they subject to easy analytical evaluation. The working group risk ~ analysts will identify to the NRC safeguards staff the structures, systems, and components that { i 8 internal initiating events are events that begin within the confines of the nuclear power plant and cause plant disruption. Two examples are inadvertent closure of the spent fuel pool cooling system suction valves and a pipe break in the spent fuel pool cooling system. External 1 . events are those events that begin outside the confines of the nuclear power plant. Two - i examples are seismic events and hurricanes. There are a few events that begin inside plants, such as intemal floods and fires, that have been characterized in some PRAs as external events. J 25 J I

1 are most important in helping assure that thel spent fuel pools do not represent an undue risk to - the public. 3.1 Beyond Design Basis Spent Fuel Pool Accident Scenarios (Intemal Events) ' The starting point for this risk assess egipalfsf developing event trees and fault trees to . model the initiating events and syste r w prient failures that lead to fuel uncovery (these trees are provided in Appendix 2). Working group estimates for initiating event frequencies considered potentially risk significant are given in Table 3.1-1. The table identifies the source of each frequency estimate (they are generic and not plant-specific). Working group estimates of preliminary conditional failure probabilities of mitigating systems and components (both active and passive) are given in Table 3.1-2. The table identifies the sources for the conditional probability estimates.. DRAFT The following are the dominant scenarios (i.e., a description of how the working group modeled - the sequences with the highest expected frequency of fuel uncovery) developed by the working group's preliminary risk assessment: Dominant Scenarios Loss of Offsite Power from PlantMa[d Grid Related Events The initiating event in these scenarios is the loss of offsite power caused by hardware failures or design deficiencies within the plant, problems with the offsite power grid, human errors, or localized weather-induced faults such as lightening. With power lost, there is no effective heat removal process for the spent fuel pool. If power were not restored quickly enough, the pool will heat up and boil off inventory until the fuel is uncovered (if there is inadequate makeup). In these scenarios, the diesel-powered fire pump is available for coolant makeup in case the ac power is . not recovered. If the diesel fails, and if offsite power is not recovered in a timely manner, offsite recovery using fire engines is a possibility. For the case of 1-year-old fuel (i.e., the last fuel was - removed from the reactor one year ago),127 hours is available for this recovery action. For 1 month-old fuel, 52 hours is available. Even given recovery of offsite power, the operator has to restart the fuel pool cooling pumps. Failure to do this or failure of the equipment to restart will necessitate other operator recovery actions. Again, considerable time is available. 4 Case 1: Frequency of fuel uncovery (FFU) = 1.3x10 per year (Case 1 is the as-found case with the last fuel transferred one year previously) In this case the dominant sequences (i.e., combinations of events that have to occur to lead to fuel uncovery) are as follows: i) (loss of offsite power) x (the diesel-driven fire pump fails to start) x (offsite power is not recovered before fuel uncovery) x (no recovery help from offsite sources such as fire engines) This sequence of events contributes to 60% of the FFU. 26

ii) (loss of offsite power with offsite power recovered, but the diesel-driven fire pump fails to start) x (operator fails to restart cooling system and electric fire pump) x (failure of recovery actions using offsite sources) = 29% iii) (loss of offsite power with power recovered and the diesel-driven fire pump is available) x (the cooling syste f is tgfeptart, either due to hardware failures or i operator errors) x (failure of re $ctibns such as use of fire pumps or offsite sources) = 11% Case 2: FFU = 4.2x104 per year (Case 2 is the as-found case with the last of the spent fuel transferred one month previously) The differences between Case 1 andE EreIdriven by less time being available for recovery of failed equipment for Case 2 where there is higher decay heat.

1) = 25%

ii) = 69% lii) = 6% DRAFT Case 3: FFU = 8.0x10-5 per year (Case 3 assumes the spent fuel pool and its support systems are configured in a much more minimal manner than found during the working group's visit to l decommissioning plant sites. It assumes that the last spent fual was transferred one year previously.) With no fire protection equipment available, and if offsite power is not recovered, the only recovery available is from offsite sources such as fire engines.

1) (loss of offsite power is recovered) x (the operator fails to restart the spent fuel pool cooling system); (failure of recovery help from offsits sources such as a fire engine)

= 95% ii) (loss of offsite oower and no recovery) x (failure of offsite recovery sources) = 5% Loss of Offsite Power from Severe Weather Events These scenarios are similar to the loss of offsite power events caused by plant-centered and grid related events. However, initiating event frequencies are smaller (0.007 per year versus 0.08 per year) while the probability of not recovering offsite power is also much larger (0.02 for 1-year fuel,0,1 for 1-month fuel, versus 0.001). Note that all initiating event frequencies and offsite power recovery probabilities are based on generic data, not plant-specific data. 27

Case 1: FFU = 14x104 per year I) (loss of offsite power occurs with no recovery) x (the diesel-driven fire pump fails to start or run) x (no recovery from offsite sources such as a fire truck) = 97% ii)(offsite poweris recovered)4 il r "Jhe spent fuel pool cooling system to re-start and run) x (operator fails to stM -diiven fire pump or electric fire pump for makeup) x (failure of offsite sources) = 2% Case 2: FFU = 9.4x104 per year Even with less time for recovery of failed equipment (because decay heat is much higher in Case 2 than Case 1), the two scenari I dominate. l) = 97% ii) = 2% 4 Case 3: FFU = 1.4x10 per year With no fire protection system to proMNk[p source for inventory makeup, the preliminary risk assessment shows scenario (ii) becomes more important.

1) (loss of offsite power with no recovery) x (no recovery from offsite sources) = 51%

ii) (loss of offsite power with recovery) x (operator fails to restart cooling system) x (no recovery from offsite sources) = 49% internal Fire The risk from this initiating event appears to be dominated by one scenario that involves a fire that is not suppressed in the building containing the spent fuel, and that is large enough to either fail the offsite power feeds, or the fuel pool cooling pumps. Spent fuel would become uncovered if the operator fails to refill the pool using the diesel-driven fire pump and if recovery using offsite sources (e.g., fire engines) fails. Note that because of the fire, recovery of offsite power or cooling pumps is assumed to not be likely and therefore was not modeled. Case 1: FFU = 8.6x104per year The initiating event frequency of 8.9x10 per year is based on generic data on pump-initiated j fires and electrical cabinet-initiated fires. Case 2: FFU = 1.0x104per year The initiating event frequency from electrical cabinet and pump initiated fires is similar to above. However, in this case, cutting and welding are assumed to be prevalent because the plant is assumed to be actively in the process of physically dismantling the support systems to the spent fuel pool and other plant systems or components. Using generic data, the initiating event 28

b L-frequency of 0.04 per year was used. Since this scenario assumes the last fuel was transferred from the reactor one month previously, it was assumed that the probability of fire suppression . was also greater because the building will more likely be occupied. The probability that a fire is not suppressed and that there is an effect on spent fuel pool cooling function is 0.05 for this cace, versus 0.1 for Case 1. Case 3: FFU = g.0x104 peryeERAFT This scenaio is similar to Case 1, except the probability of failure of fire suppression is doubled 1 - (no fire fighting equipment available) and recovery.of cooling or inventory is not possible, since j no fire pumps are available. The only recovery is through offsite sources (e.g.~, fire engines). i a of coonog DRAFT The initiating event frequency includes the loss of coolant system flow from the failure of pemps or valves (See Figure 3.0-1), from piping failures, from an ineffective heat sink (e.g., loss of heat exchangers), or from a local loss of power (e.g., electrical connections.) Operational data from NUREG-1275, Volume 12 shows that the frequency of loss of spent fuel pool cooling events in which a temperature increase of more than 20'F occurred can be estimated to be on the order of two to three events per 1000 reacti m data also showed that, for the majority of events, the duration of the loss of coc as than one hour. Only three events exceeded 24 hours, with the maximum duration being 32 hours. There were four events where the temperature increase exceeded 20 F, with the maximum increase being 50'F. For loss of cooling events, there is a lot of time for operator recovery. In the case of 1-year-old fuel,127 hours is available.' For 1-month-old fuel, the recovery time is 52 hours. The preliminary evaluation is that the risk of fuel uncovery is small. 4 Case 1: FFU = 1.5x10 per year l l} (loss of cooling) x (failure of control room a' arms) x (operator failure to notice pool steaming and level drop during walkdowns) >99% - li) (loss of cooling) x (failure to recover the regular fuel pool cooling system) x (failure of fuel pool makeup using fire pumps) <1% iii) (loss of cooling) x (alarms fail but operator notices signs of loss of fuel pool cooling during walkdown) x (failure of onsite and offsite recovery actions) <1% 29

Case 2: FFU = 1.7x10-7 peryear ' For this case the window for successful operator action becomes shorter because of increased ' decay heat loads, and therefore, recovery actions become relatively more important.. ' i) = 90% - DRAFT li) = 9%. lii) = 1% 4 ' Case 3: FFU = 1.7x10 per year. The control room alarms are assumeOMM less reliable. There is no makeup capability through the fire pumps because they are not required to be operablei L 1) = 91% li) = 7% DRAFT-iii) = 2% Loss of Coolant inventory. This initiator includes loss of coolant inventory from events such as those resulting from configuration control errors, siphoning, piping failures, and gate and seal failures. Operational . data provided in NUREG-1275, Volume 12 show that the frequency of loss of nventory events. in which a level decrease of more than one foot occurred can be estimated to be (on the order of) less than one event per 100 reactor years. Most of these events are as a result of operator error and are recoverable.' NUREG-1275 shows that, except for one event that lasted for 72 hours, there were no events that lasted more than 24 hours. Eight events resulted in a level decrease of between one and five feet, and another two events resulted in an inventory loss of between five and 10 feet. ' Using the information from NUREG-1275, it can be estimated that 6% of the loss of inventory events will be large enough and/or occur for a duration that is long enough so that isolation' of the loss is required if the only system available for makeup is the spent fuel pool makeup t. l system. For the other 94% of the cases, operation of the makeup pump is sufficient to prevent I fuel uncovery. L Case 1: FFU = 2.9x104 per year L I) (a "small" loss of inventory event) x (level not restored by makeup pumps or fire pumps due to operator error or hardware failure) x (offsite recovery, such as fire engines, unsuccessful) = 49% 30

ii) (a "large" loss of Inventory event) x (operator does not isolate loss) x (level not restored by fire pumps due to operator error or hardware failure) x (offsite recovery, such as fire engines, unsuccessful) = 23% iii) (a "small" loss if inventory event) x (failure of control room alarms) x (failure of the operator to notice condition ou r I wns) = 16% iv) (a "large" loss of inventory event) x (failure of control room alarms) x (failure of the operator to notice condition during walkdowns) = 11% Case 2: FFU = 6.0x10-5 per year This case is the same as Case 1 exct erator error events become more dominant due op$uat act to recover. The "large" break also to the shorter window within which op r becomes more important because less time is available for the operators to respond.

1) = 47%

ii) = 51% iii) = 0.8% DRAFT iv) = 0.5% Case 3: FFU = 1.3 x 104 per year 1

1) (A "small" loss of inventory) x (failure of control room alarms) x (the operator fails to notice condition during walkdowns) = 39%

ii) (a "small" loss of inventory) x (level not restored by makeup pump or fire pumps) x (offsite recovery fails) = 36% 1 lii) (a "large" loss of inventory) x (failure of the control room alarms) x (operator fails to notice condition during walkdowns) = 10% iv) (a "large" loss of inventory) x (foss is not isolated or makeup pump fails) x (offsite recovery fails) = 8% l Heavy Load Drops The working group investigated the frequency of dropping a heavy load in or near the spent fuel pool, and investigated potential damage to the pool from such a drop. Details of this evaluation can be found in Appendix 5. Based on discussions with structural engineers, the working group assumed that only spent fuel shipping casks had sufficient weight to catastrophically damage the poolif dropped. Other heavy loads are assumed to be treated using the guidelines presented in NUREG-0612, and are of low likelihood to be moved over the spent fuel pool. 31 )

For drop frequency, it is estimated that the, likelihood of a load drop into the spent fuel pool is in the range of 2.0x10 7 to 1.5x104 per year for a non-single failure proof load handling system. For a single failure proof load handling system (or a plant conforming to the NUREG-0612 guidelines), the range is reduced by about a factor of 10 (2.0x104 to 1.5x10 per year). Once 4 the load is dropped, the next question is whether the drop did significant damage to the spent

fuel pool.

' DRAFT For the failure of the pool wall, it is assumed that the load physically travels over the wall 2 percent (0.02) of the time (10 percent of the 5 to 25 percent of the total load path) and it is - assumed a conditional one-in-ten (0.1) chance of significant damage given load drop. The working group estimates a failure rate in the 2.0x10-7 to 3.0x104 per year for the non-single ' - failure proof system and 8.0x10* to 2.0x10-7 per year for the single failure proof system (or a plant conforming to the NUREG-0617 amirest (The value given in NUREG/CR-4982 for wall - failure was 3.7x10-7 'per year.) TheseMi m are bounded by other more likely initiating events. For failure of the pool floor, the working group assumes one-in-ten events (0.1) will result in4 significant damage. The working group estimated a failure rate range of 2.0x104 to 1.5x10 per year for the non-single failure proof system and 2.0x104 to 2.5x10 per year for the single failure 4 proof system (or a plant conforming ti L RFp-0612 guidelines). The upper ends of these estimates represent potentially impor kitdstors to spent fuel pool risk at decommissioning plants. 3.2 . Beyond Design Basis Spent Fuel Pool Accident Scenarios (Extemal Events) Seismic Events Spent fuel pool structures at nuclear power plants are constructed with thick reinforced concrete walls and slabs lined with thin stainless steel liners 1/8 to 1/4 inch thick. The walls vary from 4.5 to 5 feet in thickness and the pool floor slabs are around 4 feet thick. The overall pool dimensions are typically about 50 feet long by 40 feet wide and 55 to 60 feet high. In boiling water reactor (BWR) plants (BWR-3s, BWR-4s, and BWR-5s) the pool structures are located in ' the reactor building at an elevation several stories above the ground. in pressurized water reactor (PWR) plants and BWR-6s, the spent fuel pool structures are located outside the containment structure supported on the ground or partially embedded in the ground. The location and supporting arrangement of the pool structures determine their capacity to withstand -. seismic loads beyond their design basis. The dimensions of the pool structure are generally derived from radiation shielding considerations rather than structural needs. Spent fuel structures at operating nuclear power plants are inherently rugged in terms of being able to withstand loads substantially beyond those for which they were designed. Consequently, they have significant seismic capacity. In its preliminary risk analysis, the working group assumed that the spent fuel pools are robust for seismic events less than three times the safe shutdown earthquake (SSE). it is assumed that the 32

I ' high confidence, low probability of failure (HCLPF)* value for the spent fuel pool integrity is 3 x SSE.' For the majonty of sites and plants,3 x SSE is in the peak ground acceleration range of 0.4 to 0.5 g's (where g = the acceleration of gravity or 980 cm per sec.) Seismic hazard curves 2 from the latest Lawrence Livermore National Laboratory Study (NUREG-1488) show that, for most plants, the mean frequency for seismic accelerations of 0.4g to 0.5g is on the order of or g i ~ less than 2x104peryear. These hazaQt' are consistent with those from the Electric l. Power Research Institute (EPRI) studi as TE -4726 and EPRI NP-6395-D). o Using the definition of HCLPF, the working group applied a mathematical shortcut

• to get the l

. frequency of a seismic event that will challenge spent fuel pool integrity: j I 2x104 per year x 0.05 851x104 per year. l DRAFT Given such an earthquake, it is assumed that the spent fuel pool structure will fail, the pool will drain, there will be no recovery, and fuel uncovery frequency will equal 1x104 per year.: For the.- ] - cases (95% of the time) where the pool is intact, it is assumed that some part of the spent fuel L pool cooling components will fail and cooling will be lost. Recovery from offsite sources may be possible (although there would be considerable damage to the area's infrastructure at such earthquake accelerations). ' The preliminakA FT estimate for the contribution from this scenario is also 1E-6 per year. D The preliminary estimate for the total seismic contribution to the frequency of spent fuel uncovery for a plant with a spent fuel pool HCLPF three times its SSE and sited at most sites east of the 4 Rocky Mountains is about 2x10 per year. Tornadoes i The working group performed a preliminary risk evaluation of tomado threats to spent fuel pools j (details are in Appendix 4). It was assumed that very severe tomadoes (F4 to F5 tomadoes) 1would be required for significant damage to a PWR or BWR spent fuel pool. The working group 3 then looked at the frequency of such tomadoes occurring and the conditional probability that if p lsuch a tomado hit the site, it would seriously damage the spent fuel pool or its support systems. ,To do this the working group examined the frequency and intensity of tomadoes in each of the continental United States { The' working group determined, based on the buildings housing the l spent fuel pools and the thickness of the spent fuel pools themselves, that the probability of a tomado causing a catastrophic failure of the spent fuel pool is negligible and can be ignored. The l working group assumed that an F2 to F5 tomado wculd be required for possible significant i 8 A HCLPF is the peak acceleration value at which there is a 95% confidence that less L than 5% of the time the structure, system, or component (SSC) will fail. The HCLPF value is l. frequently used as part of a process of demonstrating that an SSC is sufficiently rugged for its D intended purpose.

• The working group independently confirmed that the short cut is accurate within about a factor of two, which is well within the uncertainty inherent in hazard curves for a site and the fragility estimates for the spent fuel pool.

33 1

damage to a spent fuel pool support system (power supply, heat exchanger or makeup water supply). The estimated frequency of fuel uncovery in the spent fuel pool from such tomadoes is in the range of 2x10 per year for Cases 1 and 3 and in the range of 1x10* per year for Case 2. 4 Overall, the likelihood of significant spent fuel pool damage from tomadoes is bounded by other more likely catastroph;c spent fuel pool failure and loss of cooling modes. DRAFT Aircraft i The working group analyzed the likelihood of aircraft crashing into a nuclear power plant site and l seriously damaging the spent fuel pool or its support systems (details are in Appendix 6). The l generic data provided in DOE-STD-3014-96, " Accident Analysis for Aircraft Crash into Hazardous Facilities," U.S. Department of Energy, October 1996, were used to assess the got9missioning spent fuel pool. Aircraft damage likelihood of an aircraft crash into or nt r can affect the structuralintegrity of the Fel pool or affect the availability of nearby support / systems, such as power supplies, heat exchangers and water makeup sources, and may also affect recovery actions. The working group's preliminary estimate of the frequency of significant pool damage is as 1 follows: ' The frequency of significant PWR speNelhl amage resulting from a direct hit is based on j l the point target area model for a (100 x 50) foot pool with an probability of 0.3 (large aircraft penetrating 6-ft of reinforced concrete) that the crash results in significant damage. If 1-of-2 aircraft are large and 1-of-2 crashes result in spent fuel uncovery, then the value range is 2.7x104 to 4.0.x104 per year. The frequency of a significant BWR spent fuel pool damage resulting from a direct hit is the same 4 as that for the PWR,2.7x104 to 4.0x104 per year. Mark-l and Mark-Il secondary containments generally do not appear to offer any significant structures to reduce the likelihood of penetration, i although a crash into one of four sides of n BWR secondary containment may have a reduced likelihood of penetration due to other structures being in the way of the aircraft. Mark-Ill I secondary containments may reduce the likelihood of penetration as the spent fuel pool may be considered to be protected on one side by additional structures. These frequencies of catastrophic spent fuel pool failure are bounded by other initiators. The working group finds the frequency of significant spent fuel pool support system unavailability caused by aircraft crashes to be as follows: The working group modeled loss of a support system (power supply, heat exchanger or makeup water supply) based on the DOE model. The crash consideration includes the area the wing sweeps as the plane skids and the skid length. The target area is 400 x 200 x 30-foot area with a conditional probability of 0.01 that one of these systems in the target area is hit. The frequency range is 7x10* to 1.0x104 per year. These values are significantly lower than other failure modes that can fail spent fuel pool support systems. The working group also modeled loss of a spent fuel pool cooling support system (power supply, heat exchanger or makeup water supply) based on the DOE model where the concem was about 34

hitting a 10 x 10 x 10-foot structure, in this case the results gave a frequency range of 7.3x10 to 4 1.1x104 per year with the wing and skid area. For the point target area model, the point estimate is 7.4x1048 to 1.1 x104 per year. These values are significantly lower than other potential support system failure modes. DRAFT DRAFT DRA.FT l 4 i 1 i 35 l 1. =

q;j: l g p9'~ " g'g g

• 1 n

M l v 1 9'D >qi 3 aN g" xv i TD 3U M P M g U /N t n K cP N o k P WTH "x u PP v f A U i OE wT LE rc EK t r M uA f i FM h H. l C o X b E f LT AN EA GE I v V kV O~D>Td l Fl too 3_ c 2 t -P p o gPP 'M oS M rG L u e E x ) r E P U r T n S v m ^mF a A I G ^ S E y kv w L t e 1T o s TA Ro e L E IC e UH A( ^N kV s) L c o 4 r Y o a t P E V r s c L C e p t n E D e O o d TG p F L E ) 5 E TP E M F I R D I E 5 Y l F P ( S esa w* b

