ML19178A252: Difference between revisions
StriderTol (talk | contribs) (Created page by program invented by StriderTol) |
StriderTol (talk | contribs) (Created page by program invented by StriderTol) |
||
(One intermediate revision by the same user not shown) | |||
Line 15: | Line 15: | ||
=Text= | =Text= | ||
{{#Wiki_filter:/RA/ | {{#Wiki_filter:June 13, 2019 MEMORANDUM TO: Mirela Gavrilas, Director Division of Safety Systems Office of Nuclear Reactor Regulation FROM: Victor G. Cusumano, Chief /RA/ | ||
Technical Specifications Branch Division of Safety Systems Office of Nuclear Reactor Regulation | |||
==SUBJECT:== | |||
TECHNICAL EVALUATION REPORT OF IN-VESSEL DEBRIS EFFECTS By letter dated July 17, 2015, the Pressurized Water Reactor (PWR) Owners Group submitted a licensing topical report, WCAP-17788, Comprehensive Analysis and Test Program for GSI | |||
[Generic Safety Issue]-191 Closure (Agencywide Documents Access and Management System (ADAMS) Accession No. ML15210A667). WCAP-17788 was intended to support the closure of GSI-191, Assessment of Debris Accumulation on PWR Sump Performance, and Generic Letter (GL) 2004-02, Potential Impact of Debris Blockage on Emergency Recirculation During Design Basis Accidents at Pressurized-Water Reactors, dated September 13, 2004 (ADAMS Accession No. ML042360586), by creating a methodology to define an in-vessel fibrous debris limit to respond to in-vessel questions in GL 2004-02. | |||
The review of WCAP-17788 identified several issues that were not readily resolved. As a result, the staff of the U.S. Nuclear Regulatory Commission (NRC) reevaluated the overall significance of the effects of debris on in-vessel blockage considering improvements in understanding gained in recent years. This is in accordance with Office of Nuclear Reactor Regulation efforts to ensure that agency resources are expended on issues commensurate with their safety significance. As part of that evaluation of safety significance, the enclosed technical evaluation report (TER) evaluated the effects of debris on in-vessel blockage against risk thresholds consistent with the Commissions health objectives. The NRC staff determined that the issue is generally of low safety significance. | |||
CONTACTS: Ashley Smith Stephen Smith Paul Klein Ben Parks 301-415-3201 301-415-3190 301-415-3040 301-415-0979 Matt Yoder Shilp Vasavada Andrea Russell 301-415-4017 301-415-1228 301-415-8553 | |||
M.Gavrilas In part, GL 2004-02 requested licensees to evaluate in-vessel effects for their plants. The NRC staff is evaluating compliance with Title 10 of the Code of Federal Regulations 50.46, Acceptance Criteria for Emergency Core Cooling Systems for Light-Water Nuclear Power Reactors, also requested in GL 2004-02, separately from the enclosed TER. The information contained in the TER, however, may be useful to staff and licensees while evaluating compliance with that regulation. | |||
==Enclosure:== | |||
Technical Evaluation Report | |||
M.Gavrilas | |||
** | ==SUBJECT:== | ||
TECHNICAL EVALUATION REPORT OF IN-VESSEL DEBRIS EFFECTS Dated: June 13, 2019 DISTRIBUTION: | |||
VCusumano, NRR ASmith, NRR SSmith, NRR NON-PUBLIC SVasavada, NRR RLukes, NRR MYoder, NRR SBloom, NRR ARussell, NRR LPerkins, NRR PKlein, NRR JDrake, NRR MReisi-Fard, NRR JBorromeo, NRR JDonoghue, NRR CHoxie, RES MGavrilas, NRR BBenney, NRR TZaki, RES GCasto, NRR SBajorek, RES JStaudenmeier, RES BParks, NRR ADAMS ACCESSION NO. ML19073A044 *concurred via e-mail NRR-106 OFFICE NRR/DSS/SRXB NRR/DSS/STSB NRR/DMLR/MCCB NRR/DSS/SNPB NAME ASmith* SSmith* PKlein* BParks* | |||
DATE 5/30/2019 5/29/2019 5/31/2019 6/4/2019 OFFICE NRR/DSS/STSB NRR/DRA/RILIT: NRR/DMLR/MCCB NRR/DLP/LA TL (Acting) | |||
NAME ARussell* SVasavada* MYoder* DHarrison* | |||
DATE 5/29/2019 6/3/2019 5/31/2019 6/5/2019 OFFICE QTE NRR/DSS/SNPB: NRR/DRA/APLB:BC NRR/DSS/SRXB BC (Acting) :BC (Acting) | |||
NAME JDougherty* RLukes* MReisi-Fard* JBorromeo* | |||
DATE 3/21/19 6/10/2019 6/7/2019 6/5/2019 OFFICE NRR/DMLR/MCCB: RES/DSA/CRAB: NRR/DSS/STSB:BC BC BC NAME SBloom* CHoxie* VCusumano DATE 6/5/2019 5/19/2019 6/13/2019 | |||
- OFFICIAL RECORD COPY - | |||
TECHNICAL EVALUATION REPORT OF IN-VESSEL DEBRIS EFFECTS Enclosure | |||
i Contents 1.0 EXECUTIVE | |||
==SUMMARY== | |||
.............................................................................................. 2.0 LIST OF ACRONYMS .................................................................................................. | |||
==3.0 INTRODUCTION== | |||
==4.0 BACKGROUND== | |||
............................................................................................................ 4.1 Overview of WCAP-17788 ......................................................................................... | |||
2 | ==5.0 TECHNICAL EVALUATION== | ||
......................................................................................... 5.1 System Description .................................................................................................. 5.1.1 Key Parameters ................................................................................................ 5.2 Sequence of Events ................................................................................................ 5.2.1 Availability of Emergency Core Cooling System Primary Flowpath ................. 5.2.1.1 Cold-Leg Break ......................................................................................... 5.2.1.1.1 Debris Collection at the Core Inlet ......................................................... 5.2.1.1.2 Boric Acid Precipitation .......................................................................... 5.2.1.1.3 Debris Suspended in the Core ............................................................... 5.2.1.1.4 Local Collection of Debris ...................................................................... 5.2.1.1.5 Concentration of Particulate Debris ....................................................... 5.2.1.1.6 Conclusions for the Cold-Leg Break ...................................................... 5.2.1.2 Hot-Leg Break ........................................................................................... 5.2.1.2.1 Fiber Penetration ................................................................................... 5.2.1.2.2 Boric Acid Precipitation .......................................................................... 5.2.1.2.3 Conclusions for the Hot-Leg Break ........................................................ 5.2.1.3 Upper Plenum Injection Plants .................................................................. 5.2.1.3.1 Conclusions for Upper Plenum Injection Plants ..................................... 5.2.1.4 Debris Generation and Transport .............................................................. 5.2.1.5 Chemical Effects ....................................................................................... 5.2.1.5.1 Testing ................................................................................................... 5.2.1.5.2 Test ResultsPrecipitation Timing ........................................................ 5.2.1.5.3 Staff Analysis of Precipitate Timing ....................................................... 5.2.1.5.4 Chemical Deposits on Fuel .................................................................... 5.2.1.6 ConclusionAvailability of Primary Flowpath ........................................... 5.3 Defense in Depth ..................................................................................................... | |||
ii 5.3.1 Containment ..................................................................................................... 5.3.2 Operator Actions .............................................................................................. 5.3.3 Alternate Flowpaths .......................................................................................... 5.3.3.1 Background ............................................................................................... 5.3.3.2 Simplifying Assumptions ........................................................................... 5.3.3.3 Description of Analyses ............................................................................. 5.3.3.4 Analytic Methods ....................................................................................... 5.3.3.5 Conservative Assumptions ........................................................................ 5.3.3.6 Results ...................................................................................................... 5.4 TRACE Sensitivity Studies ...................................................................................... 5.4.1 CLB .................................................................................................................. 5.4.1.1 Westinghouse Upflow Plant ...................................................................... 5.4.1.1.1 Blockage at the Lower Core Plate (99 percent) ..................................... 5.4.1.1.2 Blockage at the Lower Core Plate (99.5 and 99.9 percent) ................... 5.4.1.1.3 Blockage at the Fuel Inlet Nozzles (99 and 100 Percent) ...................... 5.4.1.1.4 Blockage at the Fuel Inlet Nozzle with Delayed HLSO (100 percent) .... 5.4.1.1.5 Core Power Sensitivity ........................................................................... 5.4.1.1.6 Axial Power Shape Sensitivities ............................................................. 5.4.1.1.7 Large Split Break at Top of CL ............................................................... 5.4.1.2 Combustion Engineering Plant .................................................................. 5.4.1.3 TRACE Study Applicability to the PWR Fleet ........................................... 5.4.2 HLB .................................................................................................................. 5.4.2.1 Westinghouse Upflow Plant ...................................................................... 5.4.2.2 Combustion Engineering Plant .................................................................. 5.4.3 Applicability to the Pressurized Water Reactor Fleet ....................................... 5.5 Integrated Decisionmaking ...................................................................................... 5.5.1 Determination of Safety Significance ............................................................... 5.5.2 Risk Insights - Initiating Event (LOCA) Frequencies ....................................... 5.5.3 Safety Margins ................................................................................................. | |||
==6.0 CONCLUSION== | |||
==7.0 REFERENCES== | |||
1.0 EXECUTIVE | |||
==SUMMARY== | |||
Generic Safety Issue (GSI)-191, Assessment of Debris Accumulation on PWR [pressurized water reactor] Sump Performance, is a longstanding issue and a brief background of the issue is included in this technical evaluation report (TER). Initial efforts by industry and the U.S. | |||
Nuclear Regulatory Commission (NRC) evaluated sump strainer blockage. More recent testing and analysis focused on the potential for fine debris to pass through the sump strainer and block the reactor vessel (RV) core inlet (in-vessel downstream effects (IVDEs)). The NRC staffs safety evaluation (SE) of the initial Pressurized Water Reactor Owners Group (PWROG) fuel assembly (FA) test program in WCAP-16793-NP-A, Revision 2, Evaluation of Long Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid, issued July 2013 (1) resulted in a core fiber limit that permitted approximately one third of operating PWRs to successfully demonstrate long-term core cooling (LTCC). | |||
A second PWROG in-vessel testing and analysis program, described in WCAP-17788, Comprehensive Analysis and Test Program for GSI-191 Closure, submitted July 2015 (2), | |||
used a design specific approach and considered phenomena timing to justify higher plant specific in vessel fiber amounts. The NRC staff reviewed WCAP-17788 but did not complete its review or determine that WCAP-17788 could be approved for use by licensees. However, during its review of WCAP-17788, the NRC staff identified a significant amount of evidence indicating that potential IVDEs have low safety significance. | |||
The evaluation documented in this TER combines the evidence in WCAP-17788 with the NRC staffs independent analyses and evaluations to provide a more definitive conclusion regarding the safety significance of IVDEs. The staff used risk insights to differentiate the type of evaluation performed based on the likelihood of a break size. For more likely (smaller) breaks, mechanistic analyses for some of the highest fiber plants showed successful mitigation of IVDEs. For less likely (larger) breaks, the NRC staff established that the safety significance is low by considering the low initiating event frequencies and thermal hydraulic (TH) analyses which show that even in the case of substantial core inlet blockage, alternate flow paths (AFPs) through the core barrel-baffle region should be effective for all operating plant designs. All evidence considered indicated that IVDEs are unlikely to compromise LTCC following a loss-of-coolant accident (LOCA). | |||
The staff evaluation considered differences in plant designs, and the differing system responses associated with various break locations. The staff considered cold leg (CL) and hot leg (HL) breaks separately because of differences in plant response and issues (e.g., boric acid precipitation (BAP)) that are more significant for certain types of breaks. Plants with upper plenum injection (UPI) were also considered separately because of their unique design features. | |||
The timing of chemical precipitation was considered in relation to the timing of post-LOCA actions. FA tests have shown that higher fiber quantities can be tolerated if no precipitates are present, while actions such as switching to HL recirculation would mitigate the impact of any precipitates by bypassing any blockage and potentially disrupting bed formation at the core inlet. | |||
This TER applied a novel approach to conclude that the safety significance of IVDEs is low since post LOCA debris is unlikely to result in a loss of LTCC. It is not intended to establish compliance with existing regulations (i.e., Title 10 of the Code of Federal Regulations (10 CFR) 50.46, Acceptance Criteria for Emergency Core Cooling Systems for Light Water Nuclear Power Reactors) (3). The NRC staff is evaluating compliance in an effort separate from this TER. The information contained in the TER, however, may be useful to staff and licensees while evaluating compliance with that regulation. | |||
2.0 LIST OF ACRONYMS ADAMS Agencywide Documents Access and Management System AFP alternate flow path ANL Argonne National Laboratory ANS American Nuclear Society ANSI American National Standards Institute B&W Babcock & Wilcox BAP boric acid precipitation BB barrel/baffle BN bottom nozzle BWNT Babcock & Wilcox Nuclear Technologies CDF Core Damage Frequency CE Combustion Engineering CFR Code of Federal Regulations CL cold leg(s) | |||
CLB cold-leg break(s) | |||
CSS containment spray system CVCS chemical and volume control system DC downcomer DEGB double-ended guillotine break DID defense-in-depth ECCS emergency core cooling system EM evaluation model EPRI Electric Power Research Institute F Fahrenheit FA fuel assembly g/FA grams per fuel assembly GL Generic Letter GSI Generic Safety Issue HPI high pressure injection HL hot leg HLB hot-leg break HLSO hot-leg switchover IBOB initial set of autoclave tests with debris not in mesh bags IVDE in-vessel downstream effects LOCA loss-of-coolant accident LOCADM Loss-of-Coolant Accident Deposition Model LP lower plenum LTCC long-term core cooling NRC Nuclear Regulatory Commission NRR Office of Nuclear Reactor Regulation PCT peak cladding temperature psi pounds per square inch PTFE polytetrafluoroethylene PWR pressurized water reactor PWROG Pressurized Water Reactor Owners Group RAI request for additional information RCS reactor coolant system | |||
***** | RES Office of Nuclear Regulatory Research RG Regulatory Guide RLBLOCA realistic large-break LOCA RV reactor vessel RWST refueling water storage tank SE safety evaluation SG steam generator(s) | ||
** | SI safety injection SSO sump switchover STB sodium tetraborate STP South Texas Project TER technical evaluation report TH thermal-hydraulic(s) | ||
* | TSP trisodium phosphate UP upper plenum UPI upper plenum injection WCAP Westinghouse Technical Report Number Preface (formerly Westinghouse Commercial Atomic Power) | ||
***** | ZOI zone of influence | ||
******}} | |||
==3.0 INTRODUCTION== | |||
Each operating PWR licensee is required to demonstrate that LTCC can be maintained as required by 10 CFR 50.46, considering that debris may have an impact on system function. | |||
This provides assurance that the emergency core cooling system (ECCS) strainers will function adequately, that the ECCS and containment spray system (CSS) pump net positive suction head margins are maintained, and that downstream components do not become blocked or damaged by debris that may pass through the strainer. | |||
The NRC staff issued Generic Letter (GL) 2004-02, Potential Impact of Debris Blockage on Emergency Recirculation During Design Basis Accidents at Pressurized Water Reactors, dated September 13, 2004 (4) to request PWR licensees to evaluate the effects of debris on LTCC as stated in 10 CFR 50.46(b)(5). Specifically, 10 CFR 50.46(b)(5) requires that the calculated core temperature shall be maintained at an acceptably low value and decay heat shall be removed for the extended period of time required by the long-lived radioactivity remaining in the core, following the initial system transient that results from a primary system coolant pipe break. | |||
Information in GL 2004-02 summarizes applicable operating experience and previous NRC generic communications, including GSI-191. Considered together, GL 2004-02, GSI-191, and the operating experience referenced within these documents, describe the impact that debris can have on LTCC. | |||
The industry response to GL 2004-02 included: (1) substantial improvements to plant safety via ECCS recirculation sump design modifications and enhancements to containment cleanliness and (2) increases in phenomenological understanding of the various issues associated with LTCC when considering the presence of debris. Examples of the former include the installation of significantly larger ECCS suction strainers and reductions of potential debris sources in PWR containments. Examples of the latter include a large body of experimental testing, advances in investigating the risk significance through probabilistic risk analysis, and investigation of downstream effects using computer codes for TH analysis. However, IVDEs remain unresolved | |||
for approximately two-thirds of the PWR fleet. This report evaluates the overall safety significance of IVDEs. | |||
The NRC staff conducted this technical evaluation in accordance with Office of Nuclear Reactor Regulation (NRR) efforts to ensure that in expending agency resources, the staff gives priority to issues in accordance with their safety significance (see Section 5.5). Given the ongoing technical issues with portions of the TH analysis in WCAP-17788, the staff reevaluated the overall significance of the effects of debris on in-vessel blockage, considering improvements in understanding gained in recent years. Alternative means of demonstrating ECCS performance and compliance with 10 CFR 50.46, as requested in GL 2004-02, are being developed separately, and may be informed by this evaluation. | |||
The NRC staff used an integrated decisionmaking process to make its conclusions regarding IVDEs. This TER relies on engineering judgment, and qualitative and quantitative evidence, including risk insights. The staff approach is provided within the following sections of the TER: | |||
* Section 4.0, Background, provides an overview of the GSI-191 history. | |||
* Section 5.0, Technical Evaluation, describes the scenario that requires evaluation of IVDEs at a high level. | |||
* Section 5.1, System Description, describes the initial plant response to a large LOCA. | |||
This section also includes the key parameters that affect the reliability of the primary flow path and/or DID risk insights. | |||
* Section 5.2, Sequence of Events, describes phenomena that apply to all LOCA scenarios. It also describes how hot-leg breaks (HLBs) and cold-leg breaks (CLBs) affect the system response for typical PWR designs and UPI designs. This section evaluates chemical effects and subscale FA testing and associated analyses. Finally, this section provides the basis for the staff conclusion that the primary coolant flow path (through the fuel inlet nozzles) will provide adequate LTCC. | |||
* Section 5.3, Defense in Depth (DID), discusses plant options to provide LTCC if debris completely blocks the core inlet. The DID topics include containment, operator actions, and AFPs. | |||
* Section 5.4, TRACE Sensitivity Studies, provides a summary of supplemental NRC calculations. | |||
* Section 5.5, Integrated Decisionmaking, defines the criterion NRC staff used to determine safety significance. Section 5.5 also establishes an expected frequency for initiating events that could challenge LTCC due to IVDEs. Finally, Section 5.5 considers margins that provide confidence that the scenarios that are considered to successfully maintain LTCC through the design flow path will be successful. | |||
In addition to the testing and analysis performed by the PWROG in support of WCAP-17788, the NRC staff used additional information and engineering judgment to inform the TER conclusions, as summarized below. | |||
* In WCAP-17788, the PWROG referred to heated rod testing (5) and brine testing (6). | |||
The staff reviewed these test programs to determine the extent to which the information | |||
from them could be used in the TER. The test programs demonstrated phenomenological behavior and are referred to solely in this respect. | |||
* The NRC staff used a combination of information from WCAP-17788 and independent NRC staff TRACE TH analyses to reach the conclusions in this TER. The WCAP-17788 analyses were performed using computer codes approved for some aspects of LOCA analysis, but not specifically for application to LTCC. The novelty of the application led to significant questions during the WCAP-17788 review. These issues and the staff approach for addressing them are discussed further in Section 5.3.3. The analyses were not typical of licensing basis LOCA calculations in that they were not plant specific and lacked rigorous qualification for some key phenomena in the models. The staff determined that the results are representative of the classes of plants being modeled. | |||