Table 3.1-1 Loss of Spent Fuel Pool Cooling initiating Events INITIATING EplENI FREQUENCY foer voar) SOURCE OF FREQUENCY loss obGsite Power-Plant 0.08 Generic frequency obtained from I centered and grid related operational events at nuclear power events DRAFT plants (INEL-96/0334 and NUREG/CR-5496) Loss of Offsite Power-Events 7.0E-03 Generic frequency obtained from initiated by severe weather operationalevents at nuclear power plants (NUREG/CR-5496) Intemal Fire 9.0E-03 (1 yr after Generic frequencies obtained using ]QK_ the methodology provided in the EPRI t 0.04 Tmo aner slutdown) Fire-induced Vulnerability Evaluation L report (EPRI TR-100370s) Loss of Pool Cooling 3.0E-03 Generic frequency from operating experience (NUREG-1275 Vol.12) Loss of Coolant inventory 0.01 Generic frequency from operating DRAFT experience (NUREG-1275 Vol.12) Seismic Event 2.0E-05 Estimate for a 0.4 to 0.5 g earthquake, obtained from NUREG-1488 i Cask Drop 2.5E-06 Estimate based on industrial data (NUREG-0612 and other studies-see l Appendix 5 for more detail) Aircraft Impact 4.0E-08 Estimate based on DOE-STD-3014-96 and other studies (see Appendix 6 for more detail) Tomado Missile 5.6E-07 Estimd6 based on data from the National Climatic Data Center-see Appendix 4 for more details) 37

n n ,P in n P P i tn i P eE L E E P .l i H N F P P H H F t 2p G T-F F S E-qn r r E D g S S o o 3-. e 2-n L e Me; eE e f f v C S .e A M M b y y b XbY,o1 N P Pn P X X t t h o o i hi t E E4 Fo U E E b b Ed :_ t e t 3 p t r s E n3 no K H H d e a d Hlu. a u L-R t e b b e Xo E eO e n M X-X .l m3 m o m3 Th +,i.mre C F vW v o F T T ue p p ue r s E et e io T t 9 d P L L ss ss L s s : f L N n-n n S A A sa e e sa Ai E E a aC h h aC h + ee A oL o t C r r t t t n oe P t t t d e n C d N eF n n n e n e e r e4amT h s sT e4 e4 e4 R eI E v3 n r i e e3 e3 e3 e iw ic a v3 v3 v3 a i c e e3is H U ain aF sl w sl O E pb n W r, Pi E o O 0 O b b Pc n0 nW n/ pb E t t E l a ot e t S eR eMe o6 ot o6 na e-e- md r i L L 9 v E L-n r sA sun d d d ua mC mC ua 9 9 9 t t d u u pv e L-eHC a s s n1 nPo e L-e L-e L-pv a sT sT a sEoiL lu w sT w sE sE sE e n sL sL i n r w - i i t e sGp aN aN aN aN u s s sC t u aA aA u bI Hh A s r x y y eF eDe bI bI bI ir f t f l l a a n n A A. 1 2 2 k k 0 0 0 s 2 i is YT E E E 2 2 2 1 2 0 T L 8 0 0 0 0 0 0 0 R R I E I 1 6 1 E E E E E 0 A 0 P F NB A 0 6 0 0 / 8 / F F EA 3, 3, 3, 8 N 2 1 N 1 1 S S VB 2, - 2, 2, 1 2 3 1 2 E O e e 1 2 3 1 1 1 h h R s s s t P e e e e e e e e e e e s s s s s s s s s s s t n n a a a a a a a a a a a i i C C C C C C C C C C C d d e e s s o U U ) n n }. ) ~ el t ) n s s ce s i r s i . r or p ise msl h n v p e s s o p le2u e e c mnb odi m l l r b n u ia ehefs5 u El uha i c c t vl s p o p l n n c p e e le f a a7a0 L-o )o e2 v r 0ia i i u u e r n vC 2bf e e Cpt ir mTr so ir 1 a f q q e i o a a f noi f f oLor f f f i si e f r f e oo s t h oar e g2 t of A e e r chl n a r ei bh er endp esh t r r o n soPn s s2 F F e sd0as soa uiEae u e5 / / n h un2libCw ( c t s s u aHinh ( t (i 1 a o t xy e e e gt oal i ol nhngo gel r yf vk rd gc n e d t i a uf io h xiet r s t g i h nt n sr gb ot nbs oia l i Me ssb e li N a is e r eo u osm b b O - ehon oht i cii t n a a a o b b T a u u t in c gt e ox r gc2 I e r en gv er x5mT r F en t t h lenb o o P s d t r e yi a obto a ep t h PST moie moao e a o nl t mTahp r I P P R n r a n r. e cdb b tt C e cst n e iF l iaoh t t al l rd t S nPvoil a e n n n a r laFriPemdt / y aFaow oA' E e e D s f h wa t r h S e x e 2. oN hSsp v v s hosa0 b "l k a iso 3 t s t o E E i eR n p h

• l T pb f ba eluh l

r t ae& b ba s p or s amimbl v t am saie aO 2-2 ~ l t t t 1 f - i@ ia n sr t r e sis s t er7n o 'n sr p e omd1 a e o sov ef t 1 1 p uc o"n r f e h sf le n o os f a a snso 3 3 d m ee e t f a ls o L =a t f l t f = a e u kt so a3awb lsoC e e r p ao ia sonrdl f e mp l smd Co is = t a si o is l n l k b b w i t n o a ss8 ( f e o f on=e rl le. a a o oe rl ooa ol u r f iob o r wl s o l T T le t u e r nr eiat f r 1 l ab2 p c r al gb o a ndw e r nla aoia t t t n ei a os n a e r r k s c l s eet smde eel er eeis n c a pv ainm pvame pv v a e le ie ai OiogcC[ ait g i Oiwid OgaC T D E Fd t E F' E M F T T P A F F F E L L N T L-L- L P P F T M M C C C N U U P T T T E P-P U L L L V K A A A E G L D' E' M P P P P P P E E E F F F H H H l)I t l

f t o h r r E n e i i t a a X e4 r - ri p p n-v3 mo-Low t t e e iE e3 it cL Oatr n r) r) dH o/0 saE Oee ole o e eXas6 sfN Cpt e t t W d b - r eeI s a2 e vPo la9 hh P-b ea ia3 a2 eo ir r r r t oi i l s r o r d sny rl i oFf eL oTi Es og pE ui4 p p-r t u a6 p - t a - .w He clo v eE e2 d vE eE ee pS d e E l f d ia aE nS N plst l s o r r a a2 1 r2 yn a t r t 3 5 o4 o e f f n nr f e o o ge,nn cvt g h m3 .I ee ey aoid f e2 t 5 t t 9 t i o ue m= o ss yi) e= e )= m= e= e= ov s f so loe soi xesio eh r r yi t t i R el sl r r t l 7 u) u7 n e 'c L t t eumt sa 2 d2 e * )4 u) u) dr r s* 2 )4 1 n o s e r i 4 d4 oot aC ba1 la5 a1 ba8 la2 a8 h f s aP s od h oim t f f e-ee c o gc5 e n ad* E f* f b b 0 o0 oo f f ad* f* t 4 w pE st h i c 4 r rai b b 01 o0 o0 o f e 3 o y X-asst h ae3 e o m3l Cr em 5oh3 ie m1 m1 t sl r 4 r / y/ y* yb pa1 a y1 y/ 1 t 0r e 1 1 pa1 T0 e - n/oeE shs t t t t f t e1 e1 e-l e

w. ila h1 l

f / e-f o6f sH ag a4 o 6 m (p (p o b( b( b( b( 9t nX ie3 rEng9 ap ap m (p(p ap ap d - u s t a o-ehr3 o iL-pv u xx p bx bx u xx bx b* eL b pP mdn/c0 i cCniE s a s ee a t a i n s c oe oe s ee oe o* sE e aN PesF ani6 aOaN y r r r s r r u a==( p= p= a== p= p= l LrS Sa29 FLt I bI r f n A ks iR 4 3 2 4 2 1 3 3 2 0 0 0 0 0 0 0 0 0 P E E E E E E E E E F A 0 5 2 2 1 2 5 0 8 / S N 1 5 4 2 9 1 3 7 1 e h 3 1 2 3 1: 2 3 1 2 3 t e e e e e e e e e e n s s s s s s s s s s i a a a a a a d C C C C C C ':a a a a C C C e i s n . r r U i

e. o a t

r e s g yo gn f r i oe nnl t na (f s sh e hi e i h d c oiae lot o i dr sT i t os tf e u tgb seheo lo oe n cemm cl e bnl l nh s iu f r x a hr folatide4 m 9 u oe oss2 s sa ownh8 nr 3 i ead qe sTpa e svbt ot vla n s r o pss F s aia loal ac si igm 3 oh e a l r sot s e3 amdiC a i pleei& mo z / n a n eo& s nr r e elaapo ne k v h g1 ef eml et o sl 1 t r it va3r viul s gm&or igs b is c yo s. s t l i i t u r it e e sr e2 a i s t b mo1 e o e o ms oot a gnse r a ef t t aaC ir. o a s y imh e L e (e h h r l b t r cCa t s 0 s s t t o o o omnC t 4 ysps e )s r s n s a f T coin P ad na emr e hdi r ei g en wler uul nf l t in ol c sob F gnbe n Ma l a l e shli h oab e lo 3 k wd A 6a A o ala ah l v oc 2& la et 1 v o e i c v a E wml a r e e1 av M 'n t R ) / e r s t d r ' s ii e oue n s a s r 2 o a air vt veeapm.D e l t h s t s Cs r i 2 s r iclnl egt ety0 h 1 e n a iopnee sic r r r s r22 3 o' her oo so oh oa o a15 r e s iEp of t l t t e 0bo t - f sa ohebeC ls o -, l ls' f r 2a r 'b iat l r cL r ies :- s y l r i ke o f t aiC a 1i a ws-f r a u f d dolu hf f r af v o r nT T oe-b Ts o t r oe ah oiLn s .r o p c ;, Aa uo u t xx t s t at s s 0 at aeeo r - not n o r r rt neh1 ec e nPi i eioh r e einp e gv s pvx2e p vEebmtc4 pf i p f Oee5b OeHboia2 Ooias t E L E C C P O O O L L L L-L L O O O O O O C C C P P P E E E H H H

^ e' 6 e e E) 9 e e R c'g 3 L n' s g Vin C 4) S-O E S E 6 a-a. Nsl R 1 i f 4 O t a s 2e o S 0x t I 3i ae an P N( I I L u W8 d did u di AR N U u-t - yf c 7 Dn A K E b e-l u l MV M M e ot ah s tr2 Tp ns ns Xdn-N r - e i o1 Sp t t EI o p-L E' V-Ei s r s r eG EA u e u e v i i H-Hin H4 N e rE Oe d h d h E 1 XX X3 I Xda t t u rR De ind ind e no no e 3 P Puth. d P-t s s P0 vU n( FE a F/ E F b,ss iRN os o n o n SH C S 9 4e 6 H S a a e d d sX4 s hd di nP3 e n L-e n3t an ed e2 e2 t s t r t n a, su s1 s1 e3s d eF 3 m eE m v0s .n n at a6) a6) 5 s 5 b0 vS 0 vN i e6s-t l f a1 b b0 e h it eI r - x x v8 ee eGi eGi dt 5 .nn 5 n9le o. 9 e e d a5 n n-e oi e t t t d t ah aEn eEn t oLs S 3 o eL d)E m d) m e e t E9 mo mRe mRe i dE u e-u m m eNre 2 td hup hup s e3 N OR sn sNp sNp s sEI s sE s h h sI e i s a8 s s s a nh E(A E(A DT Ea y a7n A b( a E E bit b(i lan 3 A 0 60 k E is 2 2 1 2 1 1 3 2 0 E R 0 0 0 0 0 0 0 3 0 6 6 P E E E E E-E E e 4 0 0 3, E E F

0. '

0 0 0 0 0 0 O 3, 5 5 S 1' 5 1 5 1 3 2 2, 2, ' 2 2 2 1 1 e 1 h 1 2 3 1 2 3 1 3 s s t e e e e e e e e e e e n s s s s s s s s s s s a a a a a a a a a a a d C C C C C C C C C C C i es U_ l y a l m s t o t s n d k c t f laf s r a m e el o l o u e e r mspo u nsio laa r w geg d o gd au n c l eos np o lee c l n o md mf t oo gro h i n o v 0 bl yed 6 emn arl i bi w m3 iv ue dymd sa at o e 4 u 3 u o s& a. co g8cee t e l n or r r s i a l r oo ae q i t t f uis i sd r e dn n ef n r noe ame1 delk o1 t is w o ehw f w t d nLa s n e udrl t F acvnuhe-a.w pe e e a olo nI st s s s e ea eu u s h vm cdf /s smi. sea ~ t f sot s as t lor sf AvC o nh e k t e0 .br e min 2. in s wt aC need t o t f n c s l r el ua h y n g 1 *6 a f yo c o m aa vo dt n dn r or r ps et oki nle e r oia o1 r t t ina l t el na laba la in pe mi den s a a e t l s t t ce f b ec e5ot b e h a a ewe olaC os o mr e t e r o w d t, n wi n f v oan T l a av dupr inmP rP wm2ee a c vi r p a a5 e e - gf K l ae n c e kst oU v s s n e e yV t a er a l F ig nt cspo 1 r w e is . a alab t nM msb v aih gl n r r r r hla i io-o i A i t e l e7 b m la r , t a-at V n0a d v f at f y f nnd E a2o t, aN 2v R n oi oeee t - p e ecI e1 a e p1 o r vvm t e t p )r g uGo o2 ) r adP af s t ee u r D 2 la e s a e a okEr' tior lomE f e en eisss h t i d ee iH ins 3 isaRl s r is0 yi yh h s 1 r s t a mUe ea o pi oe5 o rPry rt n o Nv h C t r s u c eFe e oi t t t t n e oT t lse n* lsxly ls p p e s aml 5 s. r yFs (Sv ( a o ika* ia ee ia o it l l f t f na f yoc yr d b iaoo f a ah t l t ct n cg ya a nn1 es me f r oFio im r osi o neu ner o r h r r T r ddenbh ot l t ix a uae uivi c e gl e n el t o t 5 nb t t laa seoem1 a t r ala aeo n i r r o c emrmluiwe eloa eep e qmf qP c r r r u r sd e eFn polas nesn ov ph p p p r a o r t r Onar abaa Oca OTa O Fdt F S u t. e L M P E L S M M P P P R l P U U U A T-O L F R K K K A A D M M M -S R K V V V-E C S N N N I I R R A I P P P IA C PE E E E I E H H H H E I v rf

e ss e m 4 sed t i inn s n a oe Foo e u s-n nl dm l r ee e~ epda e oeau rl l s t hf vr us n imb v t s did ig i t ghn o r n o u a so i8 i e l ri a8 wtnte t. u at y ym t s ns g ga oa t iepoFe t e 4 f oh m1 mr o l n o=Rr sC e m )r e u e l l a m e r d h o p oG v t d dpho 2 2 n elmho ae f t n o o ot iE ~deTf e r ind h h ng l 1 1 nR nD - sy g s t son 6 6 o o oU o eesmrog t n o n e0 l e0mi i a d m7 m7 9 9 V V dN dt otepia ll e2 3 3 e u m 4 4 e e e d 45 5 5 5 nm zmayds5n o1 g0 g0 n c n a 3 - n 7 7 I o6) i0 i0od C 3R R 2 2 eo rhCmtsE 5 s1 s1 b0 s n bf t P r un a w 0C C 1 1 x u-1 or L eGi R R d og g 6G G G G ed el t o t d d e e n5px t t aEn eT eTinm r 9 a e amoFpoh LE E E E s mRe vI vI it i ea R R R R m m rR nRnlddo r up t t i e iUp s o r et hb dspto E U U U U sb so t y sNp eP ePo en o nf E(A DE DEcwo S N N N N N Eo E N a o (a n la I i nA k 2 3 3 2 5 7 s 0 0 0 0 0 0 i E E E E E E 4 2 R m 0 0 0 0 0 6 5 0 P 0 E E E 8 7 3 1 2 5 F E 0 0 m 3, 3, 3, 3, 3, 3, S 5 9 4 9 2, 2, 2, 2, 2, 2, 1 e 3 1 1 1 1 1 1 h i 1 2 3 s s s s s s t e e e e e e e e e n s s s s s s s s s a a a a a a a a a i d C C C C C C C C C C es ~ U m o y s lt r f y gc e e r r r c e e g o mr t w w n t o o i n ed n u l e n - p p o e od v u e o ca 1 q v e e c in e 4 l lel em d e e s o lo e io e o st r e o f f f f o o k s F r d o o p p a ud i f / t f f f f u r e s du o o o o q dm t e eh s s s s s h eu t b s e s s s r t s t s r eo o o o o a as l e e r a i l l l b mR r r r r e a m os o o o g ns f t f f f ei b n r e r )r )r 5 gt ) ) n o r a v a a a 0 oe P T ee e e e e o dv r I y yh y y t a e d n F et ea eI e t r rt r r 4 mis e ( 0 oh (p s pe p e w yA_ (p v A t t i ( a e y a-yr y E c ce r f y R nd n e c c nP o r n f oi t i r e e v e e 2 n uy u e u u )r )r g qd qs qn q e a e 1 t e n e e e e en y y yi 3 r r r r t f a f b f ~ f rPry r n r e e a nd nd n n p t t t t e eFe l pl ee ee e e ( (Sv p b ( p r v e vt v v z a yd et ein e e y yot T c e n e c ct n na ge gn g g n neu em mc mi n5 ms e e gl ua u uoe t l t e 'a sn q qmf t t qh st en e u n ea ea t e im h ee e e r n v r r e uv n n F Fdt o nl Fp r p e ie i I L 0 S 0 V I N M 4 I E P P L L-O R L L O O C DA I F P P O O M N I T-O O P P S R I N O O O O E O I L L L L S T-E E E E E E-E 'I I I I I I I

,s f o e e Pi o /w e g ir e e f s Le r s 3 ( g a h h t i f r n2 s t t p g o ee d Ca r l o3 6 6 S hs n Ol o f n s n en 3t a i i i ci f d0 9 9 n ei xn &eC a, Lo 3 e5 4 4 I 5 5 O ocl n ele ei 2,e n 2, o3 aR r f 8 s e o a L i sn cn r sin 1 3 8 bC R R i 1 sn 1t 1hi sn sin C C N so ed swd 1 s e yo e s sn e ss e s s s / W e s e e yG r / / pl s lie G G t r er er aa st i D be Ct an u s a e a pk e pl s C )s bR E E uip pi C a )m s K n Ceo a p aU R R L sl ust s e vs c im st ss n a nC b n U U W las n y. neAI n l oN 1 io a m i r N N r ulelat a pl ep i 1 dic i d da da pn r e e m e. O e n pk u py miu nla ela i f f C e d e o o s n e l L e e s a s ey a E s a e i b ylbn abye aq s r sri U ),R UR vg 9 9 C R a bk s l U. F .sC .C it o 1 1 C M t yl Se Uiula o u

u. d n i

eli emh o t y cis n . Sam al o o s t t s e v o B B ci ed a /o f e eo e edn s r d r r a dl t n t r it t t e g ie ag t t h a n of a a aaw m (w sh u u im im m eo e e a la e a c em m ampm d muud me. ma oo m imla( r r l s mul i l t s ,p3 iC I i nt f sPo& sO sRO sO o e ig ig i i a is bs iu sscu s os it t t t isci h y Ea~b Eaob E pl ELn1 EL ECL EL Cm F F S S s ss r e lanA 3 k 0 is E-2 2 0 2 1 2 0 0 2 2 1 R 0 0 0 0 0 0 E E 1 P E e E E E E 0 0 F 0 o 0 3, 0 0 0 1 1 y S 5 t 1 2, 2 1 2 1 1 2 3 1 2 2 2 1 1 = e h 1 2 3 0 0 0 0 s e e e e e e t e e e E E E E s s s s s s in s s s a a a 0 0 0 0 a a a a a a 1 2 C C C C CC d C C C 5 1 e s s U e r sn s b o r l w e a d e c e e e e e h o i t d c r i a la n P i f, ir ir r r k e f, f, e S f. t e u O g g g g e wl. 2 n w e e 4 l e e e c e gv q ia ( ( ( ( t e ine r e f r s sw s s n l v r l F o e eo e e la e uo t cs p s do / cs cl c s s e r r r s r r r t p uu m m n r e g uu uy u u oo oo or o o v o ew r it a. sh se sh sh i f vo i r i l s e0 e v e6 e0 ( ( el l i yp it i5 r r t t m sf it s b lt 5 s3 sf e e gd i si a n inn f f p f n f o w w f f b e u f o f ia o ou oa o o op s o o o s t h gs p p og %t g gs ga r f n T ne e e ci e ne f n P ui s ic it it f s s m i s sh c slo x ul x s s o le u e n iso F t f f a e o se f f r i yn o o st ui op gs yo e c ,n A r t t e o o los oi f f lo v nP e t E eF k lb ee eb y y co n e nl e vS R l r a v la r r e a vh p n nt ni e) e it 2 e e D ia iw a ia v v o e 1 irh s r v r v ot o nd t f r v r s 3 ao ea et ea ea cn c f ou vn v v ee e t lc v e oe o e o e r v r t s o h a ei ee ei ei o o ls n e at cm cv cm cm f e f ii r r r r d a s l t t b t m t ye y f o )- o )- o )- o )- n s s s s ia lt r a i u y r t t t t lit i o u t T i p ee ee ee ee bl b )s i t G l t a r r n a e at ra b r r n r n uin ui b b n r i e lui lui u lig ig od oe pd l ig ig t vc an an an an r ir r v n Po Fe Fe Fe Fe Pg Pe OI ir C 1 2 3 E E E E O T T T T L I I I I S S S S N F F F F C W W TV F F F F E O O O O P-S D P P K E V V V V S S L F I N N NI O O W R N I I I C C C C C C C C E E E E E E E E R R R R R R R R

l s s m s in F L a nw F in 2 in O in r r n o N n o t o o we G 2 M 2 C F P f F O A r wt W wt r a a o en or E F F F C C 1 o or pr b de de D de eu a k p R C C C F F r k p K k p ri l l lao L lao r ae a o P V. V-X X X wd W wd eMv wd U K K T T T t n iv r t n t e t n K C C H H H pa C pa Rn s pa e s E e eb e M s hv o c sd P C C C C C s 1 1 1 1 1 c sd R c sd ae x ee x ee x ee n o e, l c F P / P P P c n

e. l n

e. l c

1 t a 8n ea S S F S S S en o en 1 r e 1 1 r e 4 4 1 1 r a t t t v 5 o 's ai n n n3 n n3 n ai d t t t e r e e e ai - i e s ss= e4 e4 e3 e4 e3 e4 e e r a3t sse s s sse at p a a at p S9ceb a at P v3 v3 vO v3 vO v3 n x b C Cee ERps C Ce* O O n x n* e3 e3 eW e3 eW e3 Ce e n O t D 1 sms e s sms OTs nW nW n9 nW n9 nW nte s sm5 ot o1 o-ot o-ot t a es a a a es Dic a a e 9 9 L 9 L 9 Cl ei" d - d s eie mS e eie r u eL e L-dE r r eN e L-dE d - d a e r ul mRa ul eN eL is mqe hI m mqe oSu a m m q sE sE sI sE sI sE sO a a er s a a aN aN a y a er FWih aN a aN r vc s sre *' l sr a EL s sr a bI bI bin n bI bi bI anA k 1 5 5 5 4 5 s 0 0 0 0 0 0 i 1 1 2 2 2 E E E E E E R 2 1 0 0 0 0 0 0 0 0 9 2 9 4 2 P E E E E E E E 1 1 3 1 2 2 F 0 0 0 0 0 0 0 3, 3, 3, 3, 3, 1 2 1 5 3, S 5 1 1 2, 2, 2, 2, 2, 2, e 1 1 1 1 1 1 h 3 1 2 3 1: 2 3 s s s s s s t e e e e e e e e e e e e e n s s s s s s s s s s s s s a a a a a a a a a a a a a i d C C C C C C C C C C C C C esU is e t s e hl,e a l e h ob e t t e m inoa r e h r c inSo len g r r pl e g s e ia n e leOe viv n a e h I t v a u eLr lel a h g 3 e h c n s e e 4 q - a v e N gem c is a e l d gWn x d h a u s t ml r s w Ti e c r lc m F p x e sDig t s o e m e p n /s sKs el s e n u t o i e ele en he p a r c l c reWb edo e o r t a duh gp g h t e i l a no m ls g u dCv lc c h n w n a sere oio om lo lo ia a t t b aRs ls s l oe o o f h f b d cm c p m c o o o e x l l n le e e ls o e o r t t T ia oV o e u s t s t r P t c u e f pr p p f y a e n b p e o h s e v t .~ e F m lep len o Fv x t v hl e ss e uo ui a v n A t f f a b Pl o E le u m s f W f f y ot o s s e o Sa P .is sin u ee er e r/ r en 2 le r s p r u ulv n n ip lu r h a e d a lu or e i e t b u m 1 la po a a g f k av f f o f n o k mh a 3 okia ol 2 ek eo e a e t t a u0 M s e st s h r c d e av sw s5 lswa nf P uh u s u c un l l b iage a c a e a x d a ia gi o F c cv c e a a f r nm nr s S e l n a gg f f t e T r e ng na r ni ut s o nt s r or ov o e mn o t uae a t t u ma mk m h ga adpo adwc lu ePei eP de g mh mc m F gh r r x cv c e pF xb pFn e os oh o P luc n i x OSeo OSai R Cd Cc C S Pe L S I IO O F H M L L P U F F-F R G N N W W K C C C T L D D M C C C F P K K G V-V- X X X L L E K K T T T W-W- R C C H H H C C P C C C C C E E F P P P P P R R S S S S S S

n n n i i i t F A A o 2 1 1 n n ai i F C C t s n C F F hn o t o h P P P di a D D D t ea m M M M f t r n o e e f 1 1 1 hm n2 C C C t i1 P P P nu nl. S S S r ot oo 1 s dn dV t t t n n n e ei e5 e4 e4 e4 0 s sf s7 v3 v3 v3 4 a ao a2 e3 e3 e3 7 C by b 0 0 0 1 n/ n/ n/ 1 t o dh eid eG o6 o6 o6 G 9 9 9 t t mbe aE d d - d - r air n e L-eL eL E o s R i uliu uR sE sE sE t is U n saq sU aN aN aN t y a sv e EN bI bI bI N s aar lanA 3 4 3 3 k 0 0 0 0 s E E E E i 5 5 1 0 0 0 0 9 9 9 R P E-E E 2 5 3 3 F 0 0 0 3, 3, 3, 3, S 1 1 1 2, 2, 2, 2, e 1 1 1 1 h 1 2 3 s s s s t e e e e e e e n s s s s s s s i a a a a a a a d C C C C C C C esU se icneu n n q u u r r e s r p d d n n F m a a / y s b u t t e d p r r a a n g t t i it a n s s li t lo t t b s i o o a g o ls ls b n s c ia ia i lt l [f o ru u o f r T a o 2 l d P f p n p F hp t o F la le m n c e it i f A u u A c r p t v n c f R E u e o g g R le e in f fo l u lo r D 2 D a ia c c s c l o s 1 o o f lo lo l 3 l es o o l e le e u P P lb n n a le le a n n c a a n u u T h h o F F c c m t t n n l le e m e e v v e e o p p L L C S S 1 2 F F F F P C T T O O C F F L L P P P L-L V V M M M L L P P P C C C C C P P P P P S S S S S

I r ' Table 3.1-3 ' Spent Fuel Pool Cooling Risk Analysis Frequency of Fuel Uncovery (per year) INITIATING EVENT D W 1T CASE 2 CASE 3 l Loss of Offsite Power-Plant 1.3E 4.2E-06 8.0E-05 centered and grid related events Loss of Offsite Power-Events. 1.4E-06 9.4E-06 1.4E-05 initiated by severe weather Intemal Fire DM57T 1.0E-06 9.0E-06 Loss of Pool Cooling - 1.5E-07 1.7E-07 1.7E-05 ) Loss of Coolant inventory 2.9E-06 6.0E-05 1.3E-04 Seismic Event 2.0E-06 2.0E-06 2.0E-06 l Cask Drop 2.5E-06 2.5E-06 1.5E-05 Aircraft impact M0gg 4.0E-08 4.0E-08 Tornado Missile 5.6E-07 5.6E-07 5.6E-07 j Total

1.2E-05

'8.0E-05 2.7E-04. i 3.3 Probabilistic Assessment insights 1. . The frequency of spent fuel pool fuel uncovery appears to be driven by the degree of defense-in-depth, redundancy, and diversity in the sled mounted spent fuel pool cooling systems (including makeup capability and support systems) at decommissioning plants that replaced operating reactor spent fuel pool cooling systems and backup systems (e.g., residual heat removal system). 2. Preliminary results indicate there is the possibility of a factor of 20 or more increase in the frequency of fuel'uncovery if utilities choose to reduce their spent fuel pool cooling capabilities much below that found at the four decommissioning sites visited by the . working group in preparation for the spent fuel pool risk assessment. 3. One year after the last fuel is transferred from the reactor to the spent fuel pool, given appropriate procedures, instrumentation, equipment, and operator awareness, there -

appears to be sufficient time for operators to respond to most loss of spent fuel pool

cooling initiators and terminate the event before there is a potential for offsite dose consequences. This is not necessarily true for heavy load drop and very large seismic events that have the potential to rapidly drain the spent fuel pool and uncover the fuel. If one of these initiators were to occur during the first year or two after the last fuel was

transferred from the reactor to the spent fuel pool, it appears that there would be only five to seven hours available for local emergency response. This may be too short for an

. effective evacuation to preclude some early fatalities.- 45

r 8 1 I Lack of action levels higher than the " alert" level at decommissioning plants may need to 4. be rervaluated. The working group is evaluating whether the timing of a waming should be commensurate with the potg@uences of having a particular release at a decommissioning plant. In performing its preliminary risk assessment, the working group evaluateo seismic-5. induced risk at decommissioning plant spent fuel pools. The working group is evaluating whether the risk is sufficiently low to allow exemptions for emergency preparedness in the area of seismic events if the HCLPF of the spent fuel pool can be shown to be at least ' c basis, utilities seeking exemptions might three times the site's SSE. Onigprf-i follow the efforts described belewtal e nfidence that the as-built spent fuel pool has a sufficiently robust HCLPF. Walk down the pool structure and its vicinity and note: physical conditions such as cracking, spalling of concrete, signs of leakage or a. leaching and separation of poolwalls from the grade surface, arrangement and layout of suppotting columns and shear walls, assessment of b. other loads from tributa h yte gap carried by the supporting structure of the l pool, as-built dimensior s's ntfhbppng of any existing structural cracks, ' adjacent structures that can impact the pool structure above and below the grade c. surface, supporting arrangement for superstructure and crane and potential for failure of the superstructure and the crane, the weight of the heaviest object that can drop in the pool structure and the corresponding drop height. Calculate the seismic capacity of the pool structure. Typically such a calculation consists of thefollowing: review existing layout drawings and structural dimensions and reconcile the a. differences, if any, between the as-built and as-designed information and consider the effects of structural degradation as appropriate, b. from design calculations, determine the margin to failure and assess the extrapolated multiple of SSE level that the pool structure should survive, determine whether or not design dynamic response analysis including soil-structure interaction effects are still applicable at the capacity level seismic event; if not, conduct a new analysis using properties of soil at higher strain levels and reduced stiffness of cracked reinforced concrete, determine the loads from pool structure foundation uplift and from impact of pool c. structure with adjacent structures during the capacity level seismic event, determine loads from the impact of spent fuel rack on the pool floor and the side . walls and determine the loads from dropping of heavy objects from the collapse of superstructure or the overhead crane, d. determine a list of plausible failure modes; failure of side walls due to the worst loading from the capacity level earthquake in combination with fluid hydrostatic and sloshing head and dynamic earth pressure as appropriate, failure of the pool floor slab in flexure and bending due to loads from the masses of water and the 46

spent fuel and racks, local failure by punching shear due to impact between structures and the spent fuel racks or dropping of heavy objects, DRAFT e. - the calculations to determine the lowest structural capacity can be based on ultimate strength of reinforced concrete structures due to flexure, shear and punching shear. When conducting an yield line analysis, differences in flexural yield capacities in two orthogonal directions and for the negative and positive bending moments influence the crack pattoms and several sets of yield lines may have to be investigated to obtain the lowest capacity. For heterogeneous materials, the tradition ainlysis provides upper bound solutions; consequently, consider i is r eeded to determine the structural capacity based on the yield lines that approximate the lower bound capacity. - 6. Heavy load drops are potentially the moi,t significant loss of spent fuel pool cooling _ initiators. A heavy load drop has the potential to fail the spent fuel pool, such that it would drain rapidly and would be incapable of being reflooded from onsite or offsite sources. The risk perspective from zirconium fires during their period of vulnerability can be significantly improved by control of movement of heavy loads near or over the spent fuel. P '- DRAFT 47

4.0 Supporting Information 4.1 Maintenance Rule DRAFT 4.1.1 Identification of Maintenance Rule Concepts at Decommissioning Plants Maintenance Rule (MR) concepts at decommissioning plants are expressed in 10CFR50.65, " Requirements for monitoring the e#ectiveness of maintenance at nuclear power plents." The rule also provides the fundamental regulatory requirements for maintenance rule and scoping of structures, systems, and components scommissioning status plants (DSPs). Scoping requirements have been established ... o n aclear power plant for which the licensee has submitted the certifications specified in 50.82(a)(1), this section only shall apply to the extent that the licensee shall monitor the performance or condition of all structures, systems, and components associated with the storage, control, and maintenance of spent fuel in a safe condition, in a manner sufficient to provide reasonable assurance that such structures, systems, and components are capable of fulfilling their functions..." Normally, a DSP will have already submitted the 50.82(a)(1) certification; this will satisfy the status statement for a plant as quoted Sp;from a practicable maintenance rule viewpoint we only need worry about how the DS e F rovides for. "... monitoring the performance or condition of all SSCs associated with the storage, control, and maintenance of spent fuelin a safe condition...." 4.1.2 Identification of Potential Systems, Equipment, Functions at Decommissioning Plants The Office of Nuclear Reactor Regulation (NRR) staff has conducted workshops and provided a " staff support member" for three onsite inspections of MR implementation at shutdown plants. Consequently, by consultation at the workshops and precedent at the three sites inspected, a scoping list of expected SSCs has been developed. Among those generic SSCs that may be in the scope of the Maintenance Rule for DSPs are the following: SFP cooling and cleanup (SFPCC) SFP structure and any connecting piping system seals e e SFP building Radiation monitors above or in the SFP Standby service water system (i.e., the portion that is heat sink for SFPCC heat e exchangers) For boiling water reactors, the residual heat removal system can also be used for SFP e cooling Leak detection system (this system detects leakage from the SFP, the transfer pool, and e reactor well pool liners. Portions of this system that detect leakage from the SFP could be under the scope of the Maintenance Rule) 48 i 5

Heating, ventilation and air conditioning system above the SFP e l SFP waterlevelinstrumentatio@gry SFP emergency or normal makeup water supply i e i SFP crane (i.e., may need to be included as a result of accident scenarios involving dropped fuel bundles, which could cause cladding damage and potential radiation exposure to personnel in the SFP area.) e Standby auxiliary ac power sys r . h power supply to SFPCC pumps) SFP transfer tube penetration seals or bellows, pneumatic air system which inflates the i seals, and transfer tube gate seals (i.e., failure of these seals could cause a SFP drain j down to a few feet above top of active fuel, the area around the SFP would become a j very high radinon area, and plant staff would need to evacuate the SFP area). ) This list will vary depending on the circumstances of each individual DSP and is only intended to be a representative list. MMT The MR requires that SSCs be monitored for reliability and availability by the licensee. MR policy recognizes that the longer a plant is shutdown the less likely that fuel is vulnerable to zirconium oxidation. Consequently, the stringency of reliability and availability standards are graded accordingly. q J 4.1.3 Evaluation of the Maintenance Rule Oversight at Decommissioning Plants q Background For implementation of the MR For Decommissioning Status Plants j The maintenance rule was effective July 1996, and amended August 1996, to include nuclear plants which have submitted the certifications specified in 10 CFR 50.82(a)(1) (i.e., l deccmmissioning status plants). As amended, the rule requires licensee's to monitor the performance or condition of SGCs associated with the storage, control and maintenance of spent fuel in a safe condition and in a manner sufficient to provide reasonable assurance that such SSCs are capable of fulfilling their intended functions. i Subsequent to the August revision of 650.65, the associated RG 1.160, " Monitoring The Effectiveness Of Maintenance At Nuclear Power Plants," and NUMARC 93-01," Industry l Guideline for Monitoring the Effectiveness of Maintenance at Nuclear Power Plants," were revised to address the amended rule provisions for decommissioning status plants. l (NUMARC 93-01 is the industry guideline for monitoring the effectiveness of maintenance at nuclear power plants and has oeen endorsed by tha NRC's Regulatory Guide 1.160.) Overview of implementation of the MR for Decommissioning Status Plants implementation of the maintenance rule for decommissioning status. plants has been fundamentally the same as for operating facilities; however, the population of SSCs within scope were reduced. 49 l l J

['. o \\ LRelative to the scope of SSCs that would be applicable for a docommissioning status plant, ~ licensee's need to focus on the functions necessary for the preservation of spent fuel in a safe n" - condition. For decommissioning statug Ili.T*Xplicit process related to risk ranking of n SSCs, delineated in NUMARC g3-01, may not :e necessary. Expectations For implementation of the MR for Decommissioning Status Plants . Licensees for decommissioning status plants will continue to establish SSC performance measures and condition monitoring as well as the requisite goals and preventive maintenance activities consistent with the requiremgfg An SSC would be considered to be in the (a)(1) category when the performance or condition of the SSC does not meet established goals and corrective actions are required. An SSC would be considered to be in the (a)(2) category when it has been demonstrated that the performance or condition of the SSC is being offectively controlled through the performance of appropriate preverttive maintenance. Licensee's Periodic Evaluation on g g of the MR for Decommissioning Status Effectiveness of the maintenance rule implementation process will be evaluded by licensee's at a periodicity not to exceed 24 months between evaluations. Balancing reliateility and availability are part of the periodic evaluations. Preventive maintenance safety assessments are done on an ongoing basis. . uality Assurance Q 4.2 Part 50 of Title 10 to the CFR, Appendix B quality assurance (QA) requirements apply to all-activities affecting the safety-related functions of those SSCs that prevent or mitigate the - consequences of postulated accidents that could cause undue risk to the health and safety of the public. These activities include designing, purchasing, fabricating, handling, shipping, storing, cleaning, erecting, installing, inspecting, testing, operating, maintaining, repairing, and modifying. Safety-related structures, systems, and components are those SSCs that are relied upon to. remain functional during and following design basis events to assure: (1) The integrity of the reactor coolant pressure boundary, (2) The capability to shut down the reactor and maintain it in a safe shutdown condition; or,. - (3) The capability to prevent or mitigate the consequences of accidents which could result - in potential offsite exposures comparable to the applicable guideline exposures set . forth in $50.34(a)(1) or $100.11 Upon docketing of certifications required by $50.82(a)(1) for permanent cessation of operations and permanent removal of fuel from the reactor vessel, the license under Title 10 of the CFR no 50 c L