The staff used the analyses to determine acceptable HLB fibrous debris loads, demonstrate typical flow patterns within the RV, and demonstrate AFP viability. The staff considers the use of the WCAP-17788 analysis results, in combination with the independent NRC staff TRACE results, sufficient to support the TER conclusions. | |||
This TER is an evaluation of the general safety significance of IVDEs. The staff believes this analysis is significantly conservative compared to typical operating US PWRs, but it is not an evaluation of any specific plant. It does not address methods for plants to demonstrate compliance with regulations or define licensing basis parameters for the IVDE issue. | |||
==4.0 BACKGROUND== | |||
As discussed above, GL 2004-02 required plants to evaluate the effects of debris on LTCC as defined in 10 CFR 50.46. | |||
In the NRC document, Revised Content Guide for Generic Letter 2004-02 Supplemental Responses, issued November 2007 (7), the NRC staff identified several areas for licensees to evaluate to ensure that plants have the ability to mitigate LOCAs. All plants performed evaluations to address the areas delineated in the content guide. The technical areas of the content guide fall into two general categories, strainer and downstream effects. The downstream effects are further split into in-vessel effects and effects on other equipment. | |||
Industry developed guidance to address evaluation of the strainer and downstream effects on equipment (Nuclear Energy Institute 04-07, Volume 1, Pressurized Water Reactor Sump Performance Evaluation Methodology, issued December 31, 2004 (8), and WCAP-16406, Revision 1, Evaluation of Downstream Sump Debris Effects in Support of GSI 191, issued March 2008 (9)). The NRC endorsed this guidance with modifications. | |||
All licensees addressed aspects of the effects of debris on LOCA mitigation. For example, all operating plants installed larger strainers and implemented actions to limit the introduction of debris into containment, and to ensure that debris is removed from containment prior to restarting units. The upgraded strainers installed in the plants have larger surface areas to accommodate debris that may arrive and have smaller screen openings when compared to earlier strainer designs. These smaller openings decrease the size distribution of debris that may pass downstream and enter the reactor core or affect other equipment. | |||
Most PWRs provided sufficient information to demonstrate that their strainers will function adequately under conditions caused by the most challenging LOCAs for their plant. A few are still completing evaluations, and some are supplementing deterministic evaluations with | |||
risk-informed analyses. All licensees have evaluated ex-vessel downstream effects and used NRC approved WCAP-16406-P-A (9), to evaluate the effects of debris on pumps, valves, spray nozzles, orifices, and instrumentation. to SECY-12-0093, Closure Options for Generic Safety Issue-191, Assessment of Debris Accumulation on Pressurized-Water Reactor Sump Performance, dated July 9, 2012, (10) provides a history of GSI-191. In the staff requirements memorandum (SRM) (11) associated with SECY-12-0093, the Commission directed the NRC staff to use risk insights to close GL 2004-02. The SRM also stated that based on plant actions there is adequate DID to allow the plants time to demonstrate compliance with the regulations. Industry continued testing and analysis to demonstrate that in-vessel debris loads are acceptable and to show acceptable strainer performance. In accordance with the SRM, approximately six licensees elected to use risk-information to show acceptable system performance. | |||
Industry also developed guidance (WCAP-16793-NP-A) (1) for the evaluation of the effects of debris in the RV on LTCC. The NRC approved this guidance in an SE (12). The debris limits approved in the guidance were based on bounding assumptions that cover all plant designs. | |||
WCAP-16793 was able to be used by 19 PWR units to conclude that in-vessel debris would not inhibit LTCC. The NRC staff is aware of five additional low fiber units that could use the WCAP-16793 guidance to demonstrate acceptable in-vessel performance but have chosen not to in order to gain larger margin to debris limits. The NRC staff recognizes that the WCAP-16793 limits are conservative because the methods used to develop them were greatly simplified to reduce resource expenditure and to assure that test conditions bounded the PWR fleet. Two additional units performed plant-specific TH analyses to demonstrate that LTCC would be assured, even if a greater amount of debris than approved in WCAP-16793-NP-A was transported to the core (13). | |||
The licensees for the remaining units have calculated, or assumed, that the amounts of debris that will pass through the strainer and enter the reactor coolant system (RCS) are greater than the approved WCAP-16793 limit. The PWROG developed and submitted WCAP-17788 (2) to demonstrate on a plant-specific basis that greater amounts of debris in the core would not adversely affect LTCC. WCAP-17788 incorporated results from FA testing, chemical effects testing, TH analysis, and additional mathematical modeling. It also referred to other work that increased the knowledge of the effects of debris on LTCC, particularly in the RV. | |||
WCAP-17788 evaluated chemical effects, HLBs, and CLBs separately. Instead of using a bounding analysis as in WCAP-16793-NP-A, the PWROG used design-specific information to show that larger amounts of debris could be accommodated without inhibiting LTCC. The PWROG differentiated the analyses between four reactor designs and accounted for differences in fuel design. The method also allowed plants to consider plant-specific design characteristics such as ECCS flow rates, strainer size, chemical effects, and the number of FAs in the core. | |||
The analyses made bounding assumptions for some of the inputs like decay heat curves, timing of switchover to recirculation, FA flow velocity, and debris mixture arriving at the core. The analyses relied on a relatively complex TH analysis to demonstrate that LTCC could be maintained when the core inlet flow was impeded by debris following switchover to sump recirculation. The analyses also assumed that debris eventually completely blocked the core inlet. | |||
4.1 Overview of WCAP-17788 The NRC staff did not complete its review of WCAP-17788, and thus, it is not approved for use. | |||
However, the NRC staff relied on the analyses, evaluations, and experimental data submitted in WCAP-17788 to support its evaluation of the safety significance of IVDEs. WCAP-17788 proposed a methodology to use plant-specific analyses to determine in-vessel debris amounts for each plant based on plant-specific characteristics (e.g., plant-specific decay heat, strainer characteristics, debris transported to the strainer, flow rates, time of sump recirculation, activation of AFPs, formation of chemical precipitates, and implementation of measures to prevent BAP). Once the mass of fiber entering the RV is established, it is compared to an acceptance criterion calculated for each plant and fuel design to determine if the calculated fiber load would inhibit LTCC. | |||
The NRC staff reviewed WCAP-17788 and developed numerous requests for additional information (RAIs), especially about the TH analysis. The PWROG used computer codes used in ECCS evaluation models (EMs) that had been found acceptable for short-term system response analysis and attempted to determine a bounding peak cladding temperature (PCT) to demonstrate LTCC adequacy. The staff found that the codes used for the analyses, while previously approved for short-term LOCA analyses, had not been accepted by the NRC staff to evaluate the phenomena associated with LTCC including predicting a PCT to demonstrate adequate LTCC. In some cases, the results indicated nonphysical behavior. To allow plants to resolve the issues individually despite these uncertainties, the NRC staff drafted limitations on the use of WCAP-17788. These limitations would have required significant plant-specific analyses. | |||
==5.0 TECHNICAL EVALUATION== | |||
The scenario of relevance for evaluating IVDEs is initiated by a primary coolant system pipe rupture or break that leads to a LOCA. Because reactor coolant is discharged through the pipe break, debris is generated within containment. Although the ECCS is initially aligned to a reservoir (e.g., borated water storage tank or refueling water storage tank (RWST)), once the liquid in the reservoir is depleted, the ECCS suction must be realigned for recirculation by switching the source from the depleted reservoir to the containment sump (i.e., sump switchover (SSO)), which collects primary system liquid that has spilled out of the pipe break or collected from containment spray. | |||
Under recirculation cooling, the coolant that spills out of the break and returns to the containment sump can pick up debris dislodged by the initial pipe break, as well as other debris that may be in containment. The system includes a strainer that is intended to collect most of the debris, but some debris will pass though the strainer, and be pumped into the RCS. Once within the reactor, the debris will transport through the RV. Design features like FA debris filters may collect some of the debris, while other debris may settle in more stagnant regions of the system. If significant amounts of debris collect at the debris filters, the flow resistance at the core inlet may increase, challenging the adequacy of LTCC. | |||
The fluid initially present in the RCS is borated water. Injected emergency coolant is also borated water. Coolant flows out of the break and containment spray interacts with plant materials as it is transported through the containment to the sump, creating a pool with additional dissolved species. Chemical effects include the formation of precipitates and other | |||
chemical products that could be filtered by a fibrous debris bed and further impair the flow of coolant needed for core cooling. An additional concern in the core region for CLBs is the potential for boric acid to concentrate due to boiling associated with the decay heat of the fuel, potentially resulting in BAP. A key mitigating measure is the use of HL switchover (HLSO) (or equivalent) after a period of time to flush the core by injecting coolant into the HL, rather than the CL. HLSO (or equivalent) is required by all PWR Emergency Operating Procedures with its timing dependent on plant-specific evaluations. | |||
WCAP-16793 stated that the following criteria must be met to ensure that debris in the core will not adversely affect LTCC: | |||
* Adequate flow for decay heat removal is supplied to the core | |||
* BAP is prevented The staff agrees that if these are met, LTCC will be successful. | |||
The method used by the staff to evaluate this issue includes the premise that it is unlikely a LOCA large enough to generate and transport enough debris to the core to prevent LTCC will occur. The evaluation considered the low probability of the initiating event in combination with the low likelihood of the formation of a debris bed across the entire core. Sections 5.2.1 and 5.5.2 discuss these parameters. | |||
The NRC staff also has significant knowledge of how the system will respond to varying amounts of debris (or resistance to flow) at the core inlet, including complete blockage at the fuel inlets. The knowledge base for these conditions consists of test and analytical information. | |||
In some cases, the test results were used as inputs to TH analyses. Section 5.2.1 discusses this information. | |||
Table 1 below lists some of the major phenomena associated with the evaluation of IVDEs. The table is designed to demonstrate the differences or margins between the assumptions used in the safety significance evaluation and the expected realistic behavior. To the extent possible, the phenomena are listed sequentially as a LOCA scenario would progress in the plant. Many of the phenomena overlap in time. | |||
Table 1: LOCA Phenomena Sequence of Attribute Assumption Most Probable Margin Events Initial LOCA Size 6-inch <6-inch Less debris is likely Recovery More debris settles Recirculation before reaching sump, Switchover 20 minutes >30 minutes less debris, lower decay Time heat Unqualified Time Some fail as chips, less Coating 100% Fine <100% | |||
dependent debris transports Failure Switchover Some fines and Early Debris 100% Fines trapped, including Recirculation | |||
<100% transports Transport Transport debris recirculating from CS and break Fiber <15%, decreases Depends on plant 15% | |||
Penetration with debris amount design Core Inlet More debris required Uniform Non-Uniform Fiber Bed to block core Conservative Higher flow, flow, particulate FA Debris different Debris limits from amount, Early Testing particulate testing are conservative particulate Recirculation characteristics sizing to Long-TH Analysis Assume Slow increase in Additional Term for Debris instantaneous core inlet conservatism in debris Limit blockage resistance limits Core inlet unblocked at Chemical fiber limit w/o chemical Effects 6 Hours 24 Hours effects, lower decay Timing heat when AFP is sole Long-Term flowpath Effects Hot Leg Consider actual plant 24 Hours Plant specific, most Switchover timing, bypasses debris Maximum less than 10 hours Timing bed | |||
5.1 System Description This description applies to all LOCA locations. There are two separate categories for LOCA break locations depending on whether the break is upstream or downstream of the core (CL side or HL side). The break locations are called HLBs and CLBs. Most plant responses are similar for both categories. Because of differences in plant response to breaks on the hot and cold side of the RCS, HLBs and CLBs are evaluated separately. The system responses to HLBs and CLBs are primarily described in Section 5.2. However, aspects of the LOCA response are discussed throughout this document. | |||
The initial plant response for LOCAs is similar regardless of whether they occur on the HL or CL side of the RCS. Following a large LOCA, the CSS is actuated to suppress containment building pressure, and the ECCS is actuated to replenish liquid in the core and remove decay heat. Some breaks, depending on the break size and the plant setpoint, will not initiate CSS. | |||
Initially, the water source is from stored locations (e.g., the RWST). The borated water from these stored locations refloods the core within a few minutes, decreasing the cladding temperatures to within a few degrees of saturation. Water that is pumped into the RV is discharged through the break into containment, where it collects on the containment building basement floor and in the ECCS sump(s). The time required for the stored water to be depleted depends on the size of the break, the volume of fluid in the tank, and the ECCS and CSS pump capacities. This coolant injection phase can end as early as 20 minutes following the initiating event, if maximum ECCS and CSS flow and minimal stored water volumes are assumed. | |||
After the stored water supply is exhausted, the CSS and ECCS are realigned to draw coolant from the containment sump. The coolant discharged from the RCS and CSS is recirculated into the RCS to provide for continued LTCC without the need for additional cooling water. Most plant responses are similar for HLBs and CLBs, however, Westinghouse 2-loop plants have a significantly different ECCS design known as UPI. Section 5.2.1 describes the responses of the different plant types to breaks in different locations. | |||
Debris-laden coolant may begin to enter the RCS once the ECCS suction source is switched from the RWST to the containment sump. The amount of debris that enters the RCS depends on the filtering efficiency of the strainer, debris generation amounts, ECCS configuration, break location, amount of flow through the CSS, and the amount of time coolant is being drawn from the sump. Some debris that enters the RCS will reach the RV. | |||
During CL recirculation, the flow transports debris through the downcomer (DC) into the lower plenum (LP) and into the core through the core inlet support region. After recirculation is switched to the HL, flow transports debris into the UP and into the top of the core. Depending on the break location and the injection location, some debris may flow out of the break. | |||
Debris may collect at the bottom nozzle (BN) of the FA, causing resistance to flow. In some cases, it may collect at spacer grids higher in the FA. Ultimately, if enough debris is collected on the BNs, flow through the core inlet could be impeded, compromising LTCC. | |||
All U.S. PWRs have AFPs in the RV that provide a path for coolant to reach the core if the core inlet becomes completely blocked with debris. Examples of AFPs include the barrel baffle (BB) region, pressure relief holes, upper head spray nozzles, and flow over the top of the steam generator (SG) tubes. The types of AFPs and resultant amount of flow area they provide is plant-specific. | |||
5.1.1 Key Parameters The NRC staff reviewed the effects of debris on LTCC with respect to debris in the RCS for the current operating fleet of PWRs in the U.S. The review is based on inputs in WCAP-17788, information provided by licensees, and other information available. It is considered representative of US PWRs, if not necessarily bounding. The following are the key parameters considered by the staff as they evaluated the general safety significance of the issues. | |||
Consideration of these parameters may be useful as licensees and staff evaluate plant-specific compliance. | |||
(1) Plants have a method of diluting the coolant in the core by implementing HLSO or equivalent action. This addresses the potential for BAP and concentration of other debris types in the reactor and reduces the potential for chemical effects to result in complete blockage due to a debris bed at the core inlet. | |||
(a) To allow the [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx] CLB limit, plants other than UPI initiate HLSO before chemicals may form as per Volume 5 of WCAP-17788 (14). | |||
See Section 5.2.1.5. | |||
(b) To prevent excessive concentration of particulate debris in the core, HLSO is initiated within 24 hours. See Section 5.2.1.1.3. | |||
(c) To prevent BAP, plants initiate HLSO in accordance with their licensing basis. | |||
See Section 5.2.1.1.2. | |||
(2) Chemical effects timing (tchem) relative to the earliest time that complete core blockage can occur while not compromising LTCC (tblock). See Sections 5.2.1.5 and 5.3.3.3. | |||
(3) Plant-specific post-LOCA environment is consistent with WCAP-17788 Test Group. See Section 5.2.1.5. | |||
(4) Earliest time to swap to recirculation. See Section 5.3.3.2. | |||
(5) Particulate debris amounts reaching the core pertaining to CLB. See Section 5.2.1.1.5. | |||
(6) Fibrous debris amounts reaching the core pertaining to CLB and HLB. See Sections 5.2.1.1, 5.2.1.2, and 5.3.3.6. | |||
(7) Minimum ECCS flow. See Section 5.3.3.5. | |||
(8) FA debris capture characteristics as evaluated in WCAP-17788. | |||
(9) AFP resistance values. See Section 5.3.3.2. | |||
(10) Available driving head for the CLB. See Section 5.2.1.1.1. | |||
(11) RCS liquid mass. See Section 5.2.1.1.5. | |||
(12) LOCADM results. See Section 5.2.1.5.4. | |||
(13) LOCA frequency. See Section 5.5.2. | |||
These key parameters are considered essential considerations for the NRC staffs safety significance determination. Some of these parameters, or others not listed here, may be important to individual licensees LTCC adequacy evaluations. | |||
5.2 Sequence of Events This section describes phenomena that apply to all LOCA scenarios. It also describes how HLBs and CLBs affect the system response for typical PWR designs and UPI designs. Finally, this section provides the basis for the staff conclusion that it is unlikely the primary coolant flowpath (through the fuel inlet nozzles) will become blocked to an extent that would compromise the adequacy of LTCC. | |||
The break size and location govern the characteristics of the initial system response, as well as the amount of debris generated. These must be considered in the IVDE evaluation. | |||
The IVDEs associated with larger breaks are assumed to be more challenging to mitigate than smaller breaks. A complete, double-ended guillotine break (DEGB) would have the highest energy discharge and the greatest potential to generate the most debris. In addition, the coolant pressure and inventory are lost most quickly following the largest break, and the source of pumped injection flow is hence depleted the most quickly, leading to the earliest SSO time. | |||
These dynamics lead to a scenario in which recirculating coolant would have the greatest concentration of debris, and the debris-laden coolant would be delivered to the core at the earliest time, while the core decay heat levels remain high. The resulting scenario would require the greatest coolant flow rate, yet the high concentration of debris would have the highest likelihood of causing a blockage that could inhibit adequate core cooling. | |||
Smaller breaks, by comparison, would tend to generate less debris and deplete the source of injectable coolant more slowly. As a result, the recirculation would begin later, and would feature a lower concentration of debris. Longer times to recirculation also provide additional time for debris to settle in the post-LOCA sump pool. | |||
Based on the evaluation described above, the NRC staff determined that the following assumptions were reasonable to use when evaluating the safety significance of IVDEs: | |||
* Break size and SSO time are inversely correlated; larger breaks would be associated with faster switchover to recirculation, and smaller breaks would allow for SSO when the decay heat has dissipated more. | |||
* Break size and debris generation are directly correlated; larger breaks discharge coolant at a faster rate resulting in more damaging jets and greater damage to materials in containment. | |||
5.2.1 Availability of Emergency Core Cooling System Primary Flowpath The analyses in WCAP-17788, Volume 3 (15) and Volume 4 (16), as well as the earlier work documented in WCAP-16793-NP-A, conservatively assumed that debris would collect uniformly at the core inlet because that results in the greatest headloss for a given amount of debris. | |||
However, the debris bed is realistically expected to collect non-uniformly for reasons described below. As a result, the amount of debris required to completely block the core inlet would be greater than that assumed in the analyses. | |||
If a debris bed does not form at the core inlet or other locations that affect cooling, current LTCC analyses remain applicable. However, based on testing (WCAP-16793-NP-A (1) and WCAP-17788 (2)), it was determined that a fiber bed could form at FA inlets. This testing was | |||
conducted with a single FA or portion of an FA and assumed uniform flow into all areas of the core. | |||