L 4 j l longer authorizes operation of the reactor or emplacement or retention of fuel into the reactor j vessel. After the reactor fuel has been permanently removed from the vessel and placed in the spent fuel pool, potential offsite releases are the primary safety consideration. Sutuequent to permanent relocation of all fuel to the spent fuel pool, licensees may apply the $50.5g regulatory change control process goSCs that no longer perform safety functions associated with maintaining _the reactor coolant pressure boundary or with the capability a of shutting down the reactor. Reclassification of safety-related equipment through the $50.5g ] process may considerably reduce the number of SSCs subject to Appendix B QA requirements. Licensees generally retain their approved 4 i@A programs through the decommissioning phase. Changes to QA y cor nmitments are submitted for NRC approval through the $50.54(a)(3) change control process. Although considerable simplification in the management, administration, and oversight of the QA program may be possible, the eighteen Appendix B criteria are generally applicable to any equipment or structure that performs a safety j function. Beyond decommissioning activities, licensees can use their approved Appendix B programs (following review for applicabilityMng and transportation of radioactive material (Part 71) and for storage of spent nuclear fuel and high-level radioactive waste (Part 72). An attemate approach, planned by one licensee, would be to submit a new quality assurance . program specifically applicable to decommissioning activities. If such a program were submitted, l the staff would review it for conformance with Appendix B criteria in accordance with Standard Review Plans (SRP) 17.1/17.2. Pursuant to $50.34(b)(6)(ii), this review would extend to the determination of how Appendix B requirements would be satisfied through licensee commitments to applicable regulatory guides and industry standards. The staff would review for acceptability . any proposed attematives to NRC-endorsed industry standards. 1 L Changes to technical specification (TS) administrative controls ($50.36(c)(5, 6)) related to quality assurance are reviewed in accordance with applicable standard review plans, regulatory guides, industry standards, and other regulatory guidance, such as the proposed standard technical specifications for permanently defueled Westinghouse plants (NUREG-1625). Relocation of TS g j administrative controls to the licensee's QA program description are reviewed using the guidance . contained in Administrative Letter 95-06. For design basis events which could result in potential offsite exposures comparable to the applicable guideline exposures set forth in $50.34(a)(1) or $100.11, the quality assurance requirements of Appendix B to 10 CFR Part 50 apply. l 4.3 Water-basin Type independent Spent Fuel Storage Installations l A spent fuel storage facility built separately from a reactor facility would be licensed under the regulations of 10 CFR Part 72 [Ref.1). This facility would be called an independent spent fuel storage installation (ISFSI). The a are'two types of ISFSis, a dry cask facility and a water-basin facility. Many decommissioning plants isolate the SFP area from the plant so that I -. decommissioning activities can proceed with minimalimpacts on the SFP area. The concept of a " spent fuel pool island" is similar, although not identical, in nature to an ISFSI. The working group j explored the design considerations and controls for an ISFSI to identify items that should be consistent between the two types of facilities.' 51 1

m Regulatory Guide 3.49 [Ref. 2) provides a method to comply with the design requirements for a . water-batin ISFSI by endorsing, with provisions, American National Standard Institute /American Nuclear Society standard (ANSI /ANS) 57.7-1981, " Design Criteria for an Independent Spent Fuel Storage Installation (Water Pool Type)" [Ref. 3]. The function of the ISFSI is to provide interim protective custodial storage of spent fuel. By the standard, it is expected that the facility would be completely independent or adjacent to a nh eTstation or reprocessing faciiity; however, no direct means to transferring fuel assem ies rom the nuclear facility to the storage installation can exist. This basic function is the same for a SFP at a nuclear power plant. However, the SFP is required to have the capability to be connected to the facility to receive spent fuel directly from the reactor. SFPs at nuclear power plants were designR M.EnTiicensed in accordance with 10 CFR Part 50 regulation [Ref 4). Without a demonstrated substantialincrease in the overall protection of public health and safety or the common defense and security in accordance with 10 CFR 50.109 [Ref. 5), the staff can not require the modification of or addition to systems, structures, components or design of a facility.DRAFT While the requirements from the Part 72 standard for a design of an ISFSI can not be applied to SFP facilities that are already built, it is recognized that the design considerations and controls may provide insights on what should considered important for decommlssioning plants. An evaluation of Part 72 considerations and controls will also provide consistency within the agency where possible. 4.3.1 Design Considerations There are four overall design considerations in ANSI /ANS 57-7-1981. Each of these are described below as they applies to spent fuel pools at decommissioning plants. Overall Design Consideration A: Short-lived, high specific activity radionuclides, particularly of iodine and xenon are no longer present in significant quantities in spent fuel that has aged for more than one year since discharged from the reactor core. This design consideration describes the natural decay process of radionuclides in spent fuel foMHng the discharge from a nuclear reactor. The statement also applies to spent fuel at decommissioning plants. Overall Design Consideration B: The storage of aged spent fuel is a low hazard potential activity. Very little of the radioactivity present is available in a dispersible form and there is no mechanism present to cause the release of radioactive materials in significant quantities from the installation. This criteria is used to qualitatively assess offsite consequences. An NRC Backfit Review Panel [Ref.13] determined that a significant loss of spent fuel pool water accident that results in fuel uncovery is a reasonable credible accident to consider for exemptions to some 10 CFR Part 50 regulations, such as emergency planning. If the decay heat is sufficient to increase clad 3 temperature to the point of self-sustaining oxidation, a zirconium fire could occur (see Section 2). { if a fire occurred, significant quantities of radioactivity could be released (see Section 3.3). For fuel uncovery events, considering high burnup fuel (up to 60-62 GWD/MTU) and high density 4 52 }

o t spent fuel racking in the storage pools, a zirconium fire is possible beyond one year following j permanent shutdown of the plant. i o'- Overall Design Consideration C: Decay heat is not a significant design consideration in an ISFSI because the spent fuel shall have aged a minimum of one year since discharge from the actor DRAFT For a nuclear power plant, the capability to remove spent fuel decay heat is a design criteria for . the SFP cooling system. General Design Criterion 61 requires, in part, that the system shall be designed with residual heat removal capability that reflects the importance of decay heat. SFP cooling and makeup systems at nuclear g-: m!= ( from a whole core of freshly discharged sp se MI'At re designed to remove the decay heat decommissioning plant, the amount of cooling that is needed reduces as the decay heat decreases, o Overall Design Consideration D: Because of the low potential for airborne radioactive materials above a spent fuel pool, the buildinET housing such pools need not be designed i ' foTconfineaant. DRA The buildings for spent fuel pools are not required to be designed for confinement of airbome radioactive material. This design consideration is consistent with the design criteria used for spent fuel pools at nuclear power plants. . The ANSI /ANS standard states that the sole function of the ISFSI is the interim protective custodial storage of spent fuel. While this function is the same function for a spent fuel pool at a nuclear power plant, some of the design requirements are different due to the different emphasis j that is needed for the storage of the fuel in each facility. In the design of a spent fuel pool, the decay heat removal capability must be sufficient to handle a whole core of freshly discharged

• 1 spent fuel. In contrast, s' water-basin ISFSI cannot accept spent fuel until it has decayed at least one year. Another significant difference in the operation of a water-basin ISFSI and a spent fuel pool at a decommissioning plant is that an ISFSI is expected to routinely receive spent fuel shipments, while at a decommissioning plant, except for final removal from the pool, minimal fuel movement is expected. As such, the decay heat load at an ISFSI will continuously fluctuate, while the heat load at a decommissioning plant spent fuel pool will generally only decrease. With

' distinctions such as these is mind, the working group believes that not all of the controls required for a water-basin ISFSI would be necessary for a spent fuel pool at a decommissioning plant. j 4.3.2 Defense-in-Depth

The defense-in-depth concept is used in the design of a water-basin ISFSI. The primary barrier is the fuel clad. Several secondary barriers are provided to control radioactivity. : The pool water i

itself retains radioactive particulate material and provides plant personnel shielding. Subcriticality l is provided by the design of the spent fuel storage racks and design features of the handling system ~ The pool clean up system is provided to remove dissolved and suspended radioactive material. A clean water environment is important to minimize corrosion, coolant cleaning and water chemistry monitoring are means to ensure a quality environment is maintained. A sufficient source or sources of water should be available for routine makeup and be of sufficient capacity for event mitigation activities. Indication of loss of water, including level indication and radiation monitors, are important for early indication to provide personnel time to take compensatory 53

e i actions. Further, proper load handling controls are needed to ensure the safety of the fuel. The working group believes that defense in depth should be maintained for spent fuel pools at decommissioning plants to protect from accidents that could lead to a radiological release. For an operating reactor, defense in depth is considered to be maintained through fission product - barners. These barriers include the clad, the reactor pressure boundary, and containment in a . spent fuel pool, the only fission product bagpifains is the fuel clad., which is similar to the water-basin ISFSI. 4.3.3 Accident Scenarios ~ ISFSis are designed to withstand a spectne ts. The accidents range from those that ~ - are expected to occur frequently to those tu ly considered "possible", but could have potential maximum impact on the environment. The latter of the accidents include events such as criticality, total loss of water, and a dropped cask. The loss of water accident analysis has only been evaluated for the effects from shine. Part 50 of Title 10 of the CFR provides regulations that dictate design criteria and accidents for spent fuel pools. The working group believes that a loss of water accident with @MjtraTfor seif-sustained zirconium oxidation is = - valid consideration for severe accidents at spent fuel pools. 4.3.4 Systems and Controls ~ The ISFSI design emphasizes the importance of controlling radioactivity. The following SSCs . and control programs that are identifiexf in the guidance as important for the operation of a water-basin ISFSI and may be considered important for spent fuel pools at decommissioning plants. SFP water cooling and cleaning system ' e-LSFP water makeup system . SFP water levelinstrumentation e . Radiation monitoring Heating ventilation and air conditioning system (HVAC) e On-site and off-site electrical power, instrumentation and controls, and communications e . Water chemistry program e SF cask handling program e SF handling program-Radioactive waste treatment ' 4.4 Design Basis Events Design criteria for the storage of spent nuclear fuel requires that fuel storage and handling systems be designed to assure adequate safety under normal and postulated accident conditions. In addition, these systems are to be designed with appropriate containment, confinement, and filtering systems, and be designed to prevent significant reduction in the -

coolant inventory under accident conditions. Design guidance for fuel storage facilities includes preventing the loss of SFP coolant from the pool that would result in the fuel becoming uncovered. protecting the fuel from mechanical damage, and providing the capability for limiting potential offsite exposures in the event of a significant release of radioactivity from the fuel.

i 54 l'

Unless mitigative measures are taken, loss of water from the SFP during the initial decay period l of about 3 to 5 years could cause overheating of the spent fuel and result in damage to the fuel - cladding integrity and could result in release of radioactive materials to the environment. Events considered in the spent fuel storage facility design with the potential to cause significant inventory L loss from the SFP include: DRAFT Earthquakes = Loss of cooling : Siphoning Dropping of heavy loade g Other extemal events that could c.c,]iinviiii.s the inventory of the pool Fuel storage facility features designed to prevent 'significant inventory loss from the SFP include a seismically qualified SFP structure, redundant, and in some cases, seismically qualified SFP cooling systems (some with safety related ggrg jos), seismically qualified and safety-j related SFP coolant makeup systems, anti % iu,,,, action, local and remote level indication,- local and remote radiation' alarms, pool structure designs that plan for the effects of a dropped spent fuel cask without significant leakage from the fuel storage area. Even a small amount of mechanical damage to the spent fuel stored might cause an offsite radiological release if no dose reduction features, such as iodine removal, are provided. Mechanical damage resulting in radiological releases from the spent fuel can occur from the I following events: ~ Dropping fuel assemblies during fuel handling operations L Dropping objects onto the stored fuel Impacting the stored fuel with intemally or extemally generated missiles Fuel storage facility features designed to prevent mechanical damage to the stored fuel include a pool structure design that prevents missiles from impacting the stored fuel, and physical design features and administrative controls are used that minimize the possibility of damaging the stored 4 fuel by dropping a fuel assembly or dropping a heavy object onto the stored fuel. l For SFPs not located within the reactor containment, without adequate protective features, radioactive material could be released to the environs as a result of either loss of water from the i SFP or mechanical damage to the fuel within the pool. Should radioactive material be released from the stored fuel, design requirements to provide a controlled leakage building surrounding the SFP with the capability to limit releases of radioactive material address this concem. \\ r 55-

3 5.0 . References References for the Executive Summary and Section 1.0 Benjamin, A.S.,' McCloskey, D.J., NTand Dupree, S.A, " Spent Fuel Heatup 1. Following Loss of Water During Storage," NUREG/CR-0649 (SAND 77-1371), March 1979. 2. - Sailor, et, ali, " Severe Accidents in Spent Fuel Pools in Support of Generic Safety issue 82", NUREG/CR-4982 (BNL-NUREG-52093), July 1987. U.S. Nuclear Regulatory Commissig,5tg ah$wy Analysis for the Resolution of Generic 3. Safety issue 82, "Beyond Design Busi s Mc6 de6ts in Spent Fuel Pools," NUREG-1353,- . April 1989. 4. ' U.S. Nuclear Regulatory Commission, " Residual Decay Energy for Light-Water Reactors for Long-Term _ Cooling," Branch Technical Position ASB 9-2, Rev.2, July 1981 from the " Standard Review Plan for the Rev' f a foty. Analysis Reports for Nuclear. Power. 5. American Nuclear Society, Draft A fB.6, " Decay Energy Release Rates Following Shutdown of Uranium-Fueled Thermal Reactors," October 1973. 6. U.S. Code of Federal Regulations, " Requirements for Monitoring the Effectiveness of Maintenance at Nuclear Power Plants," Section 65, Part 50, Title 10, " Energy." 7. U.S. Code of Federal Regulations, " Quality Assurance Criteria for Nuclear Power Plants and Fuel Reprocessing Plants," Appendix B, Part 50, Title 10, " Energy." 8. U.S. Code of Federa/ Regulations, " Conditions of Licenses," Section 54, Part 50, Title 10 " Energy." 9. U.S. Code of Federal Regulations, " Licensing Requirements for the Independent Storage of Spent Nuclear Fuel and High-Level Radioactive Waste," Part 72, Title 10, " Energy." References for Section 2.0 1. U.S. Code of Federal Regulations,." Changes, Tests and Experiments," Section 59, Part 50, Title 10, " Energy." 2. Benjamin, A.S., McCloskey, D.J., Powers, D.A., and Dupree, S.A,

• Spent Fuel Heatup Following Loss of Water During Storage," NUREG/CR-0649 (SAND 77-1371),

March 1979. 3. Sailor, et. al., " Severe Accidents in Spent Fuel Pools in Support of Generic Safety issue 82", NUREG/CR-4982 (BNL-NUREG-52093), July 1987; 4. U.S. Code of Federal Regulations, " Requirements for Monitoring the Effectiveness of - Maintenance at Nuclear Power Plants," Section 65, Part 50, Title 10, " Energy." 5.' ~ U.S. Code of Federal Regulations, " Quality Assurance Criteria for Nuclear Power Plants and Fuel Reprocessing Plants," Appendix B, Part 50, Title 10, " Energy." 6. U.S. Code of Federa/ Regulations, " Licensing Requirements for the Independent Storage of Spent Nuclear Fuel and High-Level Radioactive Waste," Part 72, Title 10, " Energy." j; References for Section 2.1 1. Benjamin; A.S., McCloskey, D.J., Powers, D.^., and Dupree, S.A, " Spent Fuel Heatup Following Loss of Water During Storage," NUREG/CR-0649 (SAND 77-1371), March 1979. 56 L 3; y v

2. ' Benjamin, A.S., " Dynamic Modeling of Physical Phenomena for Probabilistic Assessment of Spent Fuel Accidents," Sandia National Laboratories, SAN 097-2870C, November 1997. 3. Mellor, A.M., " Heterogenous Ignition of Metals: Model and Experiment", Aerospace and Mechanical Sciences Report No. 8@li))$GTant NsG-641, Department of Aerospace and Mechanical Sciences, Princeton University, Princeton, New Jersey, October 1967. 4. Tapscoctt, R.E., Fire Protection Handbook,18* Edition, Section 4, Chapter 16. " Metals," National Fire Protection Association, Quincy, MA, February 1997. 5. Cooper, " Review of Zirconium-Zircaloy Pyrophoricity," RHO-RE-ST-31P, Rockwell International, Rockwell Hanford Og Abbud-Madrid, A., Branch, M.C., F.._..gftpland, WA, November 1984. 6. ..i, T.J., and Daily, J.W., " Ignition of Bulk Metals By A Continuous Radiation Source in A Pure Oxygen Atmosphere," Flammability and Sensitivity of Materials in Oxygen-Enriched Atmospheres: 6* Volume, ASTM STP 1197, Philadelphia,1993. 7. Levitz, N.M., Kullen, B.J., and Steir

1. Pyrophoricity and Compaction", e(It

.m.")lanagement of Waste Cladding Hulls. Part 7, Argonne National Laboratory, February 1975. 8. Markstein, G.H., " Combustion of Metals.," Technical Report CAL-86-P, Comell Aeronautical Laboratory, November 1962. 9. Glassman, l., Combustion, Third Edition, Academic Press, New York, NY,1996. 10. Department of Interior, "The Explosivity of Titanium, Zirconium, Thorium, Uranium and Their Hydrides," NYO-1562, U.S. Atomic Energy Commission and Department of Interior, Oakridge, TN, June 1951. 11. U.S. Department of Energy, " Primer on Spontaneous Heating and Pyrophoricity," DOE HDBK-1081-94, Washington, DC,1994, 12. National Fire Protection Association, NFPA 482," Standard for the Production, Processing, Handling and Storage of Zirconium," Quincy, MA,1996, 13. Bulmer, G.H. " Recommendations on the Safe Handling or Zirconium Metal and Zirconium Alloys," United Kingdom Atomic Energy Authority, Risley, Warrington, Lancashire,1969. 14. Tapscoctt, R.E., Fire Protection Handbook,17* Edition, National Fire Protection Association, Quincy, MA. 15. Schulz, Wallace, "ARH-2351 Shear-Leach Processing of N-Reactor Fuel -- Cladding Fires," Separations Chemistry Laboratory, Research and Development, Chemical Processing Division, Atlantic Richfield Hanford Company, Richland, WA, February 15,1972. . References for Section 2.2 1. Benjamin, A.S., McCloskey, D.J., Powers, D.A., and Dupree, S.A, " Spent Fuel Heatup Following Loss of Water During Storage," NUREG/CR-0649 (SAND 77-1371), March 1979. 2. Nourbakhsh, et. al., " Analysis of Spent Fuel Heatup Following Loss of Water in a Spent Fuel Pool", NUREGICR-6441 (BNL-NUREG-52494), May 1998. 3. Smith, C.' W., " Calculated Fuel Perforation Temperatures, Commercial Power Reactor Fuels", NEDO-10093, September 1969. 4. Sailor, et. al., " Severe Accidents in Spent Fuel Pools in Support of Generic Safety issue 82", NUREG/CR-4982 (BNL-NUREG-52003), July 1987. 57 v h

{ 5. MELCOR 1.8.1 Computer Code Manual," Volume 2: Reference Manuals and j Programmers' Guides, June 1991. ' 6.' Pisano, et. al., "The Potential for Propagation of a Self-Sustaining Zirconium Oxidation Following Loss of Water in a Spent Fuel Pool" Sandia National Laboratory draft report prepared for the U.S. Nuclear Reg @Ihission, January 1984

7.

. Travis, R.J., et.al.', "A Safety and Regulatory Assessment of Generic BWR and PWR Permanently Shutdown Nuclear Power Plants," NUREG/CR-6451 (BNL-NUREG-52498), August 1997. 8. . Hermann, et.al./ Technical Support for a Proposed Decay Heat Guide Using SAS2H/ORIGEN-S Data", NUREGiCP 0 fAember 1994. L 9. U.S. Nuclear Regulatory Commissi M, " o pe idium of ECCS Research for Realistic LOCA Analysis", NUREG-1230, December 1988. Cooper, T. D., " Review of Zirconium-Zircoloy Pyrophoricity," RHO-RE-ST-31P, Rockwell 10.- intomational, November 1984. ~ ' 11. Kullen, et. al.,

• An Assessment of 7 i

Jophoricity and Recommendations for - Handling West Hulls", ANL-77-63, ; aiional Laboratory, November 1977c Referenoas for Section 3.3 1. - U.S. Nuclear Regulatory Commission, " Residual Decay Energy for Light-Water Reactors for Long-Term Cooling," Branch Technical Position ASB 9-2, Rev.2, July 1981 from the " Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power 2. Sailor, et. al. " Severe Accidents in Spent Fuel Pools in Support of Generic Safety lasue 82", NUREG/CR-4982 (BNL-NUREG-52093), July 1987. 1 3. " Code Manual for MACCS2," NUREG/CR-6613, May 1998. 4. U.,S. Nuclear Regulatory. Commission, " Severe Accident Riska: An Assessment for Five U.S. Nuclear Power Plants," NUREG-1150, December 1990.' L References for Section 3.3.2 1.- Sailor, et. al., "Seve're Accidents in Spent Fuel Pools in Support of Generic Safety ' lasue 82", NUREG/CR-4982 (BNL-NUREG-52093), July 1987.

2.

Travis, R.J., et.al., "A Safety and Regulatory Assessment of Generic BWR and PWR

- Permanently Shutdown Nuclear Power Plants," NUREG/CR-6451 (BNL-NUREG-52498), August 1997, 3. " Code Manua! for MACCS2," NUREG/CR-6613, May 1998. 4. " Calculations of Reactor Accident Consequences Version 2, CRAC2: Computer Code," NUREG/CR-2326, February 1983. 5. Charles E. Ader, U.S. Nuclear Regulatory Commission, memorandum to Robert C. Jones, _ U.S. Nuclear Regulatory Commission, "Offsite Consequence Evaluation of Steam ' Generator Tube Leaks and Induced Steam Generator Tube Ruptures," July 24,1996. 6. . Charles E; Ader, U.S. Nuclear Regulatory Commission, memorandum to Robert C. Jones,

U.S. Nuclear Regulatory Commission, "ONsite Consequence Evaluation of Steam Generator Tube Leaks, Steam Generator Tube Ruptures, and Induced Steam Generator Tube Ruptures," August 30,1996.

58

References for Section 4.3 1. U.S. Code of Federa/ Regulations, " Licensing Requirements for the Independent Storage of Spent Nuclear Fuel and High-Level Radioactive Waste," Part 72, Title 10, " Energy." 2. U.S. Nuclear Regulatory Commissi@,%Tdent Spent Fuel Storage Installation (Water-Basin Type)," Regulatory Guide 3.49, December 1981. 3. American National Standard institute / American Nuclear Society standard (ANSI /ANS) 57.7-1981, " Design Criteria for an Independent Spent Fuel Storage Installation (Water Pool Type)," issued on February.19,1981. 4. U.S. Code of FederalRegulations, ggnsing of Production and Utilization Facilities," Part 50, Title 10, "Energ: 5. U.S. Code of Federa/ Regulations, "Backfitting," Section 109, Part 50, Title 10, " Energy." 6. U.S. Nuclear Regulatory Commissien, " Operating Experience Feedback Report - Assessment of Spent Fuel Cooling," NUREG-1275, Volume 12, February 1997. 7. U.S. Nuclear Regulatory Commissi' U

G)- 800, Standard Review Plan forthe Review of Safety Analysis Reports erPowerPlants, Light WaterReactorEdition, Section 9.1.3, " Spent Fuel Pool Cooling and Cleanup System," Rev.1, July 1981.