WCAP-17788 established debris limits of [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. The HLB debris limit is based on subscale FA testing and the TH analyses performed for WCAP-17788. The CLB limit applies to all RCS designs. The HLB fiber amount referenced in this section is based on the combined RCS and fuel design with the smallest debris limit. These debris limits assume that the debris bed forms uniformly as it did during reduced-scale FA testing performed as part of the WCAP-17788 analyses. | |||
The PWROG TH analysis and the NRC TRACE evaluations (See Section 5.4) predict significant variations in flow rate and direction at the core inlet, which will prevent the deposition of a uniform debris bed. This makes complete blockage of the core inlet less likely. The flow in the center of the core will be up (from the LP to the core), while the flow at the periphery will be down, (from the core to the LP). The patterns are caused by higher decay heat in the central core and relatively low decay heat in the periphery. Contributing to this flow pattern is the downward flow entering the LP from the DC. These flow patterns will cause the deposition of fiber at the center of the core inlet first. As debris builds in the areas with upward flow, some flow redistribution to the periphery may occur. | |||
Other influences on flow will tend to prevent uniform flow to the core. For example, the lower core plate, core support plate, and other hardware within the LP disturb the flow patterns at the core inlet. These contribute to non-uniform arrival of flow and debris at the core inlet. This phenomenon was observed during brine testing (PWROG-15091, Subscale Brine Test Program Report, dated November 30, 2015 (6)) in which it was difficult to form a uniform bed at the FA inlet. | |||
The NRC staff observed similar non-uniform debris bed formation during several strainer tests, which provides evidence that similar behavior will occur at the core inlet (17) (18) (19). The velocities approaching the disks in most strainers are non-uniform because the plenum that collects the flow that passes through the perforated disks is hydraulically closer to some surfaces than others. The velocities closer to the plenums are higher. The velocity profile determines the degree of non-uniformity. Greater disruptive forces are present for the in-vessel case. Some of the flow is down, or away from the core inlet. This prevents deposition of debris in those areas. The flow velocity at the core inlet can vary considerably depending on the ECCS design and the break location. CLB flow velocities are low, while HLB velocities can be much higher if all pumps are injecting. CLB average core inlet velocities are similar to strainer approach velocities. HLB velocities are generally higher. | |||
The staff considered that chemical precipitates in the debris bed could cause the bed to deposit more uniformly over the core inlet by increasing the differential pressure across the bed and causing flow to be redirected. However, as discussed in Section 5.2.1.5, precipitation is expected to occur later in the transient. By the time precipitates may form, the debris will be depleted from the pool so that further transport to the core inlet is significantly reduced. | |||
Testing for WCAP-17788 indicates that the debris deposited at the core inlet would not cause a significant resistance to flow until the fiber limit is approached. This is consistent with testing from WCAP-16793 that showed there was a threshold fiber amount below which significant headloss would not occur. Therefore, significant redistribution of the debris will not occur early in the scenario. The design of the PWR core also permits crossflow between FAs which allows | |||
some equalization of the differential pressure in the areas where debris is deposited under the low-velocity conditions following a LOCA. | |||
When considering a full core relative to the subscale testing, it is expected to take more debris than the limits in WCAP-17788 to prevent coolant from reaching the core. In the event the core inlet accumulates enough debris to stop flow, AFPs are available to provide adequate core cooling. Section 5.3.3 discusses these AFPs. The debris limits assume that chemical effects have not occurred. Section 5.2.1.5 discusses chemical effects. In addition, once a plant initiates HL recirculation or a similar BAP prevention action, adequate LTCC is maintained. | |||
HL recirculation is also likely to dislodge portions of the debris that may collect at the core inlet. | |||
The NRC staff considers the assumption that all debris passing through the strainer reaches the core inlet to be conservative. For breaks large enough to initiate the CSS, the amount of fiber reaching the core inlet would be reduced by the amount of fiber diverted to the CSS. For smaller breaks, the CSS may not initiate; however, the overall amount of debris generated would be less. | |||
Evaluations conducted by the PWROG and the NRC demonstrated that even a very small open flow area at the core inlet will allow adequate coolant to enter the core. One evaluation conducted for WCAP-16793-NP-A (1), Appendix B, showed that up to 99.4 percent of the inlet could be blocked and flow to ensure core cooling would still be adequate. Calculations performed by the NRC Office of Nuclear Regulatory Research (RES) confirmed this evaluation (20). Later, more detailed TH evaluations performed by RES showed that 99 percent of the flow area could be blocked for HLBs and CLBs while ensuring adequate flow and prevention of BAP (21). This 99 percent flow area was converted to a core inlet resistance and applied uniformly over the core inlet area in the RES model. | |||
An additional source of information regarding debris bed formation at the core inlet is testing conducted for WCAP-17360-P, Small Scale Unbuffered and Buffered Boric Acid Nucleate Boiling Heat Transfer Tests with Sump Debris in a Vertical 3x3 Rod Bundle, dated May 31, 2012 (5). This WCAP discusses 3x3 heated rod bundle testing that was conducted with fibrous and particulate debris. The BN and first spacer grid provided the most likely places for debris to be captured. Capture of fibrous debris at spacer grids was observed during testing for WCAP-16793. During the heated rod testing, there was no debris capture at the spacer grids except for a small amount of debris captured between the outer surface of the grid and the test column. Although the testing did not include a prototypical geometry at the approach to the fuel, the test results suggest that turbulence may help to inhibit debris bed formation. The heated rod testing provided qualitative evidence that turbulence will tend to prevent debris bed formation, especially at spacer grids. | |||
As part of a DID analysis, the NRC staff considered how the LOCA response would evolve if a high-resistance debris bed were to form at the core inlet. Section 5.3.3 discusses the availability of AFPs that bypass the normal core inlet. | |||
5.2.1.1 Cold-Leg Break Following a CLB, for all PWR designs except UPI, the ECCS is injected into the CL and flows down the DC into the LP and up into the core. Only the flow necessary to make up for inventory in the core region lost to boil-off enters the core. The rest of the flow goes out of the break in the CL and spills back to the sump. The amount of debris reaching the core inlet for the CLB is | |||
much less than for the HLB because most of the debris flows out the break. For the CLB, the force driving the coolant into the core is the manometric balance between: (1) the level in the CL (up to the break elevation) and DC, and (2) the level in the core. As the water boils out of the core it is replaced because of the gravity head in the DC. There are also flow losses from liquid and vapor flows. The amount of coolant required to replace the boil-off is higher at the beginning of the event and is reduced as decay heat decreases. Flow to the CSS does not enter the core and is returned to the sump with any entrained debris. | |||
For a CLB, when injection is switched to the HL, the discharge of the ECCS pumps is aligned to the UP (with pressure limited by the height of the SG tubes). This flow will reverse the flow at the core inlet, flush concentrated debris and boric acid from the core, and disrupt any debris bed at the core inlet. | |||
The NRC staff determined that it is unlikely for a debris bed to form across the entire core inlet following a CLB. Circulating currents will prevent or limit debris deposition at the core periphery. | |||
Maintaining a small open area at the core inlet will assure adequate flow to cool the core. The concepts outlined here are the same as those discussed in Section 5.2.1 and are not repeated in detail. | |||
The PWROG submitted an evaluation of the effects of debris on a CLB in WCAP-17788 and updated the evaluation with supplemental information. This establishes a basis for a debris limit for the CLB case. In the analysis, the AFP credited in the HLB was assumed to be unavailable, and all flow to the core was assumed to flow through the core inlet. The CLB debris limit was set at [xxxxxxx]. | |||
Boiling in the core is vigorous at the beginning of recirculation, which results in sufficient liquid carryover to limit the buildup of boric acid in the RV and adequate mixing in the RV to maintain homogeneous boron concentration. The WCAP-17788 TH analysis shows upflow in the center of the core and downflow in the periphery. TRACE studies predict similar flow patterns at the core inlet (see Section 5.4). The circulating currents result in the transport of boron from the core region to the LP. The boron concentration in the core may increase until the density of the fluid in the core is high enough to result in density driven convection between the core and LP. | |||
As boiling decreases, convection can become the dominant mixing mechanism. If the boron concentration in the RV becomes too high, boron precipitation could occur which would affect the ability to cool the fuel. The TRACE studies also predict that mixing between the core and LP will occur. The LP provides a volume for dilution of boric acid and other debris that may be suspended in the core. | |||
Debris will begin to arrive at the core inlet shortly after recirculation begins and most debris will reach the core inlet within a few hours. Most will arrive relatively early because the strainer will become more efficient at filtering over time as debris collects on it. Debris will also be removed because of collection in the core or settling elsewhere throughout the sump or RCS. | |||
The following topics are addressed to evaluate the fiber limit of [xxxxxxx]. | |||
* debris collection at the core inlet | |||
- resulting in increased pressure losses | |||
- preventing transport of high-concentration boron solution from the core to the LP | |||
* debris suspended in the liquid phase | |||
- increasing pressure losses through the inner RV because of changes in fluid properties and two-phase flow characteristics and changes in heat transfer processes | |||
- causing early BAP because of reduced liquid inventory, reduced liquid carryover, or changes to the precipitation mode or precipitation location | |||
* debris deposition on spacer grids in the core increasing frictional losses or reducing heat transfer Note that UPI plants are treated differently because of the location of the ECCS injection. | |||
Section 5.2.1.3 discusses UPI plants. | |||
A greater amount of debris generation is required to inhibit flow for a CLB than for a HLB. This is based on the distribution of flow after it passes through the strainer. For a HLB, all ECCS is pumped through the core. For a CLB, only the volume of water required to replace that boiled out of the core flows to the core inlet. The remainder of the flow goes out the break, and with it any debris that is entrained in the flow. The debris that flows out of the break is returned to the sump pool and has an opportunity to enter the ECCS again. However, testing has shown that the filtration efficiency of the strainer for fiber increases quickly as fiber is captured on the strainer surface. | |||
As an example, South Texas Project (STP) performed studies to determine how much debris could arrive at the core inlet under various conditions. STP has among the highest fiber amounts in containment compared to the other U.S. PWRs because it uses only fibrous insulation in its containment. The studies were based on strainer penetration test results and modeling of the flow as it circulated through the strainer, the ECCS, and the CSS. The model found that for a design basis case with a single failure (one train of ECCS and CSS running), | |||
376 grams of fiber would arrive at the core inlet. Without CSS (only low-head safety injection and high-head safety injection running), the value was 539 grams. For a single low-head safety injection pump running using the most conservative penetration results, the debris amount arriving at the core was 1,119 grams. This amount of debris is assumed to be evenly distributed among 193 FAs. The FA loading for the single low-head safety injection case is very low at 5.8 g/FA. The NRC staff noted that it is unlikely that only a single low-head safety injection pump would be operating during recirculation, but even if that is the case, the debris loading is well below the level of concern. Pumps may be secured to reduce flow through the strainers, but this is generally done at a time when most of the debris has been removed from the sump fluid by filtering. As can be seen above, the more realistic cases have lower debris amounts. | |||
STP also performed beyond design basis sensitivity studies to determine the most challenging flow configuration for CLB in-vessel debris. STP established a strainer debris limit of 192 pounds of fine fiber based on strainer headloss testing. LOCAs that resulted in greater than 192 pounds of fine fiber at the strainer were assumed to contribute to core damage. As shown in the analysis for the license amendment request (13), 192 pounds of fine fiber could result from a 12.8-inch DEGB. However, most breaks generating this much fiber were significantly larger. A few of the largest breaks at STP could result in fine fiber amounts of up to 830 pounds. | |||
The beyond design basis sensitivity studies were based on fine fiber debris loads at the strainer of 192 pounds (design basis limit) and 800 pounds. The 800-pound value is representative of the amount of fine fiber that can arrive at the strainer from the largest LOCA breaks at STP. | |||
The sensitivity studies demonstrated that even with the largest strainer fiber loading and overly conservative flow conditions, the CLB limit was not reached for STP. It was determined that lower strainer flow rates resulted in higher core debris amounts. For example, the highest debris amount was calculated assuming one-half high-head safety injection flow and no other pumps in service. The result was 2,400 grams of fiber at the core. For STP, this results in 12.4 g/FA. It was further determined that the highest core loading would occur with the largest number of strainers in service with the lowest possible flow through each. | |||
Vogtle Electric Generating Plant performed a similar evaluation for the CLB, also using plant-specific penetration testing and modeling. Vogtle is also a very high-fiber plant. Instead of using a plant-specific debris amount in the testing, a theoretical strainer debris bed thickness of one inch was attained for each test by batching in small amounts of fiber. Testing demonstrated that additional debris did not penetrate the strainer, even before the one-inch threshold was reached. Most of the fiber penetrated the strainer early in each test when the debris load was low. Vogtle calculated that 13 g/FA of fiber would arrive at the core inlet for the CLB under the most limiting design basis conditions for pump and strainer operating combinations. As a comparison, for the Vogtle HLB case, 91 g/FA of fiber was calculated to reach the core. Note that STP did not calculate a HLB fiber amount, but conservatively assumed that the core was completely blocked relatively early in the event. | |||
Plants have different design configurations that will affect the amount of fiber that may reach the core following a CLB. However, the STP study and the Vogtle results demonstrate that small breaks do not generate enough debris to inhibit LTCC following a CLB. These examples provide evidence that even the largest CLBs in high-fiber plants are unlikely to result in greater than [xxxxxxx] fibrous debris at the core inlet. Also, because debris deposition at the core inlet will be non-uniform, it is very unlikely that debris at the core inlet will inhibit LTCC following a CLB. In conclusion, even the largest DEGB is unlikely to generate enough debris to approach the [xxxxxxx] CLB limit presented in WCAP-17788 and a large CLB is predicted to occur very infrequently (see Section 5.5.2). | |||
5.2.1.1.1 Debris Collection at the Core Inlet The fluid velocity profile across the core inlet varies depending on the relative power of the fuel. | |||
The flow patterns at the core inlet will deposit debris close to the center of the core but not at the periphery. Boiling and circulating flows will prevent a uniform debris bed from forming across the entire core inlet for the debris amounts determined by testing. Therefore, mixing between the core and LP are unlikely to be inhibited during the LOCA recovery (see Section 5.2.1). Even with an increase in density in the core region, which was not modeled by industry analyses, the recirculating currents will continue as decay heat decreases and boron concentrates in the core region. This conclusion is supported by TRACE simulations (21) that included the effects of boric acid concentration in the core. Section 5.4 discusses TRACE calculations. | |||
The TH analysis in WCAP-17788 becomes unreliable once decay heat decreases to the point where boron concentration in the core increases because of limitations of the code in accommodating buoyancy-driven flow. The TH analysis shows that the steam quality exiting the | |||
RV remains below 0.8 before the switchover to sump recirculation and that this ensures that boron solutes will not concentrate in the core up to that point. | |||
The driving force for water entering the core is the manometric balance between the DC region of the RV and the core region. The head from the DC must overcome liquid and vapor flow losses on the core side as well as any headloss caused by debris and density difference resulting from concentration of boric acid or other debris in the core region. | |||
The SE for WCAP-16793-NP-A found that at less than 15 g/FA, flow to the core would continue for the HLB condition, even if chemical precipitates were present. For the CLB, the staff found that LTCC would be assured for plants that meet the HLB criterion of 15 g/FA, but this finding did not address the effects of BAP. The staff finding also assumed that the maximum fibrous debris load arriving at the core inlet for the CLB would be roughly half of the HLB amount. | |||
For this TER, it is assumed that chemical effects, discussed in Section 5.2.1.5, are delayed until after the licensee implements HLSO (or equivalent). WCAP-16793-NP-A testing found that under CLB flows, the deposition of fibrous and particulate debris at the FA inlet resulted in very low headloss. These headlosses were found to be much lower than those occurring when chemical precipitates were included. The CLB testing for WCAP-16793-NP-A (see WCAP-17057-P, Revision 1, GSI-191 Fuel Assembly Test Report for PWROG, issued September 2011 (22)) included several tests with a fiber loading of [xxxxxxx] and varying amounts of particulate. The tests included chemical precipitates. During the testing, headlosses were allowed to stabilize and were recorded before chemical debris was added to the test. The highest non-chemical headloss recorded during the testing was [xxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. The PWROG stated that the available driving head for all plants is greater than this value. [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. Information provided by the PWROG indicates that the driving head calculation was based on a reactor design that has 14-foot FAs while all other designs have 12-foot FAs. The DC for the example plant was stated to be similar to the designs that have 12-foot FAs. This indicates that the calculation has about 2 feet of margin for most plants (approximately 1 psi). | |||
The PWROG also performed small scale boiling rod bundle tests that included fibrous and particulate debris. These tests were performed and documented in WCAP-17360 (5). The testing used a vertical 3x3 rod bundle that had dimensions and materials like those of PWR fuel. | |||
The rods were heated to achieve boiling conditions similar to those expected after a large CLB. | |||
During the rod bundle tests no debris bed formed at the fuel inlet. The debris was observed to pass through the BN and settle on top of the nozzle or pass into the heated core region. | |||
WCAP-17360 determined that the boiling conditions downstream of the FA inlet produce instabilities that inhibit debris bed formation. A small amount of fiber was observed to collect at the outer edge of the spacer grids (between the grids and the test column). The rod bundle testing was not conducted with a prototypical FA or LP, so it cannot be definitively relied on as applicable to a full core. However, it does provide evidence that turbulence from boiling will prevent significant amounts of debris from collecting at the spacer grids. | |||
The rod bundle tests demonstrated that density-driven flow occurs in the boiling condition through observation of temperature stratification in pure water tests and mixing of fluid, as determined by relative temperature between the simulated core and LP, for tests that included boric acid. Additional evidence of mixing was obtained by sampling the water in the simulated | |||
LP. Maintaining the LP as a mixing volume reduces the concentration of boric acid and other debris in the core making BAP less likely and delaying its onset. | |||
Previous studies performed by the PWROG for WCAP-16793 and RES showed that for the CLB, a single open FA provides adequate coolant to maintain LTCC. | |||
The following events are each unlikely as discussed above: | |||
(1) Large CLB LOCA (2) Transport of fibrous debris more than the analyzed limit (based on a uniform debris bed) | |||
(3) Formation of a uniform debris bed Considering the above, it is very unlikely that a high resistance debris bed will form at the core inlet such that flow into the core will be prevented. Therefore, the potential for a loss of LTCC because of debris in the core is of low safety significance for the CLB. The initiation of HL injection or an equivalent action is assumed before chemical effects may occur. HL injection bypasses the debris bed at the core inlet before headlosses resulting from chemical effects exceed the available driving head and flushes concentrated debris and boric acid from the core. | |||
5.2.1.1.2 Boric Acid Precipitation The PWROG performed SKBOR studies to estimate the effects of debris buildup at the core inlet on BAP. The simulations reflected conditions typical of design-basis safety analyses. A case was performed as a baseline assuming no core blockage and 50 percent mixing with the LP. Sensitivity cases assumed 100 percent mixing availability with the LP but assumed percentages of core inlet blockage varying from 0 to 75 percent. The study found that the cases with 0, 25, and 50 percent blockages took longer to reach the boric acid solubility limit than the baseline. The 75 percent blockage case reached the BAP limit faster than the baseline. | |||
Therefore, it was concluded that greater blockage leads to delays in the exchange flow between the core and LP. | |||
The study demonstrated that up to 50 percent of the core inlet can be blocked with debris before the flow exchange between the core and LP will be inhibited such that the licensing basis assumption of crediting 50 percent of the LP volume becomes questionable. The SKBOR studies are conservative because of the application of conditions used in design basis safety analyses. Based on the SKBOR studies, WCAP-17788 states that a relatively low steam quality (less than 80 percent) allows a significant amount of the boron to be removed from the core via entrainment and that this will delay the onset of BAP compared to most of the SKBOR cases in the study. The PWROG performed an SKBOR study that demonstrated a steam quality of 0.95 would result in significant boron concentration in the core, but that the concentration would never exceed the precipitation limit. Higher steam qualities resulted in eventual precipitation. | |||
These studies assumed that 50 percent of the LP was available for dilution of the boron. | |||
The findings from the SKBOR studies are consistent with the trends found in TRACE studies of BAP completed by RES (21). The TRACE studies are discussed in Section 5.4. TRACE found that BAP would be delayed significantly longer than in the SKBOR studies conducted by the PWROG. | |||
The PWROG conducted a brine test program described in PWROG-15091-P, Subscale Brine Test Program Report, dated November 30, 2015 (6). The testing was conducted to learn more about density-driven mass transport between the core and the LP with and without debris present at the FA inlet. The test program demonstrated the fluid exchange and showed that highly resistive debris beds cannot form at the core inlet at the established CLB in-vessel fiber debris limit. The program did not consider chemical effects because it was assumed that HL recirculation would be initiated before chemical precipitation. Testing included conditions expected to occur following a CLB and considered both Westinghouse and AREVA core inlet geometries. | |||
Fibrous debris ranges for the testing represented 2.5 to 22.5 g/FA. Some of the tests included particulate, but it was ultimately determined that the amount of particulate included did not significantly affect the test results or differential pressure across the debris bed for the tested conditions. The results from the brine testing were consistent with the subscale headloss testing that found that [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxx]. | |||
The brine testing demonstrated that the exchange flow rate (between the simulated core and LP volumes) was influenced by the amount of fiber included in the test. Larger fiber amounts blocked more open areas in the BN. Under the test conditions, the relative timing of the debris injection, the flow velocity into the FA, and the increase in density of fluid in the core region affected the ability of a debris bed to form. If a debris bed formed, the exchange between the modeled core and LP regions stopped but was reestablished when the density of the fluid in the core increased or when the flow from the LP to the core decreased and the debris bed was disrupted. The brine testing provides evidence that exchange flow can take place, even if a relatively low resistance debris bed forms at the core inlet. | |||
The NRC staff noted that the brine test results and the results of FA testing for WCAP-16793-NP-A had different headloss values for similar fibrous debris loads. The PWROG stated that the results from the brine testing and other FA tests had different headlosses because the test methods were different. It was noted that higher in-bed particulate loads can result in higher headlosses and that the tests being compared might not have similar particulate loads because the WCAP-16793-NP-A tests used a recirculating flowpath while the brine testing used a once-through system. Some of the particulate in the brine testing might have passed through the fuel inlet and avoided capture in the debris bed. The PWROG also noted that both test programs support the assertion that a [xxxxxxx] limit will not result in a significant headloss at the core inlet. | |||
The staff concluded that the headlosses that may occur at debris loads below the fiber limit are relatively low before the arrival of chemical precipitates and that both test programs demonstrated this. | |||
Based on the SKBOR studies and the brine testing discussed in this section, and the TRACE results presented in Section 5.4, the NRC staff determined that debris effects do not invalidate the current licensing bases or HLSO times for operating PWRs. The analyses indicate that BAP may be of lower risk than was previously thought. However, changes to a plants licensing basis that draw on these insights would have to be justified on a plant-specific basis. In addition, maintaining HLSO (or equivalent) is important to ensure continued LTCC by flushing concentrated debris from the core and to decrease the probability that chemical precipitates cause a significant increase in headloss. | |||
5.2.1.1.3 Debris Suspended in the Core The effects of debris in the core for the HLB are expected to be bounded by those that will occur during the CLB. This section includes some discussion of HLB behavior, but the behavior following a HLB should be similar to that following a CLB, except for the amount of debris present. | |||
Debris that passes through the core inlet or enters through an AFP will reach the core and UP. | |||
Debris in the core can affect fluid and heat transfer properties of the fluid. Because the CLB initially has the core in a boiling pot mode, any debris that reaches the heated core is assumed to stay there. For the HLB, much of the debris is flushed from the core making the CLB limiting. | |||
For the CLB, debris can concentrate in the core, similar to the way boric acid can. Most fibrous debris will reach the core in the first few hours of the event, and most particulate debris will arrive much later because most transportable particulate debris is failed unqualified coatings. | |||
The amount of fibrous debris reaching the core is much less than the amount of particulate debris that may reach the core. For the HLB, the amount of fiber that reaches the core may be greater than for the CLB. However, before significant core blockage, some of the fibrous debris will be flushed from the core. The amount of fibrous debris that may reach the core is an order of magnitude less than the potential particulate debris. | |||
Electric Power Research Institute (EPRI) testing, described in EPRI 1011753, Design Basis Accident Testing of Pressurized Water Reactor Unqualified Original Equipment Manufacturer Coatings, dated September 30, 2005 (23), demonstrated that most unqualified coatings will fail after 24 hours. Early in the event, boiling maintains turbulence in the core and keeps debris well mixed. After a few hours, depending on plant-specific procedures, plants initiate actions, like HLSO (or equivalent), to flush concentrated boric acid from the core. This action will also flush debris from the core. Because most of the debris that may enter the core following a CLB cannot arrive until after HLSO (or equivalent), the amount of debris in the core is relatively small, minimizing the effects of debris on fluid and heat transfer properties. Section 5.2.1.1.5 discusses in more detail coatings failures, particulate debris, coatings failure timing, and debris characteristics. | |||
Rod bundle testing demonstrated that debris in the heated region remained well mixed with the fluid. [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. Some deposition of debris at the outer edge of the spacer grids was observed. Similar behavior is expected for the HLB. The TH analyses discussed above demonstrate that circulating currents are present following a CLB. The currents assist with keeping debris and boric acid well mixed in the core. | |||
Debris suspended in the core may have an influence on fluid properties and displace liquid in the core. WCAP-17788 assumed a 5 percent mass limit on the debris in the core. The increase in density from this amount of debris was determined to have a small effect on headloss. The effects on fluid viscosity and surface tension were also determined to be small. The rod bundle testing was referenced to show that bubbly flow observed in the testing was not significantly affected by the addition of debris. | |||
Debris suspended in the core may also affect fluid heat transfer properties. Rod bundle testing validated that no major degradation of heat transfer occurred. Open literature, provided as | |||
references in WCAP-17788, indicates that limited amounts of particulate debris suspended in fluid will improve heat transfer. | |||
Debris in the core could affect BAP by displacing liquid and reducing the liquid volume available for mixing. Based on the small amount of liquid displaced, any changes to the rate of boron buildup in the core will be negligible. The 3x3 rod bundle testing also provided evidence that precipitation modes, and boric acid transport and mixing phenomena, are not negatively affected by debris in the core. | |||
5.2.1.1.4 Local Collection of Debris The rod bundle testing demonstrated that significant collection of fiber at the spacer grids is unlikely. However, should debris become trapped in the spacer grids, the effects would be limited to local areas. WCAP-16793-NP-A, Revision 2 (1) evaluated complete blockage around a fuel rod at a spacer grid and demonstrated that axial conduction would allow the fuel cladding to remain at acceptably low temperatures. In addition, the 3x3 rod bundle testing did not identify any significant collection of debris at the spacer grids, even with a relatively high fibrous debris load equivalent to 150 g/FA. The open lattice fuel design for PWRs permits crossflow between assemblies, which will allow flow around localized blockages. This section applies to HLBs and CLBs. | |||
5.2.1.1.5 Concentration of Particulate Debris The NRC staff evaluated the different categories of coatings and other particulate debris sources to assess the potential effects of particulate debris. The three coating categories evaluated are qualified coatings, degraded qualified coatings, and unqualified coatings. The evaluation of coatings depends on failure timing to demonstrate that an excessive amount of particulate debris will not concentrate in the core following a CLB. For a HLB, the coatings and other particulate debris are not expected to concentrate because there is significant excess flow through the core that will flush debris as it exits the reactor. | |||
Similar to boric acid, debris may concentrate in the core during a CLB because water boils away and particulate debris remains behind. Because there is not significant empirical data for particulate filtration by sump strainers, it is assumed all particulate debris passing through the strainer arrives at the core. In addition, filtering at the core inlet is not assumed. | |||
All qualified, unqualified, and degraded coatings within the ZOI are assumed to fail immediately as fine particulate that is transported to the core. Based on the review of individual plant submittals in response to GL 2004-02, the approximate range for ZOI coatings amounts is 100 to 2,000 pounds. Qualified coatings outside the ZOI are assumed to remain intact and attached to the substrate. They do not contribute to the debris source term at any time during the LOCA recovery. | |||
Unqualified coatings are assumed to fail as particulate debris. This category of coating is generally the largest contributor to the fine particulate debris source term. Some plants consider degraded qualified coatings to be unqualified. Based on the review of individual plant submittals in response to GL 2004-02, the amount of unqualified coating debris ranges from less than 100 pounds to 20,000 pounds. Based on the EPRI testing discussed below, it is estimated that less than 10 percent of the unqualified coatings will fail before 24 hours. | |||
Therefore, it is assumed that 10 percent of the unqualified coatings are available for transport to | |||
the core before the initiation of HLSO (or equivalent), which occurs within 24 hours after the initiating LOCA. Once HLSO (or equivalent) is initiated, the core will be flushed of concentrated boric acid and particulate debris. | |||
Degraded qualified epoxy coatings outside the ZOI are assumed to fail as chips. The chips are assumed to settle or be captured by the strainer and not be available to transport to the core. | |||
Degraded qualified inorganic zinc outside the ZOI is assumed to fail in particulate form. This particulate is assumed to be available to transport to the core, but the degradation of this coating category into particulate is estimated to take longer than 24 hours so it would not be transported to the core before HLSO. Many plants did not differentiate between degraded qualified coatings and unqualified coatings in their analyses. For these plants, the unqualified coating term is conservative with respect to the amount of particulate debris that may pass downstream of the strainer. The basis for the assumed debris characteristics of the degraded qualified coatings is testing performed by Keeler & Long. (24) | |||
Unqualified coatings make the most significant contribution to amounts of particulate debris in the core. To evaluate the potential effects of particulate debris in the heated core, the NRC staff referred to an EPRI report that discusses the behavior of unqualified coatings when exposed to post-LOCA conditions (23). | |||
In addition to coatings, particulate debris can originate from latent debris and particulate type insulations. According to NRC guidance, the assumption for the mass of latent particulate is generally 170 pounds (8). Guidance also recommends plants to confirm their latent debris amounts. Plants identified a range of latent particulate, with the largest amount assumed to be 500 pounds. | |||
Most plants do not have a significant particulate insulation source term. However, one licensee determined that 2,200 pounds of calcium silicate insulation could be generated by a limiting break. This plant has the largest particulate source term when calculated, as discussed below. | |||
To estimate the amount of particulate debris that might transport and collect in the core during a CLB, the NRC staff used information provided by licensees in their responses to GL 2004-02. | |||
The staff summed 100 percent of the ZOI generated coatings, 100 percent of the latent particulate, 100 percent of the particulate insulation debris calculated to reach the strainer, and 10 percent of the total unqualified coatings. A few plants were estimated to have maximum values of particulate that could reach the core between 2,500 and 3,000 pounds. These values are considered realistic for the worst breaks, and therefore conservative for other breaks. The largest calculated particulate amount calculated by the staff was 2,910 pounds. | |||
The staff assumed that 2,910 pounds of particulate would be mixed with the coolant in the core and LP regions. Mixing between the regions is discussed in the CLB discussion in Section 5.2.1.1. The amount of liquid in the core and LP is assumed to be 60,480 pounds based on 50 percent voiding in the core and 100 percent of the LP volume available for the Westinghouse 3-loop reactor design. If the entire mass of particulate is mixed in the LP and core region, the mass ratio is 4.8 percent particulate to liquid. A 3-loop Westinghouse design was used to determine the fluid mass because it has smaller volumes for the LP and active core regions than the 4-loop Westinghouse reactor design. The 3-loop and 4-loop Westinghouse designs are considered as representative of the PWR fleet. The 2-loop Westinghouse designs have smaller RCS volumes but use injection into the UP, which reduces the concern with concentration of debris in the core. The staff verified that the maximum particulate mass for the | |||
2-loop plants is less than 1,000 pounds so the mass ratio for these plants is limited to about 2 percent considering the smaller volumes for these plants. | |||
For design-basis BAP calculations, 50 percent of the UP volume, up to the bottom of the HL, is typically credited. This volume was ignored for the particulate debris calculation and provides margin in the calculation. | |||
The debris amounts in the example above are realistic maximum debris generation values for the limiting plant in the fleet. The amounts of debris generated were calculated for the worst-case breaks using accepted methods that are considered conservative. Other breaks, especially smaller, more likely breaks, would result in less debris. Also, a low liquid mass was used in the calculation by choosing a plant design with low volume core and LP regions (except for UPI plants discussed above) and assuming a high void fraction for the core region. | |||
Therefore, the calculated mass ratio is considered a conservatively high value for the fleet. | |||
Almost all breaks would generate significantly less particulate debris. In addition, the amount of particulate assumed to reach the core is based on 100 percent transport. It is likely that some would not reach the strainer or would settle in the LP, be filtered by the strainer, or be filtered at the core inlet. The latent particulate is a relatively small portion of the source term for the plants with a large amount of particulate, but a significant amount of this source would settle because of its larger size. | |||
WCAP-17788 evaluated a limit of 5 percent by mass particulate to liquid in the core. The final evaluation of this issue was included in the Westinghouse response to Request for Additional Information (RAI) 1.3 by letter dated September 28, 2018 (25). The NRC staff found the evaluation to reasonably address the potential issues that could result from particulate suspended in the core. | |||
Once HLSO (or equivalent) is initiated, any debris that has collected in the core will be diluted by mixing the coolant in the reactor back into the sump. Therefore, it is important that plants perform HLSO within 24 hours to prevent excessive unqualified coatings from concentrating in the core. | |||
5.2.1.1.6 Conclusions for the Cold-Leg Break Based on the discussion above, the staff concludes that the [xxxxxxx] fiber limit for the CLB is acceptable and that flow to the core will be maintained with this debris load if chemical precipitates do not deposit or form on the debris bed. The probability of any PWR accumulating | |||
[xxxxxxx] at the core inlet is very low because of the low volume of flow, and therefore the low amount of debris that approaches the core inlet during a CLB. If more than [xxxxxxx] is transported to the core inlet, the formation of a uniform bed is unlikely, and flow to the core will likely continue. The actual debris amount that might result in preventing adequate flow to the core is not defined because testing is based on a uniform debris bed. If the core inlet becomes completely blocked, AFPs will allow flow to reach the core. There is adequate DID to assure LTCC following a CLB for the current operating fleet of PWRs. | |||
Current plant licensing bases for BAP do not need to be revisited because of the effects of debris on LTCC. Analyses performed in WCAP-17788 and TRACE (21) show that boric acid will not concentrate to the point of precipitation in areas important to support cooling. | |||
The collection of debris in the heated core will not significantly affect LTCC. The amount of debris that enters the core following a CLB is limited by the amount of debris available before the initiation of HL injection (or equivalent) and the low amount of flow entering the core. A large amount of particulate debris may be available, but this debris will not become transportable until after HLSO is initiated. This prevents the concentration of significant amounts of debris in the core. Filtering by the strainer also reduces the amount of debris that may enter the core. | |||
Local collection of debris (at spacer grids) will not adversely affect cooling of the fuel as demonstrated by calculations for WCAP-16793-NP-A. | |||
Flow from HL injection (or equivalent) is expected to be initiated prior to BAP. Complete blockage of the core inlet is not expected for plants that initiate HL injection (or equivalent) before the formation of chemical precipitates. | |||
5.2.1.2 Hot-Leg Break Following a HLB LOCA, for all PWR designs except UPI, the ECCS is injected into the CL and flows down the DC, into the LP, and up through the core. All the injected flow passes through the core. All debris that bypasses the strainer and is pumped by the ECCS will arrive at the core inlet and may collect there causing an increased resistance to flow through this path. The driving head available for this scenario is the ECCS pump discharge as limited by the height of the SG tubes. If the core inlet resistance increases significantly because of debris, AFPs will become activated to allow coolant to reach the core and remove decay heat, helping to maintain LTCC. Flow to the CSS does not enter the core and debris in that stream returns to the sump. | |||
After a few hours, PWRs change the injection path to prevent the buildup of boric acid in the core following a CLB. This action is taken following all large LOCAs because the operators may not be able to determine if the break is on the hot or cold side of the RCS. The injection is switched to the HL or a combination of the HL and CL. For a HLB, this injects coolant into the UP where it can drain into the core to provide cooling. Excess flow spills out of the break in the HL and returns to the sump. | |||
To determine the approximate size of a break that could result in significant debris at the core inlet, the NRC staff used information from several sources. First, a debris limit of [xxxxxxxxx] | |||