8. EDO, U.S. Nuclear Regulatory Commission, memorandum to the Commission, " Resolution of SF Storage Pool Action Plan Issues," July 26,1996. 9. U.S. Nuclear Regulatory Commission, NUREG-0800, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants, Light Water Reactor Edition, Section 15.7.4, " Radiological Consequences of Fuel Handling Accidents," July 1981. 10. U.S. Nuclear Regulatory Commission, Standard Review Plan for the Review of Safety Analysis Reports for Nuclear Power Plants, Light Water Reactor Edition, Section 15.7.5, " Spent Fuel Cask Drop Accidents," NUREG-0800, Rev. 2, July 1981. 11. U.S. Nuclear Regulatory Commission, " Control of Heavy Loads at Nuclear Power Plants, Resolution of Generic Technical Activity A-36," NUREG-0612, July 1980.

12. -

U.S. Nuclear Regulatory Commission, " Spent Fuel Storage Facility Design Basis (for l Comment), Regulatory Guide 1.13, August 1975. l 13. Samuel Collins, Office of Nuclear Reactor Regulation Director, letter Mr. Michael Meisner, President Maine Yankee Atomic Power Company, dated November 6,1998. 4 59 i

6.0. Acronyms i ANSI American National Standard institute ANS-American Nuclear Socoty .ASB NRC Auxiliary Systems BranchSM%tTms Branch) ' atm atmosphere - BNL.. Brookhaven National Laboratory - BTP. Branch Technical Position - BWR Boiling Water Reactor DRAFT CFD . Computational Fluid Dynamics-CFM Cubic Feet per Minute CFR. Code of Federal Regulations DRAFT DOE Department of Energy. DSP Decommissioning Status Plant ECCS Emergency Core Cooling System EPRI Electric Power Research Institute ET Event Tree FAA - Federal Aviation Administation FFU Frequency of Fuel Uncovery FT Fault Tree gpm Gallon (s) per Minute GSI. ' Generic Safety lasue GWD-Gigawatt-Day. HCLPF High Confidence / Low Probability of Failure HVAC Heating, Ventilation, and Air Conditioning - INEL. . Idaho National Engineering and Environmental Laboratory ISFSI Independent Spent Fuel Pool Installation kW

Kilowatt LLNL Lawrence Livermore National Laboratory MR Maintenance Rule MW '

Megawatt j MWD Megawatt-Day MTU Megaton Uranium - i NEl - National Energy Institute NFPA National Fire Protection Association. NRC Nuclear Regulatory Commission 60 i n

m-NRR NRC Office of Nuclear Reactor Regulation PRA' Probabilistic Risk Assessment ' A PWR Pressurized Water Reactor 4 DRAFT QA .. Quality Assurance RES IIRC ONice of Research RG R6gulatory Guide SF-Spent Fuel SFP Spent Fuel Pool SFPCC. Spent Fuel Pool Cooling and Cleaning System SNL. Sandia National Laboratory SRP Standard Review Plan SSC Systems,- Structures, and ComBEAFT SSE Safe Shutdown Earthquake TS Technical Specification UKAEA United Kingdom Atomic Energy Authority i i 1 61

m Appendix 1 Assumptiom, Made in the Spent Fuel Pool Probabilistic Risk Assessment Case One and Two: ~ Case One and Two are based on the as-found spent fuel pool and spent fuel pool cooling a' nd support systems during the four site visitsBR)8 fiVo aroup member. 9 1. . The spent fuel pool has one level monitor, one temperature monitor, and radiation monitors and alarms in the control room. The instrumentation is maintained in an operable state.- Annunciators are in computers, which can track trends.- ~ 7 RAFT -- 2. - The certified fuel handlers are fom Er'stneor reactor operators who know the facility, the ' surrounding community, and facility maintenance personnel well. 3. The site has operable two fire pumps (one diesel-driven and one electrically driven from offsite power). DRAFT 4. . There is limited makeup capability (with respect to volumetric flow) with the exception of fire pumps, which can provide makeup via 2.5 inch diameter hoses.

5. -

The certified fuel handlers enter the spent fuel pool area once or twice per shift (8 to 12 hour shifts). 6. The site is staffed during the day shift Monday through Thursday with maintenance people, health physics staff, QA/QC staff, fuel handlers, and administrative staff. Nights and weekends there is a skeleton staff. 7.' It is known to the site personnel (through guidance and / or training) that backup sources to replenish the spent fuel pool inventory can be fro the fire protection pumps or offsite sources, such as from fire engines. . 8. The spent fuel pool water was dear and the fuel was observable. The control room monitors the spent fuel poollevel via a camera that can zoom in on a measuring stick in the pool that can alert operators to level changes. The measuring stick is about three to four feet long. 9. There is little fire protection equipment in the spent fuel pool area and it all is manual. c 10. ' Overhead cranes have stops to help prevent heavy loads from being moved over the spent fuel pool. i A1-1

_[ Case Three: Case Three is based on spent fuel pool and spent fuel pool cooling and support systems that are slightly better than the minimum required by NRC regulations. 1. The spent fuel pool has one level MlenTParature indicator in the controi room. They are not required to be maintained in an operable state. The radiation monitors and associated alarms in the control room are only required to be opsrable when fuel is being moved. Annunc:ators are analog, are not trended, and need not work. In the PRA model, the control room / spent fuel pool instrumentation is assumed unavailable 10% of the time (i.e., the fraction of time d y" hine component is in a failed state is 1 x 104) for Case 3, which is 20 ti ti r unavailability than assumed in Case 1 (i.e.,5 x 104). 2. The certified fuel handlers are not former utility employees who know the facility, but are individuals who meet the minimug irements for a certified fuel handler. 3.' There is limited makeup capability (with respect to volumetric flow). The fire pumps have been removed from the site.

4. -

The certified fuel handlers rarely enter the spent fuel pool area. It was assumed in the PRA model that the HEP is five times higher than Case 1 (i.e., the Case 3 HEP is 0.05.) 5. The site is staffed during the day shift Monday through Thursday with a minimal number of maintenance people, health physics staff, QA/QC staff, fuel handlers, and administrative staff. Nights and weekends there is a skeleton staff. Recovery probabilities are assumed to be lower for Case 3 than Case 1. 3 6. Overhead cranes had stops to help prevent heavy loads from being moved over the spent fuel pool. Cranes are operated by non-nuclear trained operators. 7. The spent fuel pool water clarity is not maintained. Control room cannot monitor the spent fuel poollevel directly. 8. There is no requirement that fire protection equipment for the spent fuel pool area be j maintained operable. Therefore, no credit is taken for the fire pump's ability to provide - inventory makeup or fire suppression. J ( A1-2 I l i

1 f .1 i Assumptions' independent of Case One Two, or Three: ' 1. An F4 to F5 tomado would be required if significant damage were to be possible to a ] PWR or BWR spent fuel pool. 2.~ An F2 to F5 tomado would be req @@Tsible significant damage to a spent fuel pool support system. f 3. - Shipping cask handling is the ' dominant heavy load operation. e" wowed over the spent fuel poolwith Spent fuel casks will be the only i YgNpod ' 4. sufficient mass to significantly dam 5. ' Crane operators will follow safe load path procedures when moving heavy loads near the spent fuel pool. Spent fuel pools are robust and wihu s smic events less than three times the safe 6. shutdown earthquake (SSE). 7. The staff used generic loss of offsite frequencies. I 8. Pumps or valves are manually operated and aligned. Every action must be accomplished by a certified fuel handler. g. The effects of criticality were not considered in the risk evaluation. Its potential for impact on risk is considered to be very low. 10. One year after the last of the reactor fuel is transferred to the spent fuel pool there is no longer any day-to-day NRC onsite oversight. We did not attempt to quantify the effect of this assumption. '11. Extemal flooding is assumed not to be a significant contributor to loss of spent fuel pool cooling. 12. The utility has removed the emergency diesel generators and other support systems such as residual heat removal and service water that could provide spent fuel pool cooling or makeup prior to the plant being decommissioned. 13. The highest action level of offsite waming for decommissioned plants is an alert. 14. The spent fuel pool cooling system (sied mounted) and any support systems all run off . the same electrical bus. 15. The utility has procedures for small leaks from the spent fuel pool or for loss of spent fuel pool cooling system. A1-3

16. The only significant technical specification applicable to spent fuel pools is the requirement for radiation monitors to be operable when fuel is being moved. 17. There are no large pipes that go into the spent fuel pool that either extend more than a few feet down into the pool or that penetrate the spent fuel pool at a level near the top of the fuel. DRAFT I 18. The evacuation is assumed to begin three hours prior to the release of fission products I to the environment. 19. Ninety-five percent of the populatiggg0 piles of the site is assumed to evacuate at a speed of four miles per hour, ] l DRAFT a s A1-4

' " ~

pehmir, F wel Puel O

C.eairq T~~-- -r-HowWevy g y Evess head kiloqFay of Systems Caamg unsg Omese g o Deusser TIten Coadre Penskan Unmg er Onsas 0 4 3 Tenes sag Sm Onsee h h e SEQUENCE DESCRIPTOR e FREQUENCY S4-45) 0 ECai FPt Cm car OLS OFB e 801 EQE ON f 802 EQECSF OK 116E403 - -. -.. - e Sol EQECSFOCS. OK - - _ _. _. ~.... - - -~ 3 00E #)2 - (E S(H EQECSFOCSOFB 6 00E+000 'g - ~ ~ - SOE EQECSI ON 100E+000 . ~ - - - EC Sol EQECSIOFB 0 60E 007 I,.-._ _ _ - - - ~... - - -. - - - -- $07 EQEFPI 100E 006 C uMnNUPRA6PP4WQE EVf 09 0F 28 OLM99 Wl4NUPRA 10 OuenWinaten Date 96/15,99 OS 06 0' TOTA CMF=194 406 SFP HilnM AI DECOMM4StuNED P. ANTS t SEl8Mic lNITIATING EVENT CASE 1 i I i.

i 1 l P ruusrum _ aseense soung suessery e - Buset hama pumpey of eguesem Cesag.- (Jamig OHAe E D Fuisuun 4Juss FREQUENCY .C,eu.te. e OEu.b.e SEM N O.~ ~ . Guesser 11isa -.E m.4 am pgE PPI cel ceF Oca OPS I 801 EQE OK f r l' SIM EQEcsF OK 1 I ( 4 00E-002 OCSL40C 0 t 903 EQECSFOCS OK t 170E@t DFDL. LOC I. SEM EQECSFOCSOF8 0 00E*000 kE C 6 f l i i l 80! EQECSI OK 2 00EM __ IE EQE t s 1 100E+000 1)NITY ' ' EE SOE EQEC830FB 9 50E 007 kgQ. 1

6 2

801 EQEFPI 100E-006 - C WWnNUPRAWFP NQE EVT Os to be 09/2549 W14NUPRA10 $FP flik AT DECOndhetSND %T6 Quanuttenen Dale 960040 0614 31 TOTAL CafF e 19tE408 SEl8AdlC INITMTING EVENT CASE 2'

- l l

k. l FuerPWM.

WWh'I Genug, vecesung g tange'm S4ERI tunsi r of Spelene Ceedng Weeg OMes E O - OmslerTaten Comme Funemen,' Onsee $mgess soumes SEQUGNCE DESCRIPTOR 9 PflEQUENCY useg er Onnes 0 4 3 Temas gee Systems S His EGE 7pe cm C0F OCs OPS 901 EQE OK 802 EQECSF OK 120E40t CCSGLOC 3 03 EQEC8FOCS OK 1 4 3 00E402 'DFDL LOC ~~ l' j'. . hg 30d EQECSFOC80FB ' 0 00E*000 4 f f' 4. .i' t i 2 00E-005 E-EQE. God EQECSI OK 100E*000 . IJNITY." kE b 80s EQEC8IOFB 9 50E407 e gyp.g 001 EQEFPI t 00lE406 C-WienNUMtA4FP-32OE EVT 08 90 22 05/28 MIS UnN.6AJPftA 19 Qeh Dese OGGenie 07 B0 M TOTAL Car e i esE40s $FP RISK AI Ul:COI4116SK)NED Pl. ANTS SEI6MIC HTETING EVENT CASE 3 1, _

Y C. My g () FID EM FWD,.... Recmery Uemp E D m the er No Eliset SEQUENCE DESCRIPTDR FREQUENCY q .~ Pf l 00P OMK OFD E $01 FIR OK i eocc.cos IEfiR $02 FIROSP OK 6 00G.001 05PflR / S03 FIROSPOMK OK 1 i.oc-*, OMK.DGFP f i $04 FIROSPOMKOFD 3 558 007 PE F------- 1 i I l l l C enknNUPRAGFP-1flR EVT 10 3518 06n0M WIN NUPRA 10 sep n,34 ar ogcouwiggougio peaurg Quenidulen Date cenHS 10 34 46 TOTAL CMF

• 8 566407 INTERNAL PIRE INITIATING EVENT CASE 1 1

m m.

FM Em rm... opereur necovery p mihe er W Eesa Recmsey IM E D Aun W en SFP Punchen Uusie ONn4e Q S Reaster SMS Diomed Sources e SEQUENCE DESCRIPTOR e FREQUENCY Fee Pumpo FR OSP OME OFD $01 FlR OK 4 ODE 402 IE. FIR ' 4 I S03 FIROSP OK 100f 402 DSP#1R 1 l Sol FIROSPOMK OK I l 2 80E401 4 OMK DGFP ~ I i ) [g 804 FIROSPOMKOFD 104E406 l C WNUPRAGFPWR EVT 1108 44 Olv10/99 WlENVPRA 10 4W P RISM AT DECOMM688tONtD PLANTS Quanuhcahon Dee 08/10/96 10 55 41 TOTAL CMF e 104E406 INTERNAL PIRE INITIATING EVENT CASE 2 L a.

p., p mesmy y i.^ Fue sums Pee.. - 8tMemy UssG E O mas s een BAsse ~ 9E7JDfCE DESCatIPTOft PMQUDICY ~Pe put CAP Chm CPD S01 FIR ON .,I' I L l 1-s cos.cos E.FR ' ' 4 '. FIROSP OK t 00E-001 13SP-FM 903 FIROSPORAK OK 100E*000 UNrfY - SO4 FIROSPOMKOFD 9 00E 005 h00E C RMhNUPRAGFP.3fm EVT 1103 33 0E#1990 N 10 " SFP RiaK AT DECOIAIA164K)NED PWTS QuenemoshonDees 98r1940110302 TOTAL CAAF = 9 00Em ~ plTERNAL FIRE JPFilATING EVENT ' CASE 3

~ t WBW Umm w pumum Castg Amam '- huessumesof 2 Roamsey at tueuses41shaus - - susewery unas a e Osmes Geween i p Ammas ' taas er Coman, e,seem - . unasFee o a J . Ceamp pungs. 8 SEQUENCE DESCRPTOR S PREQUENCY LOC CRA

1. C 0C5 OFD Ops 301 LOC OK j

r-i l 8 01 LOCOCS OK t 4 't 00E40s i OCOE40L 8 03 LOCOC80F0 OK 4 2 00E402 OFDE40C W 80d LOCOCSOFDOF8 4 Got 011 OC E f 801 LOCCRA OK t SOE LOCCRAOCS OK + 2 20E404 DCSL40C y'h*g 801 LOCCRAOCSOFO OK 3006402 OFDL40C (E f -- 80s LOCCRAOC80FDOF8 0 00E*000 SOE LOCCRAIND t 50E 007 }. C unanNUPRA4FP-ILOC EVT 103729 05/1045 WW.NUPRA 1 $ SFP RISK AT Ot!COM0dtS&KJNED PLANTS l Queemoshen Dele 05s104510 34 el TOT AL CasF e 1 $0E 007 i LOS8 OF COOUNG CASE 1

a L

cases, namm edussens of naseury se masse assneue,

puesunry uuRB g p use er emme user usume opmust Ossas neuese e O Fnsoussey secumect DescarTOn u,C ca. OC. OFO OF. 301 LOC OK w 4 Sol LOCOCS OK t I 6 50E403 'OCEE40C 803 LOCOCSOFO OK 9 00E402 OFDE40c hg# EQg 80s LOCOCSOFDOFB $ 40E408 k"[ t 8 00 LOCCRA OK t i i h' 80s LOCCRAOCS OK 1 910E402 OCSL40C h'M- - 801 LOCCRAOCSOFD OK 170E-001 OFDL40C' kg 801 LOCCRAOC80FDOFB 2 32E.000 800 LOCCRAIND 150E 007 .. ) C 1AmWeJPnAdPP.2 TOC EVT 11 et 33 cst 19e9 MNUPRA 10 app migy. At DECOMMtSSONED PLANTS . A% Dese cartsee 10 I4 48 TOTAL CMF

• 157E 007 LOSS OF COOLING CASE 2

E N N RsWWWW q g p h W. ABWoWV W hh Om Bennes E D Aamme Laos W cesans esmem unm P o e

commq,

_ p.e,,,oe.. e . esaussace caecArrost e ressouesscv loc CsIA de Oc8 OPD GPS 3 01 LOC Ox v. I - ses Lococs-on i. s uas s [ 'OC8E-LOL b 4 { SO: LOCOCSOFD OK I I k , NM .m. h$ y L ] 4 s - see LocenAOcs ox ,20E 00, OCsL LOC ~ }gf[g ~ to? LocCAAccsOFD OK WTY hECO sol LOCCRAOC80FDOFB 3 60E 007 )

== m._, .. 0, L = re:: :,,an%~::,mia;t" "'~~~ ~ -o~' LOSS OF COQLeaQ CASES l I

CeeMRl NBnIIsl thabsup fluem hetissimas of tsungsgry Legi ~ - -m ingales lemages Mahoup Olhes Gources g O O E inumatury Casamey ens Alame Imentory And inmunes Maheup Unsis Fe. s SEQUENCE DESCRIPTOR FREQUENCY Lees usag W esegman N. 41.steup puu. R M u. LOI 94 1 CRA

1. dD OIS OIL Otm OFD 801 LO OK i

l i-802 LOOL $ OK i 10f mi ~OIS-E ~ 503 LOIOtSOMK OK , -.2 i Out t

,=g;, _.

tOIOSOMKO,0 . 0E., i l SOE LOCRA OK I 4 SOE LOICRAOIS OK I 150E 001 i 015 4 ' ~1 j i g $01 LOICRAOISOMK OK 6 00E402 OMK4 RE Sod LOCRADISOMKOFD 2,eE-000 100E402 IE401 I SOE LOICRAIND 3 01E40, gL Sit LONLL OK } I. $11 LOINLLOL OK J I 102E40.1 - $12 LOINLLOILOMK OK 120Em2 DMK401 ' 'Lg ggf E $11 LONLLOILOMMOFD 181E 008 Sid LONLLCRA OK l { 811 LONLLCRAOL OK 102E401 OIMU ' Sit LOINLLCRAOILOMK OK j g-120E 002 OMK401 ~ ~ hE - Sl1 LONLLCRAOILOMKOFO 7 0sE409 I 4 -~ _ S,. LoaC_O .,,E.00, MLOis 1 6FP nisx AT DecoMuessioNED PLANT 3 Ouenemoshen EYN0617$1 70NCMFNaeE.00s LOS8 OF COOLANT INVENTORY CASE 1 l ]

r ,t i.exe p &ame unusuun . sunsu user ammust ummme apesmer Usumer necesary sum a p coment Manus tankee femem heemanns of inventer Laen menise Muses temhew Onnes sewass e O huumassy Capesty sus Amene. tousnery And memes hAmheep Uutig Pte O S isammen h W WW auu Uemy. Pumpe. e OfGUENCE DEacteTOft e PIW3UENCY 43 Rggmed Mahog Pumpe Magup punge LOI esLL CetA IND Oss oil Oam OPO I' f I 802 LOIOl8 OK 19 90E401 OSE' j 503 LOIOSOedK OK t [ e 00Em2 DMK-E 1 '%H-804 LOOlsORMOFD 3 00E406 i f I i l i 801 LOICNA OK -w ) I I SOE LOCRAOl8 OK I F ~ i I I { t 2 00E401 I l DSL^ t t .l h y" 803 LOCRAOl60MK OK 110E 001 ~0MK.L ~ ' ~ h"[- $0E LOICRAOt80MKOFO 2 73E407 100E402 IE-LO " 4-f h g' SOE LOICRAINO $ 01E407 f $1C LONLL OK i i $11 LONLLOL OK I 102E-001 OMO '" i t l l 5 12 LONLLOILOMK OK '120E402 CMKLO ' *p_ g=_ g .u LO.Nu OtOM,eFO im 8 14 LONLLCRA OK $11 LONLLCRAOlt OK 102E-001 OMU~' 1 403 $14 LONLLCRAOtLOMK OK g 120Em2 OMK-LOI ~ . (g"g ~ $11 LONLLCRADILCMKOFO 1.tt.007 ,1g g 1 to NaC=~o . m., gwenssVPRA QOf 08 80f IO 8FP RISK AT DEcomands&SK.84.D Puuvis 08 p LOSS OF GOOLANT B4VENTORY CASE 2 i i i i I t I j

asse er asse massess rmans unser apesum usmus assieun opermer --. wem p busneery caussey see Annus ensmussy ' inumumy Lass anamnes hemies Masse Omas semess g o casame samuun unhow fasom enemmens af O e umme,Fes And hismos unheup e sEOwasCE DESCmPTOR s PetEOUENCY Pun s annumen esweet Lass testsw unu .onna h Itammed tashaup Pungs ushnig Punge LO 8at CRA esD 088 Ou, Ohm OFD 801 LOI OK - 502 LOICIS OK 4 i 2 00E401 OfS-t' t Sol LOIOt80MK OK i 100E+000 f WW ~9 kE T-- sod LDIOiSOMKOFD t $7E406 t. -{ I i-SOE LOeCRA OK I 1 i I f sa LOiCRAQs OK 4 00E401~ OlS.L l 106E 001 601 LOCRAOISOWK OK 'CR4 ARM ". 100E+000_s g

1. d -

808 LOICRAOl80MKOFD 4 03E.006 pg y2~ 100E 2 00E401 800 LOICRAIND 120E 005 IND.LOL i i f $1C LOINLL OK I I r-- $11 LOtNLLOL OK i 110E401 f OMU ~ ~ p $12 LOINLLOLOMK OK "yW- , m., ..1 811 LOINLLOILOMKOFD 4 83E 005 pggggg -- LO.6ML i 8 14 LONLLCRA OK $15 LONLLCRAOL OK 110E-001 OIMU 105E401 S18 LONLLCRAOILOMK OK CA.ALARW ~ 100E+000 g ~.,- $11 (ONLLCRAOfLOMKOFD s 1eE of.6 RE SS -- ,.2 - $1E LONLLCRAIND 4 946 005 IND.LOIS C uMnNuPRAeFP-3tOi EVT es as 44 05/2see0 IMN NUPRA 1.0 SFP R66K Al DECOMidlE840NED PLANTS Quanhecshun Date 05i2840 Os De 14 TOTAL CMP = 129E404 LOSS OF COOLANT INVENTORY . CASE $

Laus er - -w mm -g w mm Puuur Fm8 Poe Pune 8lessary avstem nuesney usag Oense seween E D Peue Csaised. . tien and Run pner to. Regens teste uP Bretam 0 8 . and OneItsaged ' Pusi Unneery and fhe p SEQUE8dCE DE8CRIPTOR S PflEQUENCY Suunis LP1 DG OPR OCS OtAK OPD / p 301 LP1 OK ? 9 02 LP10C8 Cu 4 44E403 ' t DCSE LOF 901 LP10CSOMK OK j 6 i f2 00E 002 'OMK#PS s I 804 LP10CSOMKOFD 144E407 E t SOE LP10PR On 100E-003 REC 0891 R0s LP10PROMK OK 130E 001 OMK-DOFP f 801 LPtOPHOMMOFD 0 00E+000 g g f 800 LP1DG OK I I' r i bOI LPIDOOCS OK I 4 44E 003 OCSE40P $1E LP1DOOC80MK OK 190E 401 7 00E 002 MSTART OMK-EPFP (E Sit LP1DOOCSOMKOFD s ees-007 812 LPtDGOPR OK 100E-003 RECOSPf (E 513 LP1000PROFD 7 60E407 C WM4UPRAsFP-itP1 EV1 08 02 26 O!VJWet WNUPRA t o EPP RaSK Al DECOMMa$$aONED PLANTS Queekahon Does DW2emt 08 9119 TOTAL CMF

• 127E406 LOS6 OF OFFSITE POWER FROM PLANT CENTERED EVENT 4 i

CASE 1 I I

uma e oness oisem nw umme Power Gesang petor 1 eWn g y Power From Fee Pump Recovery Syesem Rocawry Useig Offsee Sawcre E D Piers Cornemd Sierl and Run Pnor le R>atest Makeep System O M MMOft FMMW and 0,18 Reemed Peeluncewery ens Rei Evems LP1 DG OPR CC8 OMK OFD 801 LP1 OK i l f i I $02 LP10CS OK f 7 $4E-003 OCSE-LOP' 9 i $02 LP10CSOMK OK s + 0 00E402 OMK FPS I I kE . 604 LP10CSOMKOFD 2 35E407 b t i 805 LP10PR OK 100E-003 REC.OSPt SOE LP10PROMK OK 2 80E 001 OMK DOFP ~~~ 8 801 LP10PROMKOFO 0 00E.000 g J. RE k i i I. SOE LP1DQ OK g l SOE LP1000C8 OK 7 94E403 OCSE LOP Stt LP1DOOC80MK OK 2 60E 401 140E401 (6 START OMK-EPFP S11 L1 WD 2 88Em REC-O St2 LP1DOOPR OK 100E 003 RE406Pi kg - - - - - 813 LP1DOOPROFD 104E406 j C WenNUPRA6FP-2tP1 EVT 95 54 00 0$f28%9 W148WUPRA 10 SFP R+SK AT DECOMMISSIONED PVWTS Quermacehon Date Ok2820 08 62 48 TOTAL CMF e 416E 000 LO68 OF OFF8ITE POWER FROM PLANT CENTERED EVENTS CASE 2

f;' ' l '. lT ems = - cameruns g

games,

,,,,n M%- , READumfy Syman Rempumpy LNMg Om soufens l h t t l Phe Commed - Sist and phen Pner to feweert taansgotyman. O a 888 GAB IIIAMEW Fusi unseaufy and phas e OEQUE8eCE DESCRIPTOR e PREQUENCY - p Guseas LP1 DG OPR Oca ORIK OFo l I l L 901 LP1 OK l-g i l } I In { Sol LP10CS OK OP~ *, ' l. ( i 8 03 LP10CSObsK OK i l I t '2 00E 002 9 Om r hE 80d LP10CSOAdKOFD 0 00E*000 \\ li l l t I soi LPiOPR OK 'l ', m 00s REC 48P1 SGI LP10PROtfK OK 190E-001 ' l OtsK-DOFP it Lp1 - pg" 301 LP10PROMKOFO 900E 000 I SGI LP1DO OK f + i. SDI LP1DOOCS OK i 1 i -{ 100E 002 DCSE LOP i l SIC LP1000CSotfK OK 100E+000 100E*000 DNfTY T DNITY ' h 811 LPIDOOCSOMMOFD 7 57E-006 I l l 812 LP1DOOPR OK 100E403 RE*M1 $13 LP1DOOPROFD 4 00E 00b C WWmMUPRAGFP-3tP1 CVT 05 11 10 OSQStat W NUPRA10 SFP RISA A1 OECOedMISSIONEiD PLANT 4 QemDme OtGMD 06 10.23 TOTAL Char e 7 e7E406 LOSS OF OFFSITE POWER PROM PLANT CENTERED EVENTS CASE 8

-,,h 6 - P - a, Puser Pius Pee PW Aeusemey Symem . Amassery desig - OWese teumos E D SEQUENCE DESCfMPTOR FREQUENCY - [ ,p 'gy LP2 DO OPR OCS CedK OPD 801 LP2 OK l I t l 303 LP20CS OK !Y 4 44E 403 i 'OCSE40P 4 $03 LP20CsoteK OK i I 2 00E-002 f OMK-FPS l F 1 f i I 80d LP20CSOMKOFD 12eE408 hgg j t 1 808 LP20PR OK f. i ' 2 OPE 402 REC-08P1 SOE LP2OPROWK OK t 190E-001 WK DOFP ' _l Sol LP2OPROMKOFO 0 00E6 - hgg ki i . SGI LP2DO. OK 1 s SOE LP2DOOCS OK I y 4 446 003 OCSE40P 61C LP2OOOC80MK ' OK 7 00f 402 t 90E40f DCLSTART 'OWicEPFP RE $11 LP2DOOCSOMMOFD S 23E408 .u LnoOOPR OK 2 00E-002 REC 06P2 kg $11 LP2OOOPROFD 1.33E-006 C wwiNUPRAerkit.P2 fvf es Os 12 Oli2Me MNUPRA 10 SFP R68K AT DECQtfMi&&lONED PWTS WMDuse 05(259B 96 0127 TOTAL CMP *1 372 000 LOSS OF OFFSTTE POWER e FROM SEVERE WEATHER EVENTS . CASE 1 'I.-,.a2 i,,.

e uum, - uimum a gause - g gymmu me,yun PEIEEF IMR . caut ans Run - Resumy 8venant Resuuup % - Oh Seesses E O EE8 % esusso Pnst to t Wusher . Fussuneevery. Ameen - tammmipagenom O a . ane Run a SEQUENDE DESCfuPTOR e F40UEffCV LPS OG OPR OCs Onst OFD 3 01 LP2 OK g l I i l s 803 LP2OCS OK l j +- I 7 tet.003 .I 'OCSE LOP $0a LP20C80MK OK 9 00E402 OMK-PPS v' I 804 LP20CSOMMOFD 135E4eg REC.068 .i ? I i 8 - 001 1.720PR OK t 6 i i '100E-001 i REC 08P1 l 801 LP2OPROteK OK l 4 2 80E401 OMK.DOFP 801 LP20 PROT #KOFD 0 00E*000 LP2 REC 05 'i I I. f-Sul LP2DO OK f 1 4 80s LP2DOOCs OK I' 1' 7 64E.003 cCst. LOP $1C LP2DOOCSOMK OK 2 60E401 140E.001 DO. START . OMK-EPFP kE $11 LP2DOOCSOMKOFD 227E 007 4 812 LP2DOOPR OK 100E 001 ' REC.08PT $11 LP2DOOPROFD 910E406 g C eMnNUPRA4FP.2tP2 EVT 07 34 41 0S'2619 MNUPRA 10 SFP RISK Al DkCOMMISSIONt;D PLANTS Quenciacamen Dees. OW2690 07.M Os TOTAL CMF = 9 35E40s LOSS OF OFFBITE POINER FROM SEVERE INEATHER EWNTS - CASE 2

Laus er - unsur us puun Feuer Geset Opereuw , som g p Pasar Peene Pee Pump fasesumey system. Reesvery Umsg Omase cources g O 8 esause Sean sad Run Puer as benen Mans, sycamen SEM DENM FREQUEDICY 1Almagist Fuss unassesy and foun tuenes LP2 DD OPR OC8 Cam OFD 801 LP2 OK t i i 802 LP20CS OK 180E-002 DCSE-LOP i 301 LP20CSOMK OK { 2 00E 002 O E FPS (E 804 LP20C80MKOFD 0 00E+000 't see LP20PR OK 2 00E402 REC-08P2 SOE LP20PROMK OK 190E-001 OMK-DGFP h"W kEM - 801 LP20PROMKOFD 0 00E*000 SOE LP2DO OK ( 500 LP20GOCS OK 199E402 OCSE-LOP Ric LP2DOOCSOMK OK 100E+000 100E+000 UNITY' W YY kEM Sti LP2DOOCSOWKOFD e m 006 8 12 LP2DGOPR - OK 1 2 00E 402 REC-OSP2 kg 813 LP2DGOPROFD 7 00E-006 - C MMeduPRA4FP 3tP2 EVT US 12 61 0478999 WIM6 *IUPRA 1 g BFP RISK AT DEconIMsS5 TONED PLANTS Quensecahon Does' 05/2849 081217 TOTAL CMP = 130E-006 LOS8 OF OFFBITE POWER FROM SEVERE WEATHER EVENTS CASE S -

) U 1 i z. i a 3 4 i 5 i, FAILURE OF SFP COOufKl SYSTEM 0. TO RE-START AND RUN @ MCClCreehon Dete: 05-19-99 Revision: 05-27-99 ~ eHec1GC IMee.NUPnA10 l -Tsai n.- - - -. _ { l 1 l' -e 1 OCaftt12 l^1 1 IDNEMrfMINIIO FeselPeelCeeARg l MST4MTAIE AuGN 's Pwnps Fat f.- sversu. sAnax 2 1 l_ HEPMO,40P4 GCSRm am() w y,,,, .ji pg,;,;qq y, - "- y. g,,,,, - Pwe rennes ones as ranws. i vmes.cce l3 I' ocanno ePcav.cce* _oc.aniss L I,\\ s as aus ( :,+ k^,t l s v n r-- f.( sFP MIeb g conensa c.ause a Pwnpa Famas 4 i ?. 4 i ePC.PasPa:P ocenisi j s so.mo i : l ["g ~ ,5 --,,-,g-Fans to seen ene Fass I. ansi i, Run Run j sPCMMP 1 sPC pup MF s j senees f..s i s eem t : i s w' L a 'l Hom &Asenger sFP' Heel sFP " Heal HUlAsenger j case ennenser casm, I omawee.ccr E - i w 6 ccP van syman ress 4 i ePcerano arc.cxv ccam f aPcerxMn g i sPccra.ccP } I t oo cos ( : su (J si i som ( s' sw is ) - J !7 1 M 8. I i r9 I f +-- .'A B L-r C. i l I

P i z i a 4 6 5 U 8 1 FAILURE TO RES5D TO 'A CONTROL ROOM ALARM 1 o Analyst MCClCreehon Dele: 05-15-99 Reveon:05-18-99 ) cr assmLLOC WWHEUPetA 14 l U T I AIN 'l t GCftA112 1 c.

• g

? . [~-- -a, r-w ,e, s ,e i Aasme Dw ggaeney Laas at inmassen GPC4VL40F SPC4VL40P { HEP-fles.ALAnM 1 l son.ms ( ; ice ons (: ) son.ma is ; s v (- l I i .83 l {. r 3 >-l4 i 4 5-i l l 6 i 8 .j! i. f b 'I t i -8 4 t sw= 9 A. . L. . B' I + C' l

c.- o i 1 i z i 3' a i gr-i JAILURE TO MITIG'AIE i I LCR,1 O-LOSS OF COOUNG EVENT ,i 1 i. Aneyst MCCl Creation Dete. 05-18-99 Revisson: 05-27-99 j l IMwec. loc WIN 4duPftA 91$ Lee 1 Swt.s a ^ l Sany ) l GLCR112 ,I' I y dst ~ .~~7eGan, e, user' " #aded Res.e.w,s,"' l i C Tiwe n me i l j2 j oLCR131. GLCR133 i L {&l 'n, roPauvoieus =~ 'ressi. i remune einanauP~ nasTons CoouNo Aswnese uswsNtomy or una or SV9 f5AA.SAALY Casang.5eny PutE PROTECTION PRIP r 3 i i _ _ _.. - -FP4snur#TF r t Marcoo64cC4 i HEP 4LTCL4 1.80e498 ( 8 _) 1.00e402 ( 3 - L80s402 ( 3 ) j r 6 1 4 f-Evt.. a t C Laas !5 oLCas.s: r i ~Pedt5 Rech~ ~Fasine to Inscover- ^ j Ceesug Sysesen Through bisheup Bretsam 6 otCani oLCaiu if; !\\ r ~ -oesnafon ras To '. ~ rail'uise To isanaupl oesuton pas To REStoRf C.oousso ESTAsuSM AL. TERNATE ; l thnftNTURY SY uSE op SYSTSAI LATE CoouNo LATE FiftE PROTECTloN PRAP, f j 7' j aiEooL4aC4 ~ b rPWWTF ' - nap 4LTCLA se.4ee fe ' * .se.4e [s ' too.4en (s') ~ g 4 1 8 i 9 i ~ ,A-W B r .C f

O i Y. i 3 i. a 4 r 5 ' ALLURE TO MAKE UP INVENTChY LIR,1

LOSS OF INVENTORY EVENTS o

'@. MCCl Creation Date: 05-19-99 Revieson: 05-27-99 am c. toc wue4suPna e o ~~WFaas to Emm.e Loi am s Esny l l: ,1. W oLait: i.- .t f, i 1J m i ) "F.inisie Esem ~Fesisia.ssG.e l em Fne P.c er 1 L ensmes= t p-i memwnomee i omas se ,2 p' l oumiss ouni:1 j ff F.} Ir ~6MFesiin r@'eFP e4] r -g,,,,57 ~F4fL'unE16'esaxEuP] - l wusw tem. t i pavtNTony ev ueE oF mesm Lane. syms. Fem meme assa.uP - ! Ft.se PnoTEctioNPesP insum asase t l 3 eeP e m w a i sFP nEoenKuP4 _i i nEP e asKuP4 FP.asKuPitF i L1osema(v) on soi i y iso.com ( s j ione.com ( : j L v v osnaksimie asses = Loi i Evans Law i f4 -J l j G'* 54: i / i 4 i "env Fne Pwnee er Fenwe se u. nee ( ~ Vesw.mamme i Lossansme = omseSowcw Nomu nenee otmist ountas _

~,.

,'5 1 i r-- i l IosilainiF'e.ssis ne'liuis~r SFP W ~Opists Paes le ~ikiLORi td MAMEUP' psvEwToRV ev uaE oF j inesw Lee { inamm isense ' ; FwiE PaoTEctioN PesP ) mounte s. sysum Fass nom. asme } i 1 61 l.6 HEPM _MKUP4 s HEP DN.heKUP4 i FP nsKUP.FTP I SFP.NEoMKUPJ 8 8 1 s one.coi : : s oo.com ( : ) oo com ( : ) (:b) l j s comen v w v p f l f opWmir~isin to i i I hangste Smse Loi j evos II l oL#tti g ~' } '4 f f. 1 reswe n i.n.e tissais.Gos i

1. ww Fee Pwnpa er Loss and inpisse

', 8 ountas ofinde Gewsps Nome Meeup -l 1 t. oumist j f\\ l% ~ ' ~ 6P aliifn55- ~ ~ ~ReguisiSFit tset'euP-~ ~OMeid Fais ii'~ ~FAiLORE YO' MAKEUP inses nesteup for iNVENYonY Sv uSE oF [' ensame Makas.for Symem Fest Smas Lassen FIRE PROTECTION PtsP. 4 l Smse Lesee 9_. HEP-INV AnKUP.Sht. FP-tsKUP-FTF t . HEP.iNW4sKUP 8ML GFP-nEGMKUP# [ , s co>co3 s s ' l ooe-col a s ooe.co3 ' S ^ 1.coe4o:. ( A <:B- -i P* C' i

7-v% U .i T-1 a 1 3 4 1 3 ^ . FAILURE TO MAKE UP INVENTORY LPR. *l 0-LOSS OF POWER EVENTS @ MCCl Creation Dele: 0519 99 - Revision: 05-27-99 ~ h>8stLoc tuupaJPRA t 0 - Pseie~Eiiisseup 8MN18 ' EP DG b'T MMS e psuus 1 i j f GLPRt12 e i-F DPERATOIIt FAILS'70" I"INVE8f70ftY SV USE OPi ILM to tiasifuP" ESTAAueH ALTERIeATE ; i,. T, GOCUNO.EARLY PIRE MIOTECT100e PRIP - 62 PP4stuPfTF 4 MP-ALTCL4P4 i 100e4ee j s } t 00 002 g 3 i s.+ v 6 .t I 3 i Faswo to maheup i N asung DG he Pung g {4 I GLPR142 I i b' OPEMf05 FAU ' d' DM Fue T i ESTAguaH AL.TEfteATE Presocemm Pump COOUNG EARLY Paan to Funchon -s 5 MEPALTCL4P4 FP OGPunaPfTF { 4 100>002 {J 100e401 ; 3 S ~. l 'l i )6 I , j i a I, ~Fasie is maneup ( 4 .WM9 ', 7 EP he pump oLPRin i i \\ - L IN, s .~0PERSTodil #AiLi T6'

• Emdens" Fee

- ESTASLISH AL.TEEINATE Pretaden Pump COOLeeG EARLv. i Paan le Funcson

g HEP 4. L_TCL4P E FP4LPuedpfTF 190e402 : 3 6 00e-002. $

m i 9 t 1 i A r B i i l i

C i

a

Appendix 3 Assumptions for Spent Fuel Heatup Estimates The working group calculated the time to heat a single fuel rod from 30 *C to 900 *C without considering any heat losses at various times following permanent shutdown. This is a very conservative, non-realistic type of calculag Giili'some cooling through conduction, convention or radiation would occur and increase the ' time to heatup. For the calculations, the working group used a decay heat per assembly and divided it equally among the pins. - Design values for a 9X9 assembly for the PWRs and a 17x17 assembly for the BWRs were used as representative asserat y heats were computed using an -5625. The decay heat in extrapolation of the decay power tables in# NUREG/CR-5625 is based on ORIGEN computer code calculations. The tables for the decay heat extend to bumups of 50 GWDMTU for PWRs and 45 GWDMTU for BWRs. It is recognized that the decay heat is only valid for values up to the maximum values in the tables, but the functional dependence of the deca power with respect to bumup for values in the table indicate that extrapolation may provide a e r%feTstimate of the decay heat for bumup values beyond the limits of the tables. The BWR decay heat was calculated using a specific power of 26.178 MWMTU. The PWR decay heat was calculated using a specific power of 37.482 MWMTU. Both the PWR and BWR decay heats were calculated for a bumup of 60 GWDMTU and include an uncertainty factor of 6 percent. PWR BWR Assumptions:. 60000 60000 burnup/ assembly (MWD /MTU) MTU/ assembly 0.47 0.17 assemblies / core 193 764 rods / assembly 264 72 pellet diameter (m) 0.0061916. 0.00906 inside clad diameter (m) 0.0063276 0.00926 outside clad diameter (m) 0.0094996 0.01077 fuellength (m) 3.66 3.6 rod peaking factor 1.166 1.166 axial peaking factor 1.2 1.2 fraction theoretical UO2 0.96 0.96 Years Decay Power per assembly 1 15.5561572 12.54475868 in kW/MTU 1.5 11.4163371055 9.33601650748 2 8.85047 7.345154884 2.5 7.252431291 6.12127658952 3 6.02689457463 5.21793956972 3.5 5.11496410863 4.60129503226 -4 4.341018 4.05752418 4.5 3.68417781163 3.57801496228 i 1 I L 1 A3-1 l