was assumed. This is taken from the WCAP-17788 analysis for the most conservative fuel and plant design combination and applies only to the HLB scenario. Debris limits for the HLB were calculated to be between [xxxxxxxxxxxxxxx] depending on the fuel and reactor design combination. Some combinations could have accommodated more, but [xxxxxxx] represents the maximum debris amount tested for WCAP-17788. [xxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxx]. | |||
Table 2 of Section 5.3.3.6 lists the HLB fibrous debris limits for the various RCS design and fuel type combinations. Section 5.2.1.1 discusses CLBs. | |||
Although the staff did not accept the WCAP-17788 analysis that determined the fiber limits, the staff considered that the limits were established at limiting flow, particulate size, and particulate to fiber ratios. The staff also compared the fibrous debris limits from WCAP-17788 to test | |||
results from WCAP-16793. The fiber limits from WCAP-17788 testing are lower than those in tests that WCAP-16793 indicate as acceptable. Even with significantly higher fiber amounts included, test results from WCAP-16793 indicate that the WCAP-17788 limits are conservative. | |||
The NRC staff assumed that information from the STP submittal could be used to estimate the HLB size that would generate and transport enough fibrous debris to reach the acceptance limit. | |||
Plants that have less fibrous insulation installed in the containment are of lesser concern and are bounded by plants with large amounts. The choice of STP as the example plant is reasonable because of the large amount of fibrous insulation, the large number of breaks evaluated around the break size of interest, and the use of the largest debris amount from the subset of breaks available at the given size. STP provided debris generation and transport information for all welds that could fail and result in a LOCA as part of a risk-informed license amendment request to resolve GSI-191 (13). The staff considers STP representative of the most challenging plants with respect to the generation and transport of fibrous debris to the strainer. | |||
To determine the maximum amount of fiber that can arrive at the strainer and still meet the most conservative fiber limit, the staff considered a plant with the minimum number of FAs (157) and an 85 percent filter efficiency at the strainers. For these conditions, the maximum fiber allowed at the core inlet is 3,202.8 grams or 7.1 pounds (20.4 grams of fiber multiplied by 157 FAs.) | |||
Using an 85 percent filter efficiency for the strainer, this implies that 47 pounds of fiber can arrive at the strainer. (UPI plants have fewer FAs but are evaluated in a different way. See Section 5.2.1.3.) | |||
The smallest break at STP that results in 47 pounds of fibrous debris at the strainer is a 6-inch DEGB (5.19-inch inner diameter). The STP debris loading includes 28.5 pounds of latent fiber. | |||
Therefore, for this case, 16.5 pounds of break generated fiber are transported to the strainer. | |||
Some 6-inch breaks generate more fibrous debris, and some generate less. The range of fibrous debris loads at the strainer for all 6-inch breaks is 29 to 63 pounds. | |||
The next smaller break in the STP analysis is on a 4-inch pipe. The limiting 4-inch break transported about 38 pounds of fiber to the strainer. All other 4-inch breaks resulted in less fiber at the strainers. | |||
A fiber pass through of 15 percent (85 percent filter efficiency) is realistic and biased conservative based on the data that licensees have provided to the NRC. Section 5.2.1.2.1 discusses strainer penetration in further detail. The value is based on testing that used 100 percent fine fiber inserted upstream of the strainer and used realistic or conservative flow rates and other test methods. The data provided strainer filter efficiency for a range of debris loads. The NRC staff does not have a large database of penetration testing that is useful for calculating penetration values at relatively low fiber amounts. Most early penetration testing was performed to determine the maximum amount of fiber that can penetrate the strainer considering a plants maximum debris load. There is some uncertainty associated with the 15 percent penetration value because of the many different strainer configurations (size, number of strainers, flow velocity, and other factors) throughout the PWR fleet. The evaluation of the smallest break size that could generate enough debris to cause in-vessel issues is centered on a relatively small fiber amount. As the amount of fiber at the strainer increases, filtration efficiency increases. | |||
For this evaluation, the pipe break that could result in a problematic debris amount for the most limiting HLB conditions is assumed to be a DEGB of a 6-inch pipe on the hot side of the RCS. | |||
Such breaks are predicted to occur infrequently (see Section 5.5.2). Smaller, more likely breaks, will be less challenging to the fuel than analyses conducted by the PWROG and NRC indicate. Smaller breaks result in lower ECCS injection flow rates. | |||
Also, as discussed above CSS may not be initiated for some plants for small breaks. The lower injection flows delay the switchover to recirculation. The WCAP-17788 analyses assumed an early switchover based on the largest LOCAs. Early switchover results in a higher decay heat assumption that is more challenging to fuel temperatures. Delaying the switchover to recirculation would result in a decrease in decay heat and require less flow to maintain LTCC. | |||
This provides margin in the analyses. | |||
Subscale FA testing performed for WCAP-17788 found that headloss across debris at the core inlet will vary depending on the size and amount of particulate debris arriving at the core inlet. | |||
The fibrous debris limit was established using a conservative size distribution and amount for particulate debris. Small amounts of particulate result in an uncompact bed with low headloss. | |||
Excessive amounts of particulate result in an unstable bed and bed breakthrough at the fiber amounts tested. Flow also affected debris bed formation, with higher flow rates resulting in bed breakthrough. The fiber limits determined by testing were based on the most limiting conditions for the parameters studied. Alternate conditions will result in higher fiber loads required to reach the same headloss. | |||
The analysis discussed in this section is considered realistic or conservative for the following reasons: | |||
* It used debris loads representative of a very high fiber plant | |||
* It used a realistic to conservative strainer filtering value | |||
* It used the most limiting fiber limit for 2 FA types and 4 plant design categories | |||
* It used the smallest number of FAs for applicable plant designs | |||
* Other margins discussed in this section and throughout the document apply 5.2.1.2.1 Fiber Penetration There is a significant database that can be used to determine plant-specific total fiber penetration amounts or total amounts per specific strainer area. Early testing did not gather periodic data to determine the penetration behavior for fiber, but simply collected a total amount at the end of the test. More recent testing collected fiber downstream of the strainer at frequent intervals. Information from this testing has been used to determine how fiber penetration changes over time or with strainer debris load. | |||
Penetration data indicates that the percentage of debris that passes downstream of the strainer at any time is related to the debris load on the strainer. Penetration percentage values taken only at the end of the test are not meaningful except to the specific condition tested or to determine a maximum fiber penetration value for a specific plant. The NRC staff has a limited database of strainer penetration testing that can be used to determine time or load dependent | |||
strainer penetration amounts. Figure 1 below (from the Vogtle test program) illustrates that fiber penetration decreases as debris amount on the sump strainer increases (26). | |||
Figure 1 - Fiber Penetration as a Function of Debris Load The staff used information from six recent strainer penetration test programs that have more refined information based on frequent downstream debris collection. Each program performed several tests to gain understanding into the factors that affect penetration. The staff determined that fiber penetration amount decreases quickly with fiber buildup on the strainer as shown in the figure above. Relatively smaller breaks may be almost as challenging for downstream effects as larger breaks because the debris initially arriving at the strainer increases the filtration efficiency. | |||
The evaluation for the 6-inch HLB in this TER considered the results from these six penetration test programs because they were the only relevant information available to the staff. The staff used a value for penetration that was conservative with respect to the available data for the debris load considered for the case. Some plants may have plant-specific strainer configurations that would allow greater penetration at the specific debris load considered. | |||
However, two of the units with recent testing have very large strainer areas (one is the largest strainer installed in any U.S. PWR) making it likely that the data would be conservative with respect to the fleet. Other considerations like approach velocity and strainer perforation size | |||
can influence penetration values. The amount of fiber that reaches the core also depends on plant configuration, including the number of strainers, and how the operating ECCS and CSS pumps are aligned to the strainers. For example, any debris that is returned to the sump via the CSS or flows out of the break, returns to the sump and may be removed by filtering on the strainer. If CSS does not initiate or a plant has lower ECCS flow, more debris is delivered to the core before the strainer efficiency increases. | |||
The amounts of debris transported to the core for the CLB are less than for the HLB because most of the ECCS flow spills out the break in the CL. WCAP-17788 assumed that less than | |||
[xxxxxxxx] of fibrous debris would arrive in the heated core and at the core inlet combined for the HLB. The Vogtle study discussed above is for a very high fiber plant. This study indicates that 90 g/FA could reach the core under limiting conditions for Vogtle. Sections 5.2.1.1.3, 5.2.1.1.4, and 5.2.1.1.5 in the CLB section discuss debris suspended in the core. | |||
Considering available data and variations in plant strainer parameters, the penetration value used by the staff is conservative for most or all plants in the U.S. PWR fleet. | |||
5.2.1.2.2 Boric Acid Precipitation BAP has historically been evaluated with respect to CLBs (Section 5.2.1.1.2 above discusses the potential for BAP to occur following a CLB). However, if debris collects at the core inlet following a HLB, to the extent that flow into the core is significantly reduced, conditions similar to those in a CLB could occur. The staff considered the conditions following a HLB and determined that the potential for BAP was bounded by the CLB conditions. For a HLB, the coolant delivered to the core before a significant amount of debris arrives is much greater than that needed to replace coolant lost to boiling. This excess flow flushes boric acid and other debris from the core more than during a CLB. Flushing of the core will continue unless a significantly resistant debris bed forms at the core inlet. Even if a resistant debris bed forms and begins to inhibit flow through the core inlet, coolant is redirected to the core through AFPs. The flow delivered to the core exceeds that required to replace boil off and the fluid in the core is well mixed by convective currents and boiling. Therefore, boric acid is removed from the core such that the concentration required for precipitation will not be approached. | |||
5.2.1.2.3 Conclusions for the Hot-Leg Break The combination of the low frequency of a break that could generate and transport enough debris to challenge LTCC, and the conclusion that a debris bed that could block core inlet flow is unlikely to form implies that it is very unlikely that debris would adversely affect LTCC. | |||
Based on the discussion above, the staff concludes that the fiber limits from WCAP-17788 (also listed in Table 2 in Section 5.3.3.6) are acceptable metrics to assist in evaluating the acceptability of HLB debris amounts. Flow to the core will be maintained with these debris loads if chemical precipitates do not deposit or form on the debris bed. If a limited amount of fiber greater than the design-specific limit is transported to the core inlet, a uniform bed is unlikely to form and flow to the core will continue. The actual debris amount that might result in preventing adequate flow to the core is not defined. If the core inlet becomes completely blocked, AFPs will allow flow to | |||
reach the core as described in Section 5.3.3. There is adequate DID to assure enough flow to maintain LTCC following a HLB for the current operating fleet of PWRs that demonstrate acceptable fiber amounts at the core inlet. | |||
Considering the above, the NRC staff concludes that it is very unlikely that debris in the RV, generated by a HLB, would inhibit LTCC for the current operating PWR fleet. In addition, the BAP evaluation for the CLB is limiting with respect to the HLB, so BAP is not a concern for the HLB. | |||
5.2.1.3 Upper Plenum Injection Plants The UPI plants have an ECCS configuration unlike other PWR designs. The operating fleet has five UPI units (R.E. Ginna Nuclear Power Plant; Prairie Island Nuclear Generating Plant, Units 1 and 2; and Point Beach Nuclear Plant, Units 1 and 2). The ECCS initially injects directly to the UP, above the core via an independent nozzle. All fluid and entrained debris are available to enter the top of the core. The initial injection into the UP is similar to HL injection for a typical PWR. Therefore, a HLB for the UPI design is like the CLB for other plants and most of the flow goes out the break in the HL. The core inlet for ECCS injection and initial recirculation is at the top of the core instead of at the bottom of the core. | |||
The top of the core is not likely to become blocked with debris because of the turbulence caused by boiling and steam exiting the core. The steam exiting the core will impact structures above the core exit and add to the turbulence. For a HLB, most of the coolant spills out the break. For a CLB, the coolant flows through the core and out the break. Therefore, debris enters the core region with the flow. The effects of turbulence on the formation of debris beds were discussed above for typical PWR designs. Turbulence is much greater for the UPI design because coolant enters the reactor at the top of the core where the velocity of steam exiting the core is relatively high. Because the top of the core is unlikely to become blocked, cooling is maintained during injection into the UP. | |||
UPI plants use something similar to HLSO to prevent BAP following a HLB. Instead of switching injection from the CL to the HL, UPI plants switch from UPI to CL injection, or a combination of UPI and CL injection. The switch to CL injection occurs several hours after the initiating LOCA. | |||
If the break is in the hot side, the switch to CL injection will flush concentrated debris and boric acid from the core. If the break is in the cold side, coolant injected into the CL will flow into the core via the DC and LP and only the coolant required to make up for boil off enters the core. By the time a UPI plant switches to CL injection, debris in the sump will be depleted so it is unlikely that a debris bed will form at the core inlet. Because of the unique design of UPI plants, it is not critical that the switch to CL injection be initiated before chemical precipitates may form. This is because a filtering bed is unlikely to form at the top of the core where injection occurs. | |||
However, it is important that CL injection is initiated to flush concentrated debris and boron from the core. | |||
Once injection is switched to the CL, debris remaining in the sump may be pumped to the core inlet. The injection into the CL can provide pressure up to the shut off head of the ECCS pumps (as limited by the height of the SG tubes) if the break is in the HL. For this condition, WCAP-16793-NP-A and the associated staff SE accepted a fibrous debris load of up to 15 g/FA at the core inlet, even when chemical effects are included in the debris headloss. | |||
The switchover to CL recirculation takes place several hours after the LOCA; therefore, it is possible for chemical precipitates to form and filter on any debris bed. By this time, filtering on the strainer, settling, or deposition in the core will have removed most of the fibrous debris from the sump. A small amount of fiber may be suspended in the pool and transported to the strainer. Most of this will be captured on the debris bed on the strainer, but a small amount may penetrate and transport to the fuel inlet nozzles. As discussed in Section 5.2.1, a continuous bed is not expected to form at the core inlet, especially with the limited amount of fibrous debris arriving at the bottom of the core for this scenario. The debris amounts likely to arrive at the core inlet this late in the scenario are much less than 15 g/FA because the source term in the pool has been reduced to almost zero by the time CL injection is initiated. | |||
After switchover to CL injection, for a CLB, coolant is delivered to the core only by the head resulting from the manometric balance between the core and the DC. This is similar to the CLB with CL injection in other PWR designs, but it occurs for UPI designs only after the switchover to CL recirculation. For UPI plant CLBs, after switchover to CL injection, only the coolant required to make up for boil off enters the core. The remainder of the coolant with any entrained debris flows out the break. Because the switch to CL injection is relatively late in the LOCA scenario, chemical precipitation may occur. For this condition, WCAP-16793-NP-A did not conclude that a debris limit of 15 g/FA was justified. After chemical effects are included in the debris headloss, testing shows that the headlosses for the CLB flow rates can increase by 1 psi above non-chemical headlosses for tests that included fiber of [xxxxxxx]. This headloss could encroach on the available driving head for CL recirculation for the UPI plants. However, because the break is in the CL and most of the debris will have been deposited on the strainer before the initiation of CL injection, the amount of debris reaching the core inlet is much less than for the HLB. | |||
The largest source of fiber may be that from erosion of larger pieces that have settled in the pool. Most of this erosion occurs relatively early in the event and decreases with time. Erosion is a relatively small source of fiber and the eroded fines are likely filtered by the sump strainer/debris bed. Most of the fluid injected into the CL flows out of the break and back to the sump where it is filtered at the strainer before being recirculated back to the core. Because the amount of fiber in the pool has been depleted by the time CL recirculation is initiated and the majority of any debris that penetrates the strainer flows out the break, the quantity of fiber reaching the core inlet for a CLB is much lower than for the HLB considered in WCAP-16793-NP-A. The amount of fiber reaching the core inlet is unlikely to have a measurable effect on headloss or flow at the core inlet. | |||
5.2.1.3.1 Conclusions for Upper Plenum Injection Plants Because of the unique design of the ECCS for UPI plants, a debris bed will not inhibit flow into the core. The initial injection is into the UP where turbulence will prevent a bed from forming. | |||
By the time a UPI plant switches to CL injection, the amount of fiber that may be transported from the pool to the strainer is very small and the filtration efficiency of the strainer is high | |||
because of the debris bed deposited on it. Therefore, the amount of fiber reaching the core inlet (FA BNs) will not affect LTCC. | |||
5.2.1.4 Debris Generation and Transport This section discusses margin related to debris generation and transport. It provides details, especially for the aspects of the analyses that are more complex. The margins are summarized in Section 5.5.3. | |||
The most likely breaks are smaller breaks that are less challenging to mitigate from several perspectives. Reasons for this are discussed below with respect to specific parameters that are considered in the evaluation. In addition to those, smaller breaks result in lower required ECCS makeup flow. The injected sources are available for a longer period and decay heat is lower when the system is switched to recirculation. Lower decay heat allows for lower flows to make up for coolant lost to boil off and greater time for operators to respond to debris blockage. | |||
Debris generation values used by licensees in their analyses are for the most challenging breaks for strainer performance. The methodology used to determine debris generation amounts contains conservatism, but also uncertainty. The staff determined that the methodology is generally conservative, with few exceptions. However, the use of the debris generation values from the most challenging breaks clearly results in a substantially conservative bias because of the very low frequency associated with these breaks. The smaller, more likely breaks will result in much lower debris generation and reduced challenges to LTCC. | |||
The transport assumptions used in this evaluation and by licensees are based on approved guidance that is considered conservative. The assumptions about transport and debris characteristics are chosen to maximize the amount of debris arriving at the strainer. In addition, the debris amounts represent the bounding breaks for each debris type. | |||
The debris characteristics assumed for the transport evaluation are chosen to increase the probability that debris will be transported to the strainer. Coating particulate is assumed to be 10-micron particles that are suspended in the fluid. Testing has shown that the coatings will fail in a range of sizes and that some of the particulate will be large enough that it will settle in the sump pool or LP of the reactor. Fine fiber is assumed to be individual strands that remain suspended in the coolant. Clumps of fiber in larger size ranges are assumed to be the smallest size within the range, making them more likely to be predicted to transport. | |||
Fluid velocities and turbulence used in transport calculations are based on flows and coolant entering the pool that are developed from the largest breaks and maximized turbulence from fluid falling into the pool. Smaller breaks would result in lower flow rates, and reduced turbulence from water falling into the pool, and they might not result in flow from the CSS if CSS is not initiated. The lower flow rates and reduced turbulence would result in increased debris settling. In addition, lower flow rates result in delayed switchover to recirculation because the RWST is depleted more slowly. The delayed switchover allows more time for debris to settle in a relatively stagnant pool. A smaller break can also result in lower flow rates to the RCS and a higher CSS to ECCS flow ratio. The higher flow ratio results in the return of more fluid to the sump, which allows any entrained debris to pass through the strainer again with the potential for filtering. | |||
The transport evaluation assumes that particulate is not filtered but continues to recirculate until it reaches the core and remains there. Testing has demonstrated that some particulate is deposited on fiber beds relatively early. As the debris bed becomes denser, the amount of particulate filtered increases. Some particulate would also settle or deposit on surfaces. Early in the scenario, the conservatism associated with the assumption that particulate debris is not filtered may be small. | |||
Credit is not given for the capture of fine debris on interceptors, walls, or other structures in containment or within the RV. These structures would capture some fine debris. | |||
It is assumed that all fibrous debris is captured at the core inlet for both the CLB and HLB cases. Some of the fiber will pass through the core inlet and remain in the heated core region or exit the core with excess flow (for the HLB). The amount of fiber assumed trapped at the core inlet is conservative. | |||
Testing conducted for a new reactor design demonstrated that some fiber will settle in the LP and not be available for transport to the core inlet. Settling of fiber is not credited. | |||
Therefore, the amounts of fiber and particulate used for this evaluation are conservative. | |||
5.2.1.5 Chemical Effects Volume 5 of WCAP-17788 contains the in-vessel chemical effects evaluation (14). The following discussion summarizes the PWROG in-vessel chemical effects testing and the NRC staff evaluation. For greater detail, the draft NRC staff SE for Volume 5 is non-publicly available in ADAMS (27). The following discussion provides a high-level summary of the PWROG chemical effects testing and the NRC staff evaluation. Section 5.2.1.2 of this report gives an overview of the WCAP-17788 methodology, including how chemical effects were considered for a HLB. | |||
5.2.1.5.1 Testing PWROG survey data were used to create post-LOCA plant materials and sump chemistry inputs for the test plan. The plant materials input consisted of metals (including aluminum and galvanized steel), insulation materials, fire barrier material, and concrete. The fibrous and particulate insulation materials included calcium silicate, E-glass (fiberglass), silica powder, mineral wool, microporous insulation, and aluminum silicate. Fire barrier material included Interam'. The sump chemistry inputs included the projected post-LOCA temperature profile, sump pool volume, pool alkalinity (pH), and buffer type. The temperature profiles were based on either a post-LOCA temperature profile provided by a plant or were estimated using temperature data provided by the plants with their survey responses. | |||
Testing was performed with independently controlled 60-liter capacity stainless steel autoclaves (pressure vessels). Two resistive heater bands were mounted on the outside of each vessel to provide the heating for the elevated test temperatures. One of these heating bands was located near the bottom of the autoclave to promote convection and more uniform temperatures. The test set-up included means to introduce pressurized air and to vent air. A stainless-steel sampling line was located at approximately the mid-height of the autoclave vessel to avoid sampling either easily settled materials or floating materials in the fluid samples. The autoclave assemblies also included pressure gauges and pressure relief valves. Thermocouples provided temperature measurement of the test fluid. A computerized data system used feedback from the thermocouples to automatically control the temperature of the autoclaves to a | |||
pre-determined profile specific to the test group. Metallic test coupons were suspended from the top of the autoclave using polytetrafluoroethylene (PTFE) cord material to electrically isolate the coupon from the stainless-steel autoclave. Depending on the type of autoclave test, fibrous and particulate debris were either placed in a bag to constrain the loose debris (i.e., in-bag tests) or were freely dispersed within the autoclave test fluid (i.e., initial battery out of bag (IBOB) and out-of-bag tests). | |||
The NRC staff visited the Westinghouse laboratory in Churchill, Pennsylvania in May 2013 and November 2013 to witness testing and to discuss the test methodology and preliminary results. | |||
The visits were intended to improve staff understanding of the tests to better support a review of the WCAP-17788 submittal. The NRC staff trip report dated June 10, 2013 (28) is available in ADAMS. The NRC staff trip report dated December 24, 2013 (29) is non-publicly available in ADAMS. | |||
The initial autoclave water chemistry included representative amounts of boric acid for all tests. | |||
Autoclave test pH depended on the plant-specific conditions. For tests that used a sodium hydroxide (NaOH) buffer, pH adjustment occurred before insertion of the test materials into the autoclaves since NaOH is inducted into the containment spray immediately following a LOCA. | |||
Therefore, the pH for plant groups with NaOH was constant throughout each test except for any small changes caused by debris reactions. For plants-specific tests using sodium tetraborate (STB) or trisodium phosphate (TSP) buffers, the sodium carbonate and the buffer were added after the coupons had been exposed to the lower of either the maximum test temperature or 200 degrees Fahrenheit (°F) for 30 minutes. This sequencing is intended to expose the test materials to a more representative initial acidic plant pH before buffer dissolution. In the event of a LOCA, the sump pool pH for plants with STB buffering increases as STB is either dissolved from retention baskets or is introduced from melting ice (in the case of plants with ice condensers). For plants with TSP buffering, the pool pH following a LOCA increases as TSP dissolves from retention baskets. Following the addition of buffer and carbonate, the autoclaves are pressurized with air (approximately 30-50 psig) and controlled to a temperature profile specific to the test group. | |||
Fluid samples were taken from the autoclave tests at various times and were immediately placed in an oven at a predetermined temperature lower than that of the autoclave. The exact oven temperature was dependent on the autoclave fluid temperature at the time of sampling. | |||
The temperature difference between the autoclave temperature and the oven temperature was intended to simulate the temperature drop of sump fluid passing through a heat exchanger before entering the RV. After a minimum oven hold time of 1 hour, the samples were analyzed by performing vacuum filtration (to determine filtration times for each sample) and by turbidity measurement. | |||
Vacuum filtration was used to determine the specific timing of the chemical precipitate product. | |||
As a screening criterion, the PWROG established a filtration time (drain time) threshold value of 100 seconds for 50 milliliters of fluid to indicate the possible presence of chemical precipitates in the test fluid. The vacuum filter membranes used were a 1.0-micron pore size PTFE fiber filter designed for high-temperature filtration. | |||
The PWROG hot filtration test objective was to detect the presence of any chemical products that could cause high pressure drops across a fiber bed in the RV during the initial 24 hours following a LOCA. Since drain times can vary for a variety of reasons not related to chemical precipitates, the PWROG selected a 100 second drain time as the screening threshold for chemical effects based on earlier testing performed by Westinghouse and experience with the | |||
first six WCAP-17788 test groups. The 100 second threshold value resulted in identification of chemical precipitates and generated false positives, when drain times exceeded 100 seconds as a result of non-chemical debris. Tests with drain times exceeding 100 seconds and uncertainty about the cause of the longer drain times were then sorted out by further experimentation (e.g., bench tests). In Section 7.1 of Volume 5 of WCAP-17788, the PWROG later evaluated and confirmed the appropriateness of a 100 second drain time threshold value. | |||
The NRC staff also evaluated the filtration test 100 second drain time threshold for chemical effects screening. The staff reviewed the filtration test results and found the 100 second drain (filtration) time acceptable as a screening threshold as discussed in the draft SE. Based on the information provided by the PWROG, the staff found the temperature profiles and oven cooling before hot filtering that were used in the test program to be acceptable. | |||
The Volume 5 test results showed that aluminum precipitates were the cause for high drain times related to chemical effects. Finding aluminum-based precipitates to be the dominant contributor to post-LOCA chemical effects in PWR environments is consistent with previous testing results, including NUREG/CR-6913, Chemical Effects Head-Loss Research in Support of Generic Safety Issue 191, issued December 2006 (30); NUREG/CR-6914, Volume 2, Integrated Chemical Effects Test Project, issued September 2006 (31); and WCAP-16530-NP-A, Evaluation of Post-Accident Chemical Effects in Containment Sump Fluids to Support GSI-191, issued March 2008 (32). This is also consistent with NRC staff observations from many vertical loop and strainer chemical effect headloss tests performed at five different strainer vendor facilities, including tests with addition of pre-mixed WCAP-16530-NP-A precipitates and tests in which dissolved chemicals were added and precipitation occurred in the test loop (33). | |||
The NRC staff finds the aluminum precipitation boundary identified in Section 7 of Volume 5 of WCAP-17788 to be an acceptable prediction tool for chemical precipitation within 24 hours of a LOCA. The staff compared Argonne National Laboratory (ANL) and WCAP-17788, Volume 5, solubility predictions at various pH and temperatures. Although the ANL solubility equations (34) are more conservative at all temperatures of interest when the pH is less than 8.5, the ANL equations over-predict solubility for the high pH tests that can be representative of plants that add NaOH to the containment spray following a LOCA. The WCAP-17788 precipitation boundary predicted all cases of aluminum precipitation in Volume 5 testing, except for Test Group 41, which was not representative of a plant condition. In contrast to the one test group exception, however, there were 39 test samples where the aluminum precipitation boundary predicted chemical effects but there was no evidence of precipitation from subsequent filtration testing. In addition, there were 38 test samples where the precipitation boundary predicted precipitation and it was detected during filtration testing. These numbers indicate that the precipitation boundary function proposed in Volume 5 is conservative within the 24-hour post-LOCA time frame of interest. | |||
Although the NRC finds the aluminum precipitation boundary function identified in WCAP-17788, Volume 5, Section 7, to be conservative, the staff also recognizes there is inherent uncertainty associated with a predictive solubility tool. The PWROG addressed the uncertainty in the solubility boundary by proposing that plants retain margin to the boundary, as shown in Table RAI 5.13-1 (35). PWROG guidance in Section 7.6 indicates that plants that have changed material quantities from the test group can demonstrate that the change is acceptable by either (1) showing that the actual plant conditions reduce precipitation risk relative to the test, or (2) demonstrating that there is still sufficient margin to precipitation if the deviation from the test condition results in an increase in the risk of precipitation. The NRC staff finds the | |||
PWROG guidance related to the margin to precipitation values acceptable since it accounts for potential errors in temperature, pH, and inductively-coupled plasma mass spectrometer measurements in a conservative manner. | |||
Based on the considerations summarized above and discussed in more detail in the draft SE, the NRC staff finds the overall technical approach in Volume 5 of WCAP-17788-P to be acceptable for determining short term (within 24 hours following a LOCA) chemical effects in the RV since this approach used representative test conditions, accounted for uncertainties, and had an overall test approach that was effective in detecting chemical precipitation. | |||
5.2.1.5.2 Test ResultsPrecipitation Timing The overall WCAP-17788 methodology relies on the time to chemical precipitation being longer than the minimum time at which AFPs into the core can accommodate complete core inlet blockage. WCAP-17788 expresses success as tchem > tblock. Section 5.3.3 further describes tblock. Volume 1 of WCAP-17788 states that tblock is less than 6 hours for all plant designs (36). | |||
At that point, flow through AFPs into the core is sufficient for LTCC according to the WCAP-17788 analysis. The autoclave and filtration testing performed for Volume 5 showed that a 6-hour precipitation timing is conservative for almost all PWRs. | |||
Figure 2 shows the timing of chemical precipitation based on the WCAP-17788, Volume 5, PWR test group data. One test group with plant representative conditions had precipitation earlier than 6 hours. | |||
Figure 2: Timing of Chemical Precipitation for WCAP-17788 Test Groups | |||
For the remaining test groups, the NRC staff concluded that one test group may have experienced precipitation as early as 8 hours, although the PWROG concluded that precipitation did not occur until 24 hours for this group based on the mass of aluminum on the filter paper. | |||
Five other test groups showed evidence of precipitation in the 24-hour samples. One of these five test groups with precipitation at 24 hours only performed an out of bag test with plant representative amounts of debris. Earlier in bag tests with extra aluminum margin had early chemical effects. The plant-representative out of bag test contained about 2/3 of the aluminum coupon surface area compared to the earlier in bag tests with early chemical effects. | |||
No precipitation was detected up through the 24-hour sampling time for 37 of 44 test groups. | |||
5.2.1.5.3 Staff Analysis of Precipitate Timing The NRC staff used the results from the WCAP-17788, Volume 5, chemical effects testing to consider precipitation timing and other post-LOCA operator actions that could impact the integrity of a fibrous debris bed at the core inlet. The NRC staff compared chemical precipitation timing to the time when safety injection flow paths are modified (e.g., switching to HL injection or the equivalent) to preclude BAP. Switching safety injection to the HL, injecting simultaneously into the HL and CL, or an equivalent modified flow path (for some plant designs) before chemical precipitation is expected to reduce the probability of a high resistance debris bed forming at the core inlet. The NRC staff observed during the brine testing (6) that a reduction in flow allowed the buoyancy-driven circulation to disrupt the debris bed. This suggests that HLSO could disrupt a debris bed at the core inlet. | |||
The NRC staff evaluation focused the timing evaluation on the seven test groups (representing nine units) that confirmed chemical precipitation during the WCAP-17788, Volume 5, testing. | |||
One test group (two units) with confirmed chemical precipitates before 6 hours represents a two-unit, low fiber plant that has closed out in-vessel concerns under the conservative fiber limits established by the NRC staff in the SE for WCAP-16793. Therefore, assuming chemical precipitates arrive at the core inlet at this plant earlier than 6 hours, there is insufficient fiber to form a bed capable of blocking flow at the core inlet. This plant calculated a post-LOCA in-vessel fiber quantity of approximately 10 g/FA, significantly less than the 15 g/FA acceptance criterion the NRC staff approved in the WCAP-16793-NP-A SE. The staff approved the 15 g/FA limit since FA testing showed that the pressure drop from this quantity of fiber, with particulate and chemical precipitates, is acceptable. The single FA tests were designed to promote uniform flow to the bottom of the FA to ensure that a fiber bed was formed. This plant design also has UPI following a LOCA. Fiber bed formation at the top of a UPI plant core following a LOCA is not expected because of turbulence from boiling (see Section 5.2.1.3). | |||
Other than the units discussed above, the results from the next earliest test group indicated that precipitation may have occurred as early as 8 hours. The staff notes that the PWROG considered the confirmed precipitation time for this plant to be 24 hours. This plant switches to HL injection no later than 6.5 hours following a LOCA. | |||
One additional test group (one unit) with confirmed chemical effects at 24 hours using plant-representative conditions, had early chemical effects during initial in bag testing (with greater than representative materials, including aluminum surface area). The final out of bag test for this group was performed with representative amounts of calcium silicate, fiberglass, | |||
and aluminum. This plant switches to simultaneous injection in the HL between 5.5 and 6.5 hours after a LOCA. Without a plant representative in bag test, the staff is uncertain if precipitation would occur before HLSO since previous tests for this group showed in bag testing produced earlier chemical effects than out of bag tests. | |||
Four other test groups confirmed chemical precipitation at 24 hours. Three of the four groups have a maximum time before safety injection flow into the HL (or other flow injection into the RV) to prevent boron precipitation equal to 6.5 hours. The final unit that confirmed precipitation at 24 hours has a HL injection time of 14 hours following a LOCA. | |||
The remaining 37 test groups did not detect precipitation within the 24-hour test period. Each of these plants is expected to flush the core within 24 hours following the assumed LOCA, thereby increasing the probability that any bed on the core inlet would be disrupted before chemical precipitation. | |||
5.2.1.5.4 Chemical Deposits on Fuel The staff previously evaluated the effects of chemical deposits on fuel cladding as part of its SE of WCAP-16793-NP-A. That SE explains why the NRC staff approved the use of the Loss-of-Coolant Accident Deposition Model (LOCADM) to evaluate deposits on fuel following a LOCA. The NRC staff reviewed LOCADM analyses for the 21 units that have closed GL 2004-02 and none of these units have results that challenge the 800 °F or 50-mil thick deposit acceptance criteria. This includes units with high fiber loading. LOCADM is a tool developed by the PWROG that can be used to validate that chemical deposits will not prevent adequate LTCC. | |||
5.2.1.6 ConclusionAvailability of Primary Flowpath The staff concludes that following a LOCA, debris is not expected to completely block the normal ECCS flowpath through the core and prevent adequate LTCC. This is supported by testing and analysis that, taken as a whole, make this an implausible event. This conclusion considers the timing of events, understanding of the TH behavior of the system, and safety margins in the analyses. Should the primary flowpath through the core inlet become blocked, the staff considered DID offered by AFPs and other considerations. Section 5.3 describes these in more detail. | |||
5.3 Defense in Depth In the preceding sections, the NRC staff developed a basis to conclude that a core-wide inlet blockage arising from IVDEs following a postulated LOCA is an unlikely scenario. If, however, the core inlet became completely blocked with a high resistance debris bed, the NRC staff identified multiple mitigation features to provide LTCC, or to assure that the potential inhibition of LTCC does not lead to unacceptable public health and safety consequences. Since the NRC | |||
staff determined that a core-wide inlet blockage is unlikely, these are considered DID. The sections below evaluate each DID mechanism identified by the staff. | |||
5.3.1 Containment Blockage at the core inlet does not affect the ability of the CSS to depressurize and remove heat from the containment. Therefore, the failure of containment is not expected to result from an in-vessel blockage and the containment barrier is assumed to remain intact to provide DID for the IVDEs under consideration in this TER. The staff recognizes that the debris being evaluated for in-vessel blockage can also block the ECCS strainers that supply water to the CSS. However, this evaluation assumes acceptable strainer performance has been demonstrated. The staff evaluates strainer performance separately. | |||
The staff considered that a loss of LTCC could result in core damage that might affect strainers located close to the RV if the vessel is breached due to molten fuel reaching the lower head. | |||
The staff determined that few plants have with this configuration, and some of these have remote strainers in addition to the ones near the RV. The staff also determined that even if the core inlet is blocked, coolant delivered to the LP will reduce the potential for molten fuel to reach the lower head and for vessel breach. Therefore, it is unlikely that IVDEs would cause strainer failure and CSS will remain functional. | |||
5.3.2 Operator Actions All PWRs use HLSO or a similar action to prevent the buildup of boric acid in the core. In response to the issue of the effects of debris on LTCC, PWRs have initiated plant-specific actions to monitor for, and respond to, loss of flow to the core. HLSO is one of the actions that the plants use to inject coolant via a flowpath that is not blocked with debris (injection to the top of the core). Use of HL injection will also tend to disrupt any bed formed at the core inlet by reversing flow through the core. This will be more effective for CLBs, although reducing or stopping flow to the core inlet was shown to remove debris beds from the BN for low-resistance beds. | |||