T' ' Decay power in watts for one - assembly: (kW/MTU)*MTU*1000 Years Decay Power *0.47*E3 Decay Power *0.17'E3 1n kwer(W) Decay Power (W) 1.5 U 11.393884 2132.6089756 2 5365.678439585 1587.1228062716 2.5 4159.7209 1248.67633028 3 3408.64270677 1040.6170202184 3.5 2832.6404500761 887.0497268524 4 . 3+310561 782.2201554842 4.5 0.27846 689.7791106 Calculate adiabatic heatup time from 30 C to 900 C in hours for one fuel rod: Years ggg(hours) Heatup time (hours) 1.5 6.32 7.7 2 8.15 9.79 2.5 9.94 11.75 3 11.96 13.78 3.5 14.1 15.64 4 16.62 17.72 4.5 19.57 20.11 i A3-2

Appendix 4 Structural Integrity of Spent Fuel Pool Structures Subject to Tomadoes and High Winds Tomado or high winds damage, resulting from missile neration, can affect the structural integrity of the spent fuel pool or affect the availability oM rt systems, such as power supplies, cooling pumps, heat exchangers and water makeup sources, and may also affect recovery actions. A - set of site specific evaluations for tomados and high winds was documented in NUREG/CR-5042, " Evaluation of Extemal Hazards to Nuclear Power Plants in the United States," Lawrence Livermore National Laboratory, December 1987. It is noted that this study was performed to assess core damage frequencies at operating plants. In 944,"Tomado Damage Risk Assessment," Brookhaven National Laboratory, Septembei m thodology and assessment of tomado risk was developed. The National Climatic Data Center'(NCDC) in Asheville, N.C., keeps weather records for the U.S. for the period 1950 to 1995 (Ref: http://www.n a C. These data are reported as the annual average number of (all) tomadoes per 10, Tnies per state, and the annual average number of strong-violent (F2 to F5) tomadoes per square mile per state, as shown in Figures A4-1 and A4-2. Figure A4-1 Figure A4-2 Anray Annual Rr bar dStrangNicient (P2-PS) Annual AverageNunbar orTanadcarper Tanadoes per 10,000 Sgnare Mies by State 105000 Sgure Mies by State. 1950 1995 " 33 h 88 ' 83 04 ^Y g4 -l25 }24 T 28 Y"u 2 %g sp 9 o H n .3 n a to % i} [ n" u is a n n a" at nu u ,g n w,"""" u l A comparison of the site-specific evaluations (from NUREG/CR-5042) and general regional values from the NCDC database is presented in Table A4-1. The NCDC data was reviewed and a range of frequencies per square mile per year was developed based on the site location and neighboring state 4 (regional) data. ' In general, the comparison of the NUREG/CR-5042 tomado frequencies for all J tomadoes to the NCDC tomado frequencies for all reported tornadoes shows good agreement between the two sets of data. A4-1 )

i i Table A4 Tomado and High Wind Data Summary NUREG/CFgp& NCDC data ) seu u i, Frequency Frequency Tornado Tomado High wind Tomado 1950-1995 1950-1995 frequency strike damage ' damage - average for average for (per mi8 frequency frequency frequency FO-F5 (per F2-F5 (per Site - year) (per year) (per year) (per year). mi2-year) mi8-year) Indian Pt. 2 1.00x10d 1.00x104 Mh li.0x104 1.2-2.2x10d 0.2-0.7x104 Indian Pt. 3 : 1.00x10d 1.00x10d - 1.80x104 <1.0x1&' 1.2-2.2x1&d. 0.2-0.7x10d Limerick 1-2 1.13x10d 2.30x10d 9.00x10* <1.0x104 2.2-3.4x10d 0.7-1.3x10d ( <F1 ) - Millstone 3 1.87x104 1.87x10d GRAF Ti.0xiO4 2.8-3.4xi0d 0.2-i.1xi0d Oconee 3 2.50x10d 3.50x10-8 Low <1.0x104 2.8-3.4x10 d 0.7-0.9x10d 1 mirad. Seabrook 1-2 1.26x104 7.75x104 <3.89x104 2.06x10* 1.8-3.8x10d 0.4-1.1x10d LOSP & RWST Zion % 1.00x10-8 1.00x108 - N.A. <1.0x104 3.4-5.4x10d 1.2-2.0x104 4 GSI A-45 PRAs Regional w/o recovery of offsite power Local ANO 1 5.18x10d 1.53x10-8 5.69x104 2.53x10d 3.7-7.5x10d 1.7 2.4x10d 4.37x10d Point Beach 1-2 6.98x10d 5.38x10d. 1.00x104 5.00x104 3.4-4.7x10d 1.2-1.5x10d 4.11x10d Quad Cities 1-2 5.18x10d 1.04x10 8 <<1.0x10* 5.08x10# 3.4-5.4x10d. 1.2-2.0x10d 5.44x10d St. Lucie 1 6.S8x104 1.70x10d <<1.0x104 1.61x104 8.4x10d ' 1.2x104 1.20x108 Turkey Pt. 3 - 3.37x10d 1.70x10d 3.30x104 2.54x104 8.4x10d 1.2x10d 5.83x10-8 The Storm Prediction Center (SPC) raw data, for the period 1950 to 1995, has been used to develop a data base for this assessment. There have been about 121 F5, and 924 F4, tomadoes recorded between 1950 and 1995 (an additional four in the 1996 to 1998 period). It is estimated that about 30 percent of all reported tomadoes are in the F2 to F3 range and about 2.5 percent are in the F4 to F5 range. A4-2

DOE-STD-1020-94, " Natural Phenomena Hazards Design and Evaluation Criteria for Department of Energy Facilities," January 1996, Department of Energy, provides some insights into wind generated missiles. debris, a 2x4 inch timber plank weghing 1J lb}s viable threat, to acc For sites where tomadoes are notg IEf s is considered as a missile for straight winds and hurricanes. With a tacommended impact speed of 50 mph at a maximum height of 30 ft above ground, this missile would break annealed glass, perforate sheet metal siding and wood siding up to to 3/4-inch thick. For weak tomadoes, the timber missile horizontal speed is 100 mr i 4e a height of 100 ft above ground and a vertical speed of 70 mph.- A secor is considered: a 3 inch diameter steel pipe weighing 75 lbs with an impact velocity of 50 mph, effective to a height of 75 ft above ground and a vertical velocity of 35 mph. For the straight wind missile, an 8 inch CMU (concrete masonry unit) wall, single wythe (single layer) brick wall with stud wall, or a 4 inch concrete (reinforced) is considered uste to prevent penetration. For the - tomado missile, an 8 to 12 inch C@4f le wythe brick wall with stud wall and metal ties, or a 4 to 8 inch concrete (reinforced) slab is considered adequate to prevent penetration (depending on the missile). (Refer to DOE-STD-1020-94 for additional details.) For sites where tomadoes are considered a viable threat, to account for objects or debris, the samt 2x4 inch timber is considered but for heights above ground to 50 ft. The tomado missilos are (1) the 15 lbs, 2x4 inch timber with a horizontal speed of 150 mph effective up to 200 ft above ground, and a vertical speed of 100 mph; (2) the 3 inch . diameter,75 lbs steel pipe with a horizontal speed of 75 mph and a vertical sped of 50 mph effective up to 100 ft above ground; and (3) a 3,000 lbs automobile with ground speed up to 25 mph. For the straight wind missile, an 8 inch CMU wall, single wythe brick wall with stud wall, or a 4 inch concrete (reinforced) is considered adequate to prevent ' penetration. For the tomado missile, an 8 inch CMU reinforced wall, or a 4 to 10 inch concrete (reinforced) slab is considered adequate to prevent penetration (depending on the missile). - (Refer to DOE-STD-1020-94 for additional details.) The winds associated with hurricanes and other storms are generally less intense and lower in magnitude than those associated with tornadoes. Generally, high winds from wind storms and hurricanes are considered to be the controlling wind level at a higher frequency but at a lower magnitude. A4-3

l Recommended values for risk 4nfonned assessment of spent fuel pool The tomado strike probabilities for each F-Scale interval was determined from the SPC raw data on a state basis. For each F-Scale, the probability was obtained from the following equation, for the point strike probability: DRAFT i y<a>[ 1 P. = Equation 1 (. A. Yu P,, = strike probability for F-Scale <a>r = tomado area, mi8 2 A. = area of observation, m (state land area) Y, = interval over which observations were made, years ) In = sum of reported tomados in t e observation The tomado area, <a>r. was evaluated at the mid-point of the path-length and path-width intervals shown in Table A4-2. Table A4-2 Tomado characteristics F-Scale Damage Path-length Length Path-width Width and wind speed scale (miles) scale (yards) O Light Damage 0 < 1.0 0 < 18 yds. (40-72 mph) s 1 Moderate Damage 1 1.0 - 3.1 1 18 - 55 yds. (73-112 mph) 2 Significant Damage 2 3.2 - 9.9 2 ' 56 175 yds. (113-157 mph) 3 Severe Damage 3 10.0 -31.9 - 3 176 - 527 yds. (158-206 mph) 4 Devastating 4 32 - 99.9 4 528 - 1759 yds. Damage (207-260 mph) 5-Incredible Damage 5 100 > 5 1760 yds > (261-318 mph) The tomado area, <a>r, was then corrected using data from NUREG/CR-2944 (Table 6b, page 19 and Table 7b, page 21) to correct the area calculation for the variations in the tornado intensity along the path length and width. A4-4 }

r The contral U.S. region (12 states. Kansas, Oklahoma, lows, Missouri, Arkansas, Illinois, Indiana, Ohio, Kentucky, Tennessee, Mississippi and Alabama), was used to develop the tomado strike probabilities for each F-Scale interval; This region has the highest frequency of ) tomado occurrences and is about equal to the central region defined in NUREG/CR-2g44 (region A in Figure 2, page 12.) DRAFT The SPC raw data for each state was used to determine the F-Scale, path-length and path-width i characteristics of the reported tomadoes. Equation 1 was used for each state and the summation of the 12 states performed. The results are shown in Table A4-3. f% n A f***T* 1.s I mF i Table A4-3 Point Strike Probabilities F-scale Tomado strike pe er 0 1.2x10-8 1 . 3.5x10-5 2 4.0x10-5 3 2.5x104 4 7.7x104 5 6.3x10-7 S' ignificant pool damage ' An F4 to F5 tomado would be needed to consider damage to the spent fuel pool from a tornado missile. The likelihood of a tomado of this magnitude even striking the plant is estimated to be 7.7x10 per year, or lower. In addition, the spent fuel pool is a multiple-foot thick' reinforced 4 concrete structure housed within another (normally) foot thick reinforced concrete structure (i.e., the building housing a spent fuel pool normally is designed to mitigate an F4 or F5 tomado.). Based on the DOE-STD-1020-94 information, it is very unlikely that a tomado missile would be - able to penetrate the spent fuel pool. The working group concludes that tomado missile strikes from F4 and F5 tomadoes add negligibly to the risk from spent fuel pools at decommissioned plants. Support system availability - An F2 or larger tornado would be needed to consider damage to a support system, such as power supplies, cooling pumps, heat exchangers and water makeup sources. The likelihood of this magnitude tomado striking the plant is estimated to be 4x10-8 per year, or lower. A4-5

a Recovery from loss of support systems if the spent fuel pool support system (s) are damaged by an F2 or larger tomado, then offsite recovery is considered to be available (e.g., fire trucks) to provide makeup to the spent fuel pool. The assumed conditional probability of ugpyfsite recovery is 0.02 for Cases 1 and 3, and 0.1 for Case 2. Combining the frequency of an F2 orlarger tomado with failure of offsite recovery results in a frequency of fuel uncovery in the spent fuel pool in the range of 8x10 per - 4 - year from damage to support systems caused by tomadoes for Cases 1 and 3 and in the range of4x104 peryear for Case 2. The SPC data is summarized in Table A4h. RAFT Table A4-4 SPC Tomado Summary by State (1950 - 1995)' Fujita damage scale Fraction in F-range State. Total F0 F1 F2 'F3 F4 F5-F4-F5 F2-F5 AL 1031 165 364 323 129 36 14 0.049 0.49 AR 1007 198 298 331-149 31 0 0.031 0.51 l AZ 160 90 57 11 2 0 0 0.000 0.08 CA 223 142 58 21 2 0 0 0.000 0.10 CO 1172 616 441 99 15 1 0 0.001 0.10 CT 65 9 29 20 5 2 0 0.031 - 0.42 DC 1 1~ 0 0 0 0 0 0.000 0.00 DE-55 20 23 11 1 0 0 0.000 0.22 FL 2148 1156 665 293- '30 4 0 0.002 0.15 g GA 1032 147-537 266 65 17 0 0.016 0.34 IA 1607 478 506 421 119 74 9 0.052 0.39 ID 124 63 53 8 0 0 0 0.000 0.06 IL 1342-431 440 -316 113 39 3-0.031 0.35 IN 1038 246-336 263 108 77 8 0.082 0.44 KS~ '2363 1111 610 404 168 54-16 0.030 0.27 KY 483 79 168 133 65 35 3 0.079 0.49 LA-1254 225' '620 268 123 -16 2 0.014 0.33 MA

138 24 72 31 8

3 0 0.022 0.30 MD-172 49 92 26 5 0 0 0.000 0.18' ME 82 21 44 17 0 0 0' O.000 0.21 MI 807-195 308 210 57 30 7 0.046 0.38 MN 953 372 336 158 53 28 6 0.036 0.26 MO-1367 298-577 334 109 48 1 0.036 0.36 MS 1268' 226 468 369 136 59 10 0.054 0.45 MT 253 174 42-33 4 0 0 9.000 0.15 NC' 687-153 321 143 44 26 0 0.038 0.31 1 A4-6

Fujita dam =ga scale Fracton in F-range State Total F0 F1 F2 F3 F4 F5 F4-F5 F2-F5 ND 830 490 211 91 28 7 3 0.012 0.16 NE 1818 827 585 255 105 42 4 0.025 0.22 NV 49 41-8 0 0 0 0 0.000 0.00 NH 75 24 34 '15 2 0 0 0.000 0.23 NJ 127 42 58 23 4 0 0 0.000 0.21 NM 400 261 104 31 4 0 0 0.000 0.09 NY 268 101 106 35 21 5 0 0.019 0.23 OH 733 157 321 166 53 27 9 0.049 0.35 OK 2580 845 808 626 209 83 9 0.036 0.36 OR 49 31 15 3 0 0 0 0.000 0.06 PA 506 93 220 143 1 2S 22 2 0.047 0.38 RI 8 3 4 1 0 0 0 0.000 0.13 SC 516 136 234 100 31 15 0 0.029 0.28 SD 1172 651 259 197 57 7 1 0.007 0.22 TN 596 107 241 139 76 29 4 0.055 0.42 TX 5934 2632 1837 1067 317 76 5 0.014 0.25 UT 79 53 19 6 1 0 0 0.000 0.09 VA 318 84 132 68 28 6 0 0.019 0.32 VT 33 7 14 12 0 0 0 0.000 0.36 WA 56 24 17 12 3 0 0 0.000 0.27 WI 949 204 378 276 62 24 5 0.031 0.39 WV 87 27 36 16 8 0 0 0.000 0.28 WY 444 247 145 43 8 1 0 0.002 0.12 Total 38459 13776 13251 7834 2553 924 121 0.027 0.30 A4-7 j

Appendix 5 Structural Integrity of Spent Fuel Pool Structures Subject to Heavy Load Drops A heavy load drop onto the spent fuel pool wall or into the spent fuel can affect the structural integrity of the spent fuel pool. Previously, the staff evaluated heavy loads as Generic Technical Activity A-36, which resulted in the publication of NUREG-0612, " Control of Heavy Loads at Nuclear Power Plants," U.S. Nuclear Regulatory Commission, July 1980. The working group assumed that shipping cask handling is t c no eavy load operation. In NUREG-0612, the staff evaluated Navy data for cranes and hoists for the period February 1974 to October 1977. There were 74 load drop events identified. Of these,23 percent were attributed to design and maintenance errors, 73 percent to operator (or human) error and 4 percent to rigging errors. The working group also reviewed several,D R A FT more recent studies concerning heavy load drops: 1. Notice No.125 - March 27,1984, U.S. Department of the Interior, Minerals Management Service (MMS), Gulf of Mexico OCS Region The Risk and Safety Analysis Unith N[ reviewed 50 crane-related accidents for the period January 1,1971 to June 30,1983. The major findings included the following: . The major contributing cause of crane accidents has been employee negligence and/or error, About 44 percent of the crane accidents involved some type of equipment failure due to poor maintenance and/or overloading of the crane. f 2. " Findings and Recommendation of the Crane Accident Workgroup" B. Hauser, B. Lewis, W. Rhome, Engineering & Operations Division, October 16,1998 (U.S. Department of the Interior, MMS) The workgroup reviewed 34 incidents for the period 1995 to 1998. (1) 17 incidents were listed as equipment failure (EF) and 12 listed as human error (HE). (2) Fatalities were reported in 1 of the 12 HE incidents and in 5 of 17 EF incidents. (3) Major property damage was reported in 1 of 12 HE incidents and in 6 of 17 EF incidents. 3. " Independent Oversight Special Study of Hoisting and Rigging Incidents Within the Department of Energy," October 1996, Offien of Oversite, Office of Environment, Safety and Health, U.S. Department of Energy AS-1

The Department of Energy (DOE) Occurrence Reporting and Processing System (ORPS) served as the principal information source for incidents relating to hoisting and i rigging (H&R) operations. An initial set of 491 occurrence reports, corresponding to the { October 1,1993, to March 31,1996, period, describing incidents related to H&R. There l were 131 relevant hoisting and rigging incidents between October 1993 and March 1996. The data is summarized in Table A5-1 and Table A5-2. Table A5 Distribution of hbMbIgging incidents and accidents Equipment Number of Number of incidents as a Accidents as a Accidents as a incidents Accidents Percent of Total

• Percent of Total
• Percent of incidents
• Crane 66 49 50/ORAFT 51 %

74 % Forklift 40 36 31 % 38 % 90 % Other** 25 11 19% 11 % 44 % Total 131 96 l100 N >A g T 100% 73 %

• Rounded to the nearest whole number,~ '

" includes manual and power-operated hoists, chainfalls, and block and tackle. Table AS Root cause of hoisting and rigging incidents by equipment type

• Root Cause Crane Forklift Other inattention to Detail 20 %

23% 8% Work Organization and Planning 18% 3% 27 % Procedure Not Used or Used incorrectly 9% 15% 0% Policy Not Adequately Defined, Disseminated, or 9% 10 % 4% Enforced inadequate or Defective Design 5% 5% 19 % Defective or inadequate Procedure 9% 5% 0% inadequate Administrative Control 9% 0% 4% Defective or Failed Part 5% 5% 8% Other Management Problem 3% 3% 12 % Other Human Error 3% 3% 0% inadequate Work Environment 0% 10 % 0% A5-2

Root Cause Crane Forklift Other Lack of Procedure 2% 3% 4% insufficient Refresher Training 3%- 3% 0% Insufficient Practice or Hands-On Experience 5% 0% 0% Communication Problem 2% 3% 4% ) r3 D A C--

w. u u i

inadequate Supervision 0%- 3% 4% Error in Equipment or Material Selection 0% 3% 4% ) Weather 0% 3%' 0% No Training Provided' 0% 0% 4%

• Rounded to the nearest whole number.