In addition, industry developed actions that include monitoring and responding to blockage at the core inlet. These measures include monitoring core exit thermocouples, reactor levels, and containment radiation levels from the control room. If core blockage is detected, the plant operators have multiple options to mitigate the condition. Mitigative measures include refilling the RWST and injecting via different flowpaths, reducing RCS pressure, restarting reactor coolant pumps, and initiating early HL injection. Therefore, in the unlikely event that a LOCA occurs, a high-resistance debris bed forms at the core inlet, and AFPs are not successful in providing flow to the core, operator action options exist to maintain LTCC. | |||
5.3.3 Alternate Flowpaths If the core inlet becomes completely blocked, each plant has AFPs to provide coolant to the core for LTCC purposes. These AFPs include design features such as flow passages in the BB region, upper head spray nozzles, and pressure relief holes between the BB region and the core | |||
periphery. A series of TH analyses described in Volume 4 of WCAP-17788 (16) and associated RAI responses (37) (38) (39) evaluates the efficacy of the AFPs to provide LTCC. | |||
The material described in this section relates to debris accumulating at the core inlet (specifically, the BN or spacer grid) after a HLB. Another consideration is that debris present in the coolant not captured at the core inlet, passes through the AFPs, and continues to the heated core region. This can happen in two ways: (1) resistance at the inlet builds to the point where AFPs are activated, directing flow through the AFP to the core or (2) operator action occurs to initiate HLSO, in which a fraction of the ECCS flow is transferred from the CL to the HL to mitigate possible BAP. The combined amount of fibrous debris that can accumulate at the core inlet and within the core without affecting LTCC is assumed to be [xxxxxxxx], which is discussed in Volume 1, Section 6.4, of WCAP-17788. The staff evaluated in-core fibrous and particulate debris as described in Section 5.2.1.1.3 of this TER. The remainder of this section considers the effect of debris blocking the core inlet. | |||
5.3.3.1 Background Once a plant depletes its injectable inventory, its ECCS is aligned for recirculation via SSO. At that point, fibrous and particulate debris may enter the RCS. Such debris may collect at the core inlet as the recirculating coolant travels from the LP into the core. Over time, as debris continues to collect, the hydraulic resistance at the core inlet will increase. The presence of chemical effects, such as precipitates, may also contribute to increased inlet flow resistance. If the flow resistance continues to increase, a greater fraction of liquid will enter the core through the AFPs as opposed to the core inlet. If the core inlet becomes completely blocked, the AFPs become the sole source of liquid entering the core to replace that which is boiled off or spilled out of the break. | |||
Many variables can influence the scenario described above, including the following: | |||
* the depletion time of the source of injectable emergency coolant, | |||
* the amount of debris entering the RCS, | |||
* the fraction of that debris that can be transported to the core inlet, | |||
* the amount by which the hydraulic resistance at the core inlet increases over time, | |||
* the ability of, and time at which, transported debris may cause a complete core inlet blockage The scenario is further complicated by plant-specific considerations, including design features such as flow connections between the BB and core regions or ECCS flow capacities, and the particular operating characteristics (e.g., core power distribution) at the time of the initiating event. | |||
5.3.3.2 Simplifying Assumptions To distill this complex scenario into a tractable problem, the PWROG made several simplifying assumptions in evaluating the efficacy of the AFPs. Some of these assumptions are summarized below: | |||
* break location - The AFP analyses simulated a HLB because it results in the largest fraction of recirculating ECCS flow traveling to the core inlet, where debris could then collect to form a bed. | |||
* SSO time - The AFP analyses assumed that SSO would occur 20 minutes following the initiating event. This is intended to represent a bounding minimum that would cause most debris to be transported into the RCS early, while the decay heat remains high. | |||
* core operating characteristics - The AFP analyses assumed the limiting FA was operating with a top-peaked power shape, similar to that assumed in small-break LOCA analyses, to maximize a second heatup if the core uncovers during a core inlet blockage. | |||
* reactor design differences - Four separate sets of analyses make up the AFP analyses, each set intended to represent a different class of operating PWR reactors. These included: (1) an upflow-configured, 4-loop Westinghouse plant, (2) a downflow configured, 3-loop Westinghouse plant, (3) a Combustion Engineering (CE) plant, and (4) a Babcock & Wilcox (B&W) plant. | |||
* analytic methods - The AFP analyses were performed using EMs that were derived from NRC-approved EMs used to analyze short-term ECCS performance. | |||
* AFP treatment - Each class of plant has different AFPs that were modeled. The AFP resistances were hand calculated using physical plant geometry and, via scaling analyses, were adjusted to ensure that core power was considered as well (see below). | |||
The treatment of rated thermal power level also warrants additional discussion. In WCAP-17788, the thermal power levels selected for each category were not necessarily bounding for each plant in that category. Instead, the PWROG intended to account for variations in core power in the way that the AFP resistances were modeled. The PWROGs response to RAI 4.2 (38) provided a scaling analysis to demonstrate that the core power was considered by scaling the AFP resistance for the modeled plant. Although the NRC staff did not review this RAI response in sufficient detail to make a regulatory finding, it provides information to indicate that the analyses included a means to account for variations in core power level without modeling an explicitly bounding, high-power level. | |||
Although these simplifying assumptions enabled a reasonably consistent set of analyses among the four design classes, it should be noted that the CE analyses were performed using slightly different approaches, in some cases, because the minimum ECCS flow per FA is lower than for the other plants. As appropriate, the sections below note these differences. Furthermore, because the CE plant analyses used a different, and generally less limiting, set of assumptions, the NRC staff performed supplemental analyses for the CE plants using the TRACE computer code, as discussed in Section 5.4. | |||
5.3.3.3 Description of Analyses The goal of the WCAP-17788 analyses in Volume 4 (16) was to determine if AFPs were adequate to maintain LTCC after switchover to recirculation and formation of a debris bed blocking the core inlet. While 10 CFR 50.46(b)(5) requires that the core be maintained at an acceptably low value to remove decay heat from the core, the PWROG opted for an | |||
acceptance criterion of 800 °F as a more quantitative, acceptably low value to show that LTCC is not compromised. | |||
In this TER, however, the NRC staff evaluated the AFP analyses differently. The NRC staff considered the AFP analyses, insofar as the analyses provided evidence that the AFPs would ensure LTCC if a complete core inlet blockage occurred. It should be noted that, while the AFP analyses were performed using methods based on NRC-approved models, the analytic results were not considered to provide an exact prediction of the PCT, or to provide a bounding prediction of behavior for the entire PWR fleet. As discussed below, some parameters, such as the assumed plant power level, do not bound all plants in the category analyzed. However, the NRC staff considers the analyses to establish the concept that the AFPs can provide an additional means to cool the core thus preventing core damage beyond that which occurs during the blowdown and recovery phases of the event. The NRC staff determined that this conclusion is appropriate, given the low likelihood of complete core blockage following the already unlikely initiating event. | |||
The AFP analyses described in WCAP-17788 include three variablesi that characterize the effects of debris blockages at the core: | |||
(1) Kmax is the maximum resistance due to fibrous and particulate debris bed formation that can be tolerated at the core inlet. This value is compared to results of subscale headloss testing to establish an upper bound of the amount of fibrous debris that can be tolerated at the core inlet. A Kmax value corresponding to the headloss testing results for each FA was applied beginning at the SSO time, although the CE analyses modeled Kmax as a piecewise linear ramp beginning 800 seconds after the SSO time. This difference was justified in the response to RAI 4.26 (39). In actuality, Kmax is a function of ECCS flow rate. The limiting cases are derived from the lowest ECCS flow rates. | |||
(2) tblock is the earliest time that complete core blockage can occur while not compromising LTCC. While Kmax was applied at SSO time (except for the CE analyses identified above), the core inlet flow resistance was increased to a prohibitively high value to prevent flow through the core inlet and establish the AFPs as the primary source of core cooling. For the CE plants, the analysis for tblock was separate from the Kmax analysis. | |||
(3) tchem is the time at which chemical precipitates form, as determined by autoclave testing. | |||
The methodology assumes that after tchem, the resistance through the debris bed is greatly increased because of chemical precipitate formation. Due to this increased resistance, the WCAP-17788 methodology requires the timing of complete core blockage, tblock, to be less than the timing of chemical precipitates, tchem. If tchem is less than tblock, then the method in WCAP-17788 is not valid and further analysis is needed. | |||
To simplify the calculations and provide margin, the Kmax values were determined at limiting conditions as discussed above. The Kmax values are converted to debris limits as described in Volume 6 of WCAP-17788 (40) and discussed further in Section 5.2.1.2. | |||
i In a separate series of analyses, the PWROG also evaluated the parameters ksplit and msplit, which related to the core inlet flow resistances and flow rates required to divert flow into the AFPs, but these are not evaluated in detail in this TER. They merely assure that, at any given ECCS flow rate, some portion of flow will divert to the AFPs as the core becomes blocked. | |||
The Kmax and tblock values are determined iteratively. That is, for Kmax, the value was increased until the PCT limit was approached when the resistance was applied at SSO (except for the CE analysis which is discussed above). For tblock, the time of application of a near infinite resistance was decreased until the PCT limit was approached. | |||
5.3.3.4 Analytic Methods The PWROG used three separate computer codes to perform the AFP analyses: | |||
(1) The Westinghouse upflow and downflow analyses were performed using WCOBRA/TRAC. The NRC staff approved this code for use as a part of the Westinghouse Code Qualification Document (41) and Automated Statistical Treatment of Uncertainty Method (ASTRUM (42)) ECCS EMs. | |||
(2) The CE analyses were performed using S-RELAP5. The NRC staff approved S-RELAP5 for use, among other analytic methods, as a part of the Realistic Large-Break LOCA (RLBLOCA) EM documented in EMF-2103(P)(A), Revision 1, Realistic Large Break LOCA Methodology for Pressurized Water Reactors, issued January 2015 (43), | |||
and Revision 3, issued June 2016 (44). | |||
(3) The B&W analyses were performed using RELAP5/Mod2-B&W (R5M2BW) (45). The NRC staff approved R5M2BW for use, among other analytic methods, as a part of the Babcock & Wilcox Nuclear Technologies (BWNT) LOCA EM (46). | |||
WCOBRA/TRAC and S-RELAP5 are approved as elements of realistic ECCS EMs, in which the computer code is applied with a statistical methodology that accounts for uncertainties, while R5M2BW is approved as an element of an EM conforming to 10 CFR Part 50, Appendix K, ECCS Evaluation Models, in which the EM relies on a series of bounding, conservative assumptions to account for uncertainties in the results. Regardless of the computer code however, all the AFP evaluations were performed in a deterministic sense, with reliance on conservative assumptions to offset the effects that uncertainties could have on the predicted results. No attempt was made to quantify uncertainties using the statistical features of the Code Qualification Document, ASTRUM, or RLBLOCA. | |||
5.3.3.5 Conservative Assumptions As described above, these analyses relied on a set of generally conservative modeling assumptions to offset some uncertainties that were not explicitly addressed. While this evaluation is not intended to review these conservatisms exhaustively, examples include the following: | |||
* Decay heat was modeled similar to the requirements in 10 CFR Part 50, Appendix K. | |||
The Westinghouse upflow and downflow analyses, as well as the B&W analyses, used the 1971 American Nuclear Society (ANS) Proposed Standard, Decay Energy Release Rates Following Shutdown of Uranium Fueled Thermal Reactors (47) with a 1.2 multiplier. The CE analyses for tblock used the 1979 standard American National Standards Institute (ANSI)/ANS-5.1-1979, Decay Heat Power in Light Water Reactors (48), for decay heat with a 1.1 multiplier, while the CE Kmax analyses used the 1979 ANS standard for decay heat with a 1.2 multiplier. | |||
* Owing to variability in conceivable ECCS recirculation flow rates, the PWROG studied a range of flow rates on a per-FA basis. The limiting results for the Kmax/tblock analyses were obtained from minimum recirculation flows, which are the basis for this evaluation. | |||
* For the limiting analyses summarized in Section 5.3.3.6, Kmax was applied at the time of SSO, with the exception of the CE Kmax analysis, which applied a piecewise linear ramp beginning at 2,000 seconds (i.e., 800 seconds after SSO time, as described in the response to RAI 4.26 (39)). The response to RAI 4.26 also showed a reasonable amount of conservatism in the CE Kmax modeling approach. | |||
* For the Westinghouse upflow and downflow analyses, as well as the CE analyses, a complete blockage was modeled instantaneously at tblock. This is conservative because a debris bed would take time to build at the core inlet, especially to the point where inlet blockage occurs. Modeling these blockages with a more gradual ramp would result in earlier establishment of the AFPs and could even prevent a secondary heatup. | |||
However, instantaneous blockages were modeled to obviate the need to justify a particular time-dependent ramp of the core inlet flow resistance. | |||
* Because the B&W plants have low resistance AFPs, resulting in greater coolability, they were evaluated assuming a complete blockage at the core inlet beginning at the time of SSO. | |||
5.3.3.6 Results In aggregate, the limiting analyses for each plant indicated that, if the core inlet blocked with a finite amount of fibrous and particulate debris at SSO, followed by a complete blockage at some later time, a secondary cladding heatup would occur. In the Westinghouse analyses for Kmax, the heatup occurred coincident with the application of the Kmax blockage at the time of SSO. In the Westinghouse tblock analyses, a heatup of similar magnitude occurred with a complete core blockage at time tblock. In both cases, Kmax and tblock were modeled to ensure that PCT did not exceed 800 °F. In the CE analyses, there was not a significant heatup in the Kmax analysis, but there was a heatup to about 545 °F after tblock. Meanwhile, the B&W analyses indicated that, because of the ample flow communication between the BB and core region for that category of plants, LTCC would be assured with no secondary heatup, even if the core inlet is completely blocked at the time of SSO. Altogether, the analytic results indicate that the AFPs are an available source of DID, in the unlikely event that the entire core becomes blocked with debris. | |||
The limiting results from the analyses in WCAP-17788 are summarized in Table 2. | |||
In WCAP-17788, Volume 4, and in response to NRC staff RAIs, the PWROG also performed sensitivity studies, which demonstrated the conservative nature of the analyses. For example, for the Westinghouse upflow analyses, the PWROG also simulated a linear ramp in core inlet flow resistance in two cases. These cases illustrated more benign system behavior compared to the Kmax analyses, and the PWROG stated that these linear ramp cases were more realistic regarding the timing at which debris is expected to arrive at the core inlet. As discussed elsewhere in this TER, the NRC staff agrees with this assessment because the debris would | |||
require transport to the core from the sump, such that an instantaneous blockage at the time of SSO is unlikely. | |||
Table 2: Debris Limits and the Associated tblock Plant Design Fuel Type Debris Limit (g/FA) tblock (min) PCT (°F) | |||
Westinghouse [xxxx] | |||
Westinghouse Upflow 143 780 Framatome [xxxx] | |||
Westinghouse Westinghouse [xxxx] | |||
260 790 Downflow Framatome [xxxx] | |||
Westinghouse [xxx] | |||
CE 333 545 Framatome [xxxx] | |||
Westinghouse [xxx] | |||
B&W 20 + | |||
Framatome [xxxxx] | |||
*These represent maximum values based on the debris load tested. | |||
+The B&W analysis did not indicate a secondary cladding heatup. | |||
As discussed in Section 5.3.3.3, the PWROG applied a PCT acceptance criterion of 800 °F or less to demonstrate adequate LTCC. The NRC staff did not find that the results demonstrated, with high probability, that LTCC would be ensured. However, the NRC staff accepts the analyses as a demonstration that AFPs provide a measure of DID. The specific results are quantified above to illustrate the variability among analyses, as well as to correlate the results to in-plant debris limits, specific tblock values, and to provide a comparison to the TRACE supplemental analyses (21) discussed in Section 5.4. | |||
These analyses assume that in-vessel debris is limited to a value that would correspond to a core inlet flow resistance that was modeled as Kmax for each plant type. The PWROG performed subscale headloss testing, as described in Section 6 of Volume 1 of WCAP-17788, to correlate specific fiber limits to the Kmax values modeled in the AFP analyses. | |||
Tables 6-3 and 6-5 of WCAP-17788, Volume 1, provide these fiber limits, which are summarized above in Table 2. | |||
5.4 TRACE Sensitivity Studies RES conducted TRACE analyses (21) to explore the effects of debris on LTCC. These analyses apply to both HLB and CLB conditions. Details of the various studies performed are shown in Table 3. None of these are confirmatory calculations in the traditional sense in that they were not intended to validate or verify specific industry results. The first two columns include studies performed in the 2015/2016 timeframe. They were completed before WCAP-17788 was submitted to the NRC for review. The second two columns are a result of the current effort to aid in determining the safety significance of IVDEs. | |||
As shown in Table 3, the TRACE analyses were performed for two specific plant models that are representative of or conservative compared to the operating PWR fleet. Modeling was not performed using plant-specific conditions but used realistic inputs and assumptions to provide insights regarding LTCC behavior considering the effects of debris. The analyses also provided | |||
insights into how changes in important assumptions affect behavior. The analyses show that flow within the core and LP regions will promote non-uniform debris distribution at the core inlet. | |||
They also show that liquid entrainment will limit the concentration of boric acid within the core region such that BAP is not a concern if licensees implement CL injection as required in their licensing bases. The TRACE work provides a separate and independent source of knowledge regarding the effect of IVDEs on LTCC. | |||
Table 3: TRACE Sensitivity Studies CLB BAP HLB BAP/AFP HLB AFP CLB AFP Purpose Evaluate effects of Evaluate effects Evaluate PWROG Evaluate debris buildup at of debris buildup conservatisms PWROG core inlet on BAP at core inlet on conservatisms BAP and LTCC Model DEG, 4-loop W upflow, mid-peaked axial, DEG, CE, top-peaked axial, ANS 79 ANS 79 Cases/ 99.9% blocked - 99.5-100% -Decay heat -Base case LCP blocked core only Sensitivities inlet -Bypass 100% blocked nozzle -Blockage timing | |||
-HLSO time, decay -Rate of blockage heat, axial skew -HPI flow 5.4.1 CLB For the CLB, the TRACE analyses address both the effects of blockage at the core inlet and the potential for BAP. | |||
The TRACE runs conducted by RES (21) for the CLB indicate that it is unlikely for BAP to occur, or for LTCC to be inhibited because of a loss of flow, even if a debris bed forms at the core inlet. | |||
TRACE runs were performed for a Westinghouse 4-loop upflow design PWR and a CE design PWR. The majority of the CLB analyses were conducted for the Westinghouse design and the discussions in this section concentrate on those analyses. The staff concluded that for the CLB, the mechanical design of the reactor would not have a large influence on the analysis. | |||
In the Westinghouse simulations, TRACE was modified to account for higher densities and viscosities that occur because of the concentration of boric acid within the core region following a CLB. The TRACE runs confirmed the temperature driven circulation patterns that were demonstrated in industry modeling. The simulations assumed that recirculation started at 1,200 seconds to ensure the use of high decay heat values. Several cases were run. A | |||
reference case that assumed no blockage was performed for comparison to the cases that simulated blockage. The reference case and most other cases assumed a mid-core peak and nominal decay heat based on the 1979 ANS draft Standard ANS-5.1/N18.5, Decay Energy Release Rates Following Shutdown of Uranium Fueled Thermal Reactors (48). The reference case found that up until about 10,000 seconds, the boric acid concentration in the core region was limited because of entrainment of liquid in the steam flowing to the SGs. As decay heat (and entrainment) decreased, the concentration of boric acid in the core increased. The concentration of boric acid in the core and LP was uniform. Mixing with the LP was caused by the upflow in the central core and downflow in the periphery. The difference in boric acid concentration between the two volumes was not significant enough to cause buoyancy-driven circulation. However, the boric acid concentration in the LP remained slightly lower than that in the core. It was assumed that HL injection was initiated at 6 hours, which eliminated the increase of boric acid concentration in the core. The baseline simulation was continued to 7.5 hours, at which time the core was covered, and boric acid concentrations returned to low levels. | |||
5.4.1.1 Westinghouse Upflow Plant The following list includes the reference case assumptions for the Westinghouse Upflow Plant TRACE CLB analysis. | |||
* DEGB of the CL | |||
* No debris blockage | |||
* SSO at 20 minutes | |||
* Blockage applied from 1,200 seconds to attain full resistance at 5,000 seconds, linear ramp | |||
* HLSO at 6 hours | |||
* Nominal decay heat | |||
* End of simulation at 7.5 hours | |||
* Flat axial power peaked at mid-core | |||
* 212 °F recirculation temperature Additional TRACE cases simulated progressively increasing levels of debris blockage. The blockage was simulated at the lower core plate, below the area where blockage would occur in the plant. This is a conservative scenario because simulating the blockage at this location prevents flow through the AFPs credited in the PWROG analyses for the HLB from WCAP-17788. Other sensitivity studies, discussed later in this section, applied blockage at the core inlet which allows flow to the AFPs. The blockage was applied by closing valves to the assumed blocked area. RES checked the resistance (k/A2) value for the valves against the expected debris resistance value to ensure the modeling was valid. The blockage was applied over time starting at 1,200 seconds and was ramped to the full blockage for the case being run | |||
over the next 3,800 seconds, or just longer than an hour, which is considered reasonable. The full blockage value for the case was assumed to occur at 5,000 seconds. | |||
Each of the sensitivity studies for the Westinghouse upflow plant are summarized in the following sections. | |||
5.4.1.1.1 Blockage at the Lower Core Plate (99 percent) | |||
In the case where 99 percent of the core area was blocked, the core remained covered and no temperature excursions occurred. Boric acid concentration was elevated but remained well below the concentration where BAP could occur. Localized BAP was observed only in the UH in areas where trapped fluid boiled away leaving the boric acid solids behind. There was no BAP predicted at the core inlet, BB regions, or areas that could impede coolant flow to the core. | |||
The simulation of HLSO effectively ended the transient, even when the resistance at the lower core plate was assumed to remain in place. | |||
5.4.1.1.2 Blockage at the Lower Core Plate (99.5 and 99.9 percent) | |||
The case for 99.5 percent blockage had slightly elevated boron concentration compared to the 99 percent case, but LTCC was maintained. Cases where core inlet blockage exceeded 99.9 percent resulted in cladding heat up as blockage at the core inlet prevented adequate coolant from entering the core. In this simulation, boric acid concentration in the core was found to be below the precipitation limit, even when the PCT excursion occurred. This finding implies that fuel failure caused by loss of coolant flow will occur before failures caused by BAP. The boric acid concentration did not approach the solubility limit in any of the blockage sensitivity studies. | |||
5.4.1.1.3 Blockage at the Fuel Inlet Nozzles (99 and 100 Percent) | |||
The TRACE simulations investigated the effect of debris blockage location by simulating blockage at the fuel inlet nozzles instead of the lower core plate. Blockage at this location prevents (or reduces) coolant flow entering the core directly but allows flow to the BB region. A 99 percent blockage was applied at the fuel inlet nozzles. Results of this simulation showed that LTCC continued and boric acid concentration remained well below precipitation limits. Flow from the BB region into the core near the top nozzles mixed with the coolant in the core and diluted boric acid in the core. | |||
For the 100 percent blockage case, the only flowpath to the core was through the BB region. | |||
Core coolability was maintained and boric acid concentration was maintained well below the BAP limit. Flow through the BB region and into the core exceeded boil off from decay heat. | |||
Because of the low resistance assumed at the core inlet in WCAP-17788 CLB cases, flow through the BB region was not credited to ensure LTCC. The identification of the BB flowpath by the TRACE simulations indicates that the diverse flowpaths credited for the HLB can also provide the required flow for the CLB. The flow is through the same pathways credited for the HLB analysis, so all reactor designs evaluated have flowpaths available to allow LTCC for a CLB if the core inlet becomes completely blocked. | |||
5.4.1.1.4 Blockage at the Fuel Inlet Nozzle with Delayed HLSO (100 percent) | |||
A case was run to evaluate the effects of a much later switchover to HL injection. Instead of assuming 6 hours for the initiation of HL injection, the case was run for 36 hours. The case assumed 100 percent blockage at the FA inlet. The boric acid concentration increased over time but did not exceed the solubility limit. The concentration plateaued at around 30 hours. | |||
The boric acid in the core remained well mixed. Some precipitation was predicted in the UP after 30 hours. This likely resulted in maintaining the core concentration below the solubility limit. | |||
5.4.1.1.5 Core Power Sensitivity A sensitivity case using Appendix K decay heat assumptions was run by increasing the reference case decay heat by 30 percent. A 99 percent blockage was assumed at the lower core plate. The higher power case resulted in a more highly voided core and UP, but no heatup occurred. Compared to the similar case at the reference decay heat value, the boric acid concentrations were significantly lower and the rate of increase in concentration remained low for a longer time. This is caused by entrainment of the boric acid in the liquid exiting the core. | |||
5.4.1.1.6 Axial Power Shape Sensitivities Top and bottom skewed axial power shape cases were run. Both cases assumed 99 percent blockage at the lower core plate. There were differences in the very early PCT results during blowdown and reflood, but the LTCC results were very similar. | |||
An interesting result from the decay heat sensitivity is that higher decay heat results in lower boric acid concentrations. Greater liquid entrainment resulted in more boric acid removal from the core. This is counterintuitive unless entrainment, which is not credited for design basis calculations, is considered. The TRACE simulations also provide more realistic results than the SKBOR methodology discussed in Section 5.2.1.1.2 in that they account for the physical effects of liquid carryover in the steam and for viscosity and density changes resulting from concentrated boric acid. | |||
5.4.1.1.7 Large Split Break at Top of CL A sensitivity case was run to examine the effects of a split break at the top of the CL. The results were like those of the DEGB cases. | |||
For all Westinghouse CLB cases, there is liquid carryover at the beginning of the event so that the buildup of boron in the core region is limited. Later in the event, boron may begin to concentrate in the core. However, the increased density of boric acid in the core region will promote buoyancy driven convection and mixing with the LP. This convection, convection caused by heating of the fluid, and the agitation caused by boiling results in mixing within the core and between the core and LP. Early in the LOCA recovery, heat-driven convection and boiling will drive the mixing. Later, convection resulting from density increases from boron concentration may become more dominant. | |||
5.4.1.2 Combustion Engineering Plant The CE cases were run mainly to gain understanding of the behavior of the CE design following an HLB. The CE cases did not model concentration of boric acid in the core. The single CLB case run for the CE design used a decay heat multiplier of 1.1 times the ANS 1979 decay heat curve and a top peaked axial power shape. The case simulated full core blockage at 20,000 seconds. No cladding heatup was predicted. This case shows that the AFPs can provide cooling for the CE design even if a full core blockage occurs. The conclusion is valid for the conditions simulated but is dependent on several variables including the time blockage occurs. | |||
5.4.1.3 TRACE Study Applicability to the PWR Fleet The Westinghouse and CE HLB TRACE studies are discussed in Section 5.4.2. | |||
Applicability of the BAP TRACE study to the PWR fleet considered the following: | |||
* The TRACE runs modeled only two reactor designs (Westinghouse 4-loop upflow and CE), and most of the CLB cases were run for the Westinghouse upflow design. | |||
* The boron concentration aspects of the model were validated to Primrkreislauf-Versuchsanlage (commonly known as PKL) data with good correlation, but additional validation of modeling for boron concentration and density was not performed because of a lack of available data. See Section 4 of the TRACE report (49). | |||
Other parts of the model were more rigorously validated. | |||
* The simulations did not use design-basis assumptions but performed sensitivity studies to understand how assumptions can affect the results. On the other hand, the studies considered core inlet blockage, which has not been considered in design-basis calculations. | |||
* The results are not considered design basis type calculations, but they indicate that BAP is much less likely to occur than previously postulated and provide independent information regarding the PWROG calculations for core cooling and BAP. | |||
The TRACE results also identified that circulatory currents are present within the RV. These currents will prevent uniform debris bed formation at the core inlet. The PWROG simulations showed that these currents were present early in the LOCA recovery period but, unlike the TRACE studies, did not attempt to demonstrate that they would be present later in the recovery because the concentration of boric acid in the core was not modeled. | |||
RES concluded that the qualitative results of the TRACE analysis should be applicable to all plants that have bypass holes that will not become blocked by debris. UPI plants are significantly different so are not comparable. Variations in key plant parameters will affect the quantitative results. Westinghouse and B&W designs have core bypass flow paths that are qualitatively similar to those of the CE plants. B&W and some Westinghouse designs have AFPs that connect the bypass region to the core at lower elevations so that coolant can enter the core without filling up the entire bypass flow volume. The BAP results from the Westinghouse plant CLB TRACE analysis are applicable to the CE and B&W plants since there are significant mixing volumes above the core and in the HLs for all designs. The CLB debris | |||
limits are the same for all fuel and reactor designs and the response of the system is similar enough for the results to be considered applicable. | |||
The NRC staff considers the Westinghouse plant selected for boric acid concentration studies conservative relative to the PWR fleet due to its upflow configuration in the core bypass region. | |||
The upflow design reduces the mixing volume available for dilution of boric acid. The design also delays flow from the AFP to the core if a 100 percent blockage at the core inlet nozzles occurs because the BB must fill to the top before flow to the core is established. | |||
The Westinghouse CLB TRACE analysis found that core uncovery occurs before BAP takes place. Because core uncovery is the central issue during long-term cooling, additional calculations were made for the representative CE plant. These calculations showed that even with a conservative blockage at the fuel inlet nozzles, uncovery is not expected to occur for this design, similar to the Westinghouse plant findings. The calculations and sensitivity studies performed in the TRACE report indicate that even with a highly conservative blockage at the core inlet nozzles, neither core uncovery nor BAP is expected for the CLB case. | |||
In the unlikely event that the fuel inlets become completely blocked, TRACE simulations show that the BB will fill with coolant and coolant will flow into the core via AFPs. TRACE demonstrates that the core remains covered. The AFP through the BB region that was credited for the HLB (see Section 5.3.3) is also viable for the CLB, although industry analyses for the CLB do not credit it. | |||
5.4.2 HLB RES performed a series of HLB sensitivity studies using the TRACE code for both the Westinghouse upflow and CE plant categories as described below. | |||
5.4.2.1 Westinghouse Upflow Plant RES performed Westinghouse upflow sensitivity studies to determine if a break in the HL could result in high boric acid concentrations in the core and RV. The plant selected for simulations was a Westinghouse four loop PWR with a core power of 3,411 megawatts thermal. The plant model (representative of the Seabrook Station) is an upflow plant, meaning that core bypass flow enters the BB region near the core inlet and exits near the fuel top nozzle elevation. No communication was modeled between the BB region and the active core. This configuration is assumed to be conservative for investigations involving core inlet blockage with debris. Coolant cannot reach the core unless it flows completely though the BB region and back into the core near the core exit, unless it flows through the high-resistance path of the debris. | |||
For the HLB calculations, the blockage was simulated to occur at the bottom fuel nozzles using a resistance ramped from 1,200 to 5,000 seconds. The total blockage ranged from 99 percent to 100 percent. At 100 percent blockage, no flow could go directly into the bottom of the core, and any coolant reaching the core was forced to flow through the BB region and enter at the top of the core. While the BB region filled, sufficient inventory remained in the core region so that the rods remained covered and cladding temperatures remained at saturation temperature. | |||
After the BB region refilled, sufficient coolant entered at the top of the core to maintain core coverage and prevent cladding heat up. | |||
At 100 percent blockage, the core boric acid concentration was only slightly greater than that obtained in the calculations with 99.5 percent blockage or less. Flow through the BB region was sufficient to remove decay heat and keep the two-phase level at the HL elevation. | |||
Both HL calculations (99.5 and 100 percent blockages) maintain the two-phase level to the level of the HL, and as a result the carryover fraction is nearly 1.0. This is because when the two-phase level is at the level of the HL, coolant can easily flow into the HL and to the break. | |||
Essentially, the vessel is full and a mass balance exists, and the total ECCS flow is matched by the total flow out of the vessel. Excess coolant simply flows to the break. Thus, the HL calculations do not indicate the minimum amount of carryover necessary to prevent boron precipitation. | |||
5.4.2.2 Combustion Engineering Plant This section compares a TRACE analysis for CE to the PWROG analysis that used S-RELAP5. | |||
Objectives include: (1) determining how the dependence of the long-term core depression and heatup varies with core blockage ramp times, (2) determining whether there is a ramp rate that would preclude a core uncovery and heatup altogether, and (3) understanding the sensitivity of the results to assumptions about key parameters and boundary conditions. | |||
The TRACE plant model used for these calculations is similar to the St. Lucie Plant. The analyzed event consists of a double-ended guillotine HLB with only high-pressure injection (HPI) available and a containment pressure of 1 atmosphere during long-term cooling. The core inlet is assumed to be completely blocked by debris at some point in time and core cooling water must flow through the core bypass flow path and empty into the top of the core. Sensitivity calculations include varying key input parameters and assumptions including decay heat, bypass flow resistance, HPI flow rate, core blockage time, and blockage rate. | |||
The analyses showed that there is no significant fuel heatup in the long term as the result of core blockage. Sensitivity calculations showed that decay heat, bypass flow resistance, HPI flow rate, core blockage time, and blockage rate could be changed significantly with little sensitivity to the core level after the blockage. The results are qualitatively similar to those of the PWROG analysis. | |||
5.4.3 Applicability to the Pressurized Water Reactor Fleet TRACE sensitivity studies were not performed for the Westinghouse downflow or B&W plant categories. The qualitative results of this analysis should be applicable to all plants that have bypass holes large enough to not be blocked by debris after the switchover to sump recirculation. WCAP-17788 tested openings down to 1/4-inch to demonstrate that these sized openings would not become blocked with debris. Openings less than 1/4-inch were not credited in the WCAP-17788 analyses. Modeling of different plant designs would reveal quantitative differences that depend on the details of the key plant parameters such as specific plant dimensions, core power, bypass resistance, and ECCS flow rate. However, based on the industry TH analyses and the TRACE studies performed for the Westinghouse upflow and CE plants and the similarities in PWR designs, the NRC staff does not believe that additional TRACE studies are necessary. | |||
5.5 Integrated Decisionmaking 5.5.1 Determination of Safety Significance The NRC staff prepared this TER using integrated decisionmaking which includes an assessment of traditional engineering evaluations and probabilistic information. Integrated decisionmaking considers information from a variety of sources including risk insights, traditional engineering evaluations and insights, operational experience and historical plant performance, engineering judgment, and current regulatory requirements. In addition, evaluation of DID and safety margins are an important part of the integrated decisionmaking process. | |||
While a quantitative analysis of the risk associated with IVDEs following a hypothetical LOCA at each U.S. PWR is impractical, the NRC staff combined risk insights with its engineering evaluation, including the low initiating frequency of LOCAs that can challenge LTCC via IVDEs to determine the safety significance of IVDEs. This section discusses the NRC programs that were used to evaluate the safety significance of the issue. | |||
The NRC staff considered guidance that is presented in: LIC-504, Revision 4, Integrated Risk-Informed Decision-Making Process for Emergent Issues (50); NUREG/BR-0058, Regulatory Analysis Guidelines of the U.S. Nuclear Regulatory Commission (51); RG 1.174, Revision 3, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, January 2018 (52); and the Significance Determination Process (SDP) (IMC 0609) (53). Table 4 provides a summary of the risk metrics associated with those programs. The guidelines for assessing risk in the established programs discussed in Table 4 were derived from the Commissions Safety Goal Quantitative Health Objectives (QHOs) and the values were selected to be consistent with the Safety Goal QHOs (see, for example, SECY 97-077, (54)). All these processes use integrated decisionmaking and include considerations of DID and safety margin in addition to risk information to assess appropriate regulatory response relevant to the process (e.g., reactive inspection, license amendment request approval, etc.). As described in NRC guidance, each of these processes are tailored to the context of their regulatory use. For this reason, the extent to which they use the risk assessment in conjunction with the other elements of integrated decisionmaking varies. | |||
Further, the risk assessments used for these processes vary in scope, level of detail, and technical rigor. | |||
For example, RG 1.174 is used to evaluate plant licensing changes. In this process, the change in risk and its comparison against risk acceptance guidelines is one of five key principles that are considered by staff in its evaluation. NRC guidance for performing regulatory analysis for substantial safety enhancements includes risk thresholds for identification of such enhancements (51). The thresholds are starting points and the evaluation includes consideration of available safety margins and DID, which together with the thresholds influence staff decisions. LIC-504 is NRC staff guidance for evaluating emergent issues and refers to the same five key principles defined in RG 1.174 as important inputs to be used in addition to risk values calculated in the process. The SDP uses PRA analyses to evaluate the change in risk resulting from a degraded condition associated with a licensee performance deficiency and relates the significance to established thresholds. The SDP process also includes consideration of qualitative factors as part of the assessment and consideration of safety margins and DID is prominent in cases where the impact of the performance deficiency cannot be readily captured | |||
using PRA analyses. The values in Table 4 are simplified representations of the thresholds from these processes. | |||
Table 4: Risk Metric Lower Risk Risk Metric ( CDF/Rx Year) Greater Risk Reference 10-7 or Lower 10 10-6 10 10-5 10 10-4 10-4 or Higher RG 1.174 Acceptable Evaluate Unacceptable SDP Green White Yellow Red LIC-504 No Action Evaluate (Shut Down if CDF > 10-3) | |||
NUREG/BR-No Action Evaluate Take Action 0058 As stated above, the NRC staff did not quantify the risk attributable to IVDEs following a LOCA. | |||
However, as discussed in Section 5.5.2 the staff determined that the order of magnitude of the occurrence frequency of LOCAs that can challenge LTCC via IVDEs would be 1 x 10-6 per year or less. Assuming that an initiating event (LOCA) that can challenge LTCC via IVDEs leads directly to core damage will result in an order of magnitude of 1 x 10-6 per year or less for core damage frequency. This is consistent with the low end of the quantified risk spectrum for existing NRC programs and corresponding guidance referenced in Section 5.5.1. Additionally, given the high probability of the containment remaining intact following a LOCA including IVDEs, an increase in the frequency of large early releases attributable to IVDEs is not expected (see Section 5.3.1). | |||
Therefore, based on its evaluation summarized in this TER, including the low initiating frequency of LOCAs consequential enough to challenge LTCC via IVDEs (see Section 5.5.2), | |||
DID, and the available safety margins, the staff concludes that a quantitative assessment would indicate that the core damage frequency attributable to the types of IVDEs considered in this TER are of low safety significance. Using integrated decisionmaking, the evidence considered by the staff indicates that the safety significance associated with IVDEs is expected to be in the acceptable range for several existing NRC processes shown