Since crane and hoisting operations are gmy U.S. Occupational Safety and Health Administration (OSHA) requirements, aC, m is generally available. However, there is only limited data available on the number of operations, or lifts, for developing load drop per operation data. In NUREG-0612, the staff evaluated the Navy data for the period February 1974 to October 1977. Based on some assumptions regarding the number of cranes and hoist in service, the' number of lifts per year, and the 74 load drops identified in the period studied, it was estimated that frequency of a heavy load drop is in the range of 2.5x10 to 3.0x104 per 4 a operation. in NUREG/CR-4982, " Severe Accidents in Spent Fuel Pools in Support of Generic Safety issue 82," Brookhaven National Laboratory, July 1987, the staff estimated the human error contributor to a heavy load drop as 6.0x104 per lift, based on data concoming human reliability in the positioning of heavy objects. i A more recent study on heavy load drops was performed for the DOE Savannah River Site to quantify hurnan errors, " Savannah River Site Human Error Data Base Development for Nonreactor Nuclear Facilities," Westinghouse Savannah River Co., WSRC-TR-93-581, February 28.1994. The likelihood of a heavy load drop, when using cranes or hoists, was quantified using a generic data base. An operation was defined as " lift, move, and setting down" the load. Based on 200 drops in 2,000 crane years of' operation and additional crane data from nuclear power plants, the drop rate was estimated as 1.5x10d per operating-hour. No error factor information was provided in the study. If it is assumed that it takes 1-hour for an operation (lift, move, and setting down) and using an error factor of 10 from the nominal value (1.5x104), then the estimated values for a human error resulting in a heavy load drop are provided in Table A5-3. AS-3 3

Table A5-3 Department of Energy Recommended Heavy Load Drop Frequencies Frequency Error Factor Type of Load Low mean value 1.5x10-5 / operation 10 Standard L Nominal mean value 1.5x10-4 / operBRAFT 10 Typical High mean value 1.5x10-3 / operation 10 Unusual While the DOE nominal mean value for th of bei drop frequency is a factor of 2 to 4 lower than the values reported in previous evalt ati), te nupporting data base is believed to be representative of the expected cask handling operations in a spent fuel pool, or movement of heavy loads near the spent fuel pool. The DOE recommends that the high value be used for i id til i t and that . unusual, unevenly balanced loads with standard construct on or n us r a equ pmen the low value be applied to standardized I r spotter present. The number of cask operations per year and the time span during which a drop may impact the spent fuel pool wall or fall into the spent fuel pool need to be considered as part of a risk-informed evaluation. The numerical values presented in Table AS-1 have an implied time duration of 1 hour per operation. 1 For heavy loads not involving cask operations, it is expected that a " safe load-path" procedure will be followed. If movement of objects over the spent fuel pool, other than casks, is to be considered, then a description of the assumptions used for the risk-informed evaluation needs to be documented, including items such as weight, number of movements, potential spent fuel pool failure modes (e.g., crush or fail the wall, or puncture the bottom of the pool), and the drop rate frequency assigned to the particular movement.[GOES INTO RECOMMENDATION SECTION] In reviewing the drop per lift data, it was determined that the ranges of va!ues presented in the 1994 DOE study were nearly identical to those used in the 1980 NUREG-0612 study. Therefore, the working group used NUREG-0612 as the bases for developing estimates of heavy load drop frequencies. Non-sinale fM!ure oroof load handlina system Figure B-2 (pg. B-16) in_ NUREG-0612 provides a quantified fault tree for load handling over a spent fuel pool for a load handle system that is not single failure proof. The staff estimated the handling system failure rate to be 1.0x104 to 1.5x108 per lift. This value was based on a nuclear power plant estimate expected to be in the 1.0x104 to 1.5x104 per lift range, including a factor-4 d of-2 improvement over the Navy data evaluated (reported to be in the 2.5x10 to 3.0x10 per lift range), and included an assumption that failure of the interlocks and/or failure to follow an approved load path lead to a factor-of-10 reduction in the load handling system reliability (the staff estimated the interlock / load path failure range to be 2.0x10 to 1.0x104 per year.). The per 8 A5-4

~ lift range from the DOE Savannah River study,'1.5x104 to 1.5x104 for this type of load handling, is nearly the same as the expected NUREG-0612 range. Only some fraction of the load drops will lead to significant spent fuel pool damage, in NUREG-0612 the staff assumed that the only 10 percent of the critical load path was over the j spent fuel.' The working group estimated that between 2 to10 drops would occur for 200 lifts per year resulting in an estimate of 2.0x104 to 1.0x104 drops per year into the spent fuel pool. Therefore the working group estimates thiQalf prpp. o be in the range 2.0x104 to 1.5x10d per. year (applying the NUREG-0612 methodss tKahlem ptions). If the NUREG-0612 assumption on the load handling system reliability is not considered (for example, the load drop is not related to failure of the interiock or the load path) then it was estimated that the load drop is in the range 2.0x104 to 1.5x104 per year. Single-failure-oroof load handhna system D R A F T l Figure B-3 (pg. B-17) of NUREG-0612 provides a quantified fault tree for load handling over spent fuel pool for a single-failure-proof handling system: The staff estimated the failure of the handling system to be in the range of 4.0x104 to 1.0x104 per year. The staff estimated the i likelihood of the drop occurring over spen t te40x104 to 2.5x104 per event (using an i estimate of between 5 percent and 25 pei Te load' path - based on the 10 percent estimate of the total path). The resulting range is 2.0x104 to 2.5x104 per year, a reduction from l the non-single failure proof load handling system range of 2.0x104 to 1.5x104 per year. The single failure proof handling system reduced a load drop by a factor of 10. ~ Calculated values for risk-informed assessment of spent fuel pool The working group estimates the likelihood of a load drop into the spent fuel pool to be in the . range of 2.0x10 to 1.5x104 per year for a non-single failure proof load handling system. For a j 4 single failure proof load handling system, the range is reduced by about a factor of 10 (2.0x104 to 2.5x104 per year). An estimate of the likelihood of substantial damage (rapid pool draining) given the drop is on the pool wall or into the pool is provided. For the failure of the pool wall, it was assumed that the load physically travels over the wall 2 percent (0.02) of the time (10 percent of the 5 to 25 percent of the total load path) and it was assumed that a one-in-ten (0.1) chance of significant damage given load drop. The working 1 group estimated a failure rate in the 2.0x104 to 3.0x10 per year for the non-single failure proof 4 system and 8.0x104 to 2.0x104 per year for the single failure proof system. (The NUREG/CR-4982 value for wall failure was 4 3.7x10 per year.) For failure of the pool floor, it was assumed one-in-ten ' events (0.1) will result in significant ~ damage. The working group' estimated a failure rate range of 2.0x104 to 1.bx104 per year for the non-single failure proof system and 2.0x104 to 2.5x104 per year for the single failure proof system. 1 A5-5 1

I { Insights to be considered in assigning a drop rate to a particular operation, the following information should be considered. L It appears that the human error contribution to heavy load drops has not changed since the original work documented in NUREG-0612 in 1980. The following conclusions were reported in the 1996 DOE, Office of Oversite, study: Human error is the major cause oMEnIrigging (H&R) incidents. 1. Human error, whether directly associated with supervisors or equipment operators, is the principal cause of H&R incidents. Factors not related to human performance, such as equipment failure and weather, are responsible for only 6 percent of H&R incidents- -~i. (33 percent) collectively account for 68 management (35 percent) and pef *into percent of all H&R incidents, as report RPS. 2. _ Management shortcomings and workers' inattention to detail account for a large proportion ofincidents. Further analysis shows that deficib N kaIning (43 percent) and inadequate definition, dissemination, and enforcement of policy (24 percent) are responsible for two ' thirds of the incidents attributable to management deficiencies. Inattention to detail (56 percent) and not following procedures (28 percent) account for 84 percent of H&R incidents caused by personnel error. Furthermore, inattention to detail is the most prevalent cause of all 131 H&R incidents, accounting for about one in every.five incidents. Additionally, there are no indications that certain root causes are becoming less frequent over time, are being remedied, or are being replaced with other causal factors. 3. Work planning is a significant factor in non-forklift incidents. Work organization and planning require more attention in operations involving cranes and "Other" hoisting equipment than when forklifts are utilized. This is evident by the fact that inadequate work planning was the cause of 18 percent of all incidents involving cranes,27 percent of the incidents involving "Other" hoisting (i.e., non-forklift) equipment, and only 3 percent of all forklift incidents. 4. Training-related deficiencies were not identified as a major problem. Training-related deficiencies were not identified as a significant problem. Procedure-related problems, including applying procedures incorrectly, defective or inadequate procedures, or procedures not used, are responsible for 18 percent of crane incidents. They were not found as causal factors for incidents involving "Other" equipment. Communication, lack of procedures, and defective or failed parts cause incidents with approximately equal frequency for all equipment type categories, although it is the greatest for "Other" equipment (e.g., hoists', chainfalls, block and tackle). A5-6 i

5. The use of mobile cranes by subcontractors is expected to increase, heightening the need for effective oversight of subcontractors' safety performance. Discussions with H&R experts within the DOE (Federal workers, contractors, and subcontractors) indicate that as production-related operations are curtailed and superseded with activities directed at waste management, environmental restoration, and facility dismantlement, the need for stationary or overhead cranes will be reduced, and mobile units will be in more degd bile cranes owned and operated by g subcontractors are often used to 1:s..mn is teirial handling tasks of varying complexity, whereas overhead cranes are generally operated by contractors and are used to perform maneuvers that are relatively simple and often routine. Independent evaluations performed by the Office of Oversight, in addition to information reported into ORPS and the Department's Computerized Accident / incident Reporting System (CAIRS), have highlighted deficiencies in oversigggyactor activities. Therefore, the additional risks posed as more H&R tasks involving cranes are performed by subcontractor personnel heightens the concern over H&R safety and the need for effective oversight of subcontractor performance. Information contained in ORPS does not explicitly and formally identify whether an H&R incident is associated with a contractor or subcontractor activity. While it was M nol possible to resolve this issue for the entire

• M dis review to make this determination for some of the 131 incidents analyzeO sample.

The following standards describe the guidelines currently in use based on requirements of the U.S. Occupational Safety and Health Administration (OSHA) and the American National Standards Institute (ANSI): d DOE-STD-1090-99, " Hoisting and Rigging (Formerly Hoisting and Rigging Manual)." U.S. Department of Energy, March 1999. MIL-HDBK-1038, " Weight Handling Equipment," Department of Defense Handbook, March 06,1998. NAVFAC P-307, U.S. Navy, June 1998. The DOE standard occasionally goes beyond the minimum generalindustry standards established by OSHA and ANSl; and also delineates the more stringent requirements necessary to accomplish the extremely complex, diversified, critical, and oftentimes hazardous hoisting and rigging work found within the DOE complex. In doing so, it addresses the following items which are not covered in detail in the general industry standards: 1. Management responsibility and accountability 2.' Operator / inspector training and qualification requirements 3. Definition of critical lifts and the additional requirements for making them ) 4. The need and responsibilities of a person-in-charge for critical lifts A5-7

5. The need and responsibilities of a designated leader for ordinary lifts 6. The definition and special requirements for pre engineered production lifts t Special requirements for the testing, inspection, and maintenance of hoisting equipment 7. in hostile environments Nondestructive testing /nondestruMMnItion requirements for such items as 8. hooks, welds, and spreader bars Special requirements for inspection and load-testing of hoisting and rigging 9. equipment / accessories Hook latch requirements for cranes, shngs,FT DRA end rigging accessories 10. 11. Design standards for such equipment as cranes, forklifts, and hooks r s rations 12. Operating practices for hoisting ar 13. Rigging information and load tables 14. Good and bad rigging practices. AS-8

m Appendix 6 Aircraft Crashes into Spent Fuel Pools Y T The staff used the genenc data provided in DOE-STD-3014-96, " Accident Analysis for Aircraft ? Crash into Hazardous Facilities," U.S. Department of Energy (DOE), October 1996, to assess the L likelihood of an aircraft crash into or near a M%i{ned spent fuel pool. Aircraft damage can affect the structural integrity of the spent fuel pool or affect the availability of nearby support systems, such as power supplies, heat exchangers and water makeup sources, and may also affect recovery actions. The staff obtained the frequency of an aircra "l}ve aircraft target area model: o a site, F, from the four-factor formula in . DOE-STD-3014-96, and will refer to this as t ' F = { N, P, f,(x,y) Ali . Equation 1 m wheref .l DRAFT N,- estimated' annual number of site-specific aircraft operations (nodyr) ' -= aircraft crash rates (per takeoff and landing for near-airport phases) and per P, = flight for in-flight (non airport) phase of operation f,(x, y) = aircraft crash location probability (per square mile) ' site-specific effective area for the facility of interest including skid and fly-in A, , = effective areas (square miles) (index for a flight phase): i=1,2, and 3 (takeoff, in-flight, landing) i = -J =, (index for aircraft category, or subcategory) ' k. (index for flight source): there could be multiple runways and non airport = operations The site-specific area is further defined as: eff

• f + ^s Equation 2 where:

A = (WS + R)-(H cotq)+ 2 L W WS-+L W f R A, = (WS + R) S where: A., = total effeuNe target area H = height of facility Ai = effective fly-in area L = length of facility. A, = effective skid area W = width of facility WS = wing span S = aircraft skid distance cote = mean of cotangent of aircraft R = length of facility diagonal - impact angle Altematively, a point target area model can be defined as just the area (length times width) of the facility in question.~ In Table A6-1 the staff has sunimarized the generic aircraft data and crash frequency values for five aircraft types from Tables B-14 through B-18 of DOE-STD-3014-96. A6-1

Table A6-1 Generic Aircraft Data Aircraft Wingspan Skid distance cot 8 Crashes per mi2-yr Notes: DRAFT Min Ave Max 4 2x10d 3x10-8 General 50 1440 10.2 1x10 aviation Air carrier 98 60 MM r-7x104 4x104 2x10'* r-T Air taxi 58 60 8Y '4x104 1x104 8x104 4 4 Large 223 780 7.4 6x104 2x10 7x10 takeoff military Small 100 447 DR94F F4x104 4x104 6x104 landing military The working group used the data presented in Table A6-1 to determine the frequency of aircraft hits per year for various building sizes (length, width, and height) for the average and maximum crash rates. The resulting frequencies are presented in Table A6-2. The product N,*P,*f,(x,y) for Equation 1 is taken from the crashes per mi2-yr and A,is obtained from Equation 2 based on the aircraft characteristics. Two sets of data were generated: one including the wing and skid lengths using the effective aircraft target area model and a second case which considered only the area (length times width) of the site using the point target area model. The working group chose the building or facility characteristics to cover a range typical of a spent fuel pool to that of the PWR auxiliary building or the BWR secondary containment structure. j f A6-2 l j

1 Table A6 Aircraft Hits per Year Building (L x W x H)(ft) Effective Area (mi2) Average Hits / year Maximum Hits / year j With the DOE effective DRAF T: I aircraft target area model 100 x 50 x 30 1.0x10 2 2.1x104 3.1x104 200 x 100 x 30 1.8x14 A r 3.7x104 5.5x104 T T 400 x 200 x 30 3.5x10 7.0x10-8 1.0x10d 200 x 100 x 100 2.5x10-2 5.0x104 7.6x10-5 400 x 200 x 100 4.7x10-2 9.5x104 1.4x104 80 x 40 x 30 9.0x1M I 1.8x10* 2.7x10 5 With the point target area-j model l 100 x 50 x 0 1.8x104 3.7x104 5.4x10-7 200 x 100 x 0 7.2x104 1.5x10-7 2.2x104 400 x 200 x 0 2.9x104 5.9x10-7 8.6x104 .80 x 40 x 0 1.1x104 2.4x104 3.5x10-7 i The working group compared the DOE effective aircraft target area model, using the generic ( data in Table A6-1, to the results of two evaluations reported in "Probabilistic Safety Assessment i and Management," A. Mosleh and R.A. Bari (editors), PSAM 4, Volume 3, Proceedings of the 4-th International Conference on Probabilistic Safety Assessment and Management,13-18 September 1998, New York City, USA. The first evaluation of aircraft crash hits was summarized by C.T. Kimura, et al., in " Aircraft Crash Hit Analysis of the Decontamination and Waste Treatment Facility (DWTF) at the } Lawrence Livermore National Laboratory (LLNL)." DWTF Building 696 was assessed. It is a l 254 feet long by 80 feet wide,1 story,39 feet high structure. The results of Kimura's study are j shown in Table A6-3. i A6-3

Table A6-3 DWTF Aircraft Crash Hit Frequency (per year) Period Air Carriers Air Taxies ygergpviation Military Aviation Totalm 1995 1.72x10-7 2.47x104 225k105 5.03x10-7 2.76x104 4 2.82x10-8 6.47x10-7 3.16x104 1993-1995 1.60x10-7 2.64x10 1991-1995 1.57x10-7 2.58x104 48pxyt_ 7.23x10-7 3.23x104 5 4 1986-1995 1.52x10'7 2.41x104 N.'89x10" 8.96x10-7 3.23x10 Note (1): Various periods were studied to assess variations in air field operations. bn a frequency range of 4.4x104 hits per Applying the DOE generic data to the D year to 6.6x104 hits per year for the effective aircraft target area model. For the point target area model, the range was 1.5x10-7 to 2.2x10-e per year. The second evaluation was presented in a paper by K. Jamali, et al., " Application of Aircraft Crash Hazard Assessment Methods to Various Facilities in the Nuclear Industry," in which additional facility evaluations were summarized. For the Seabrook Nuclear Power Station, Jamali's application of the DOE effective aircraft target area model to the Final Safety Analysis Report (FSAR) data resulted in an impact frequency 2.38x104 per year. The Millstone 3 plant area was reported as 9.5x10 square miles and the FSAR aircraft crash frequency was reported 4 to be 1.6x104 per year. Jamali applied the DOE effective aircraft target area model to information found in the Millstone 3 FSAR. Jamali reported an impact frequency of 2.74x104 per year using the areas published in the FSAR and 2.31x104 per year using the effective area calculated by the effective aircraft target area model. When the working group used the generic DOE data in Table A6-1, it resulted in an estimated impact frequency range of 2x104 to 3x10-5 per year for the point target area model, and 1.6x10-8 to 2.4x104 per year for the effective aircraft target area model. The staff documented a site-specific evaluation for Three-Mile Island Units 1 and 2 in NUREG/CR-5042, " Evaluation of External Hazards to Nuclear Power Plants in the United States," Lawrence Livermore National Laboratory, December 1987. The NUREG estimated the aircraft crash frequency to be 2.3x10 accidents per year, about the same value as would be 4 predicted with the DOE data set for the maximum crash rate for a site area of 0.01 square miles. In NUREG/CR-5042, the staff summarized a study of a power plant response to aviation accidents. The results are presented in Table A6-4. A6-4

( i i Table A6-4 Probability of Penetration as a Function of Location and Concrete Thickness ] g

- T--

Probability of penetration Thickness of reinforced concrete Plant location Aircraft type 1 foot 1.5 feet 2 feet 6 feet \\ s 5 miles Small s 12,000 lbs. m0g-0 0 0 i Large > 12,000 lbs. OU96 O.52 0.28 0 > 5 miles Small s 12,000 lbs. 0.28 0.06 0.01 0 Large > 12,000 lbs. 1.0 1.0 0.83 0.32 LJN/W I Y The working group has reasonable assurance that the DOE model and generic data provide a - range of aircraft crash hit frequencies that would be consistent with plant-specific evaluations. At this level of effort, the resulting damage from an aircraft crash cannot be fully evaluated on a plant-specific bases. ' A detailed structural evaluation is beyond the scope of this effort. In general, PWR spent fuel pools are located on, or below grade, and BWR spent fuel pools, while generally elevated about 100 feet above grade, are located inside a secondary containment structure. The vulnerability of support systems (power supplies, beat exchanges and makeup water supplies) requires a knowledge of the size and location of these systems, information not readily available. Calculated values for risk-informed assessment of spent fuel pool Significant pool damage PWR The working group calculated a value for significant PWR spent fuel pool damage resulting from a direct hit. It is based on the point target area model for a (100 x 50) foot pool with a conditional probability of 0.3 (large aircraft penetrating 6-ft of reinforced concrete) that the crash results in significant damage. If 1-of-2 aircraft are large and 1-of-2 crashes result in spent fuel uncovery, then the estimated range is 2.7x104 to 4.0.x104 per year. BWR The working group calculated a value for significant BWR spent fuel pool damage resulting from a direct hit. It is the same as that for the PWR,2.7x104 to 4.0x104 per year. Mark-l and Mark-ll secondary containments do not appear to offer any significant structures to reduce the likelihood of penetration, although on one side there may be a reduced likelihood due to other structures. Mark-Ill secondary containments may reduce the likelihood of penetration as the spent fuel pool may be considered to be protected by j additional structures. ) J A6-5

/ v i i These values for significant damage to spent fuel pools are significantly less than estimates for other initiators. 1 Loss of support system The working group estimated the value fo@p(fs}pport system (power supply, heat exchanger or makeup water supply) based on the DOE model including wing and skid area for a (400 x 200 x 30) foot area with a conditional probability of 0.01 that one of these systems is hit. The estimated value range is 7x10 to 1.0x104 per year. 4 The working group estimated the value fo i fs}Ipport system (power supply, heat exchanger or makeup water supply) base on e DOE model including wing and skid area for a (10 x 10 x 10) foot structure. The estimated value range is 7.3x10-7 to 1.1x1.04 per year with the wing and skid area, and 7.4x1048 to 1.1 x10 per year for the point model. 4 These expected values of support systemggg than those for other initiators. qq A6-6 m}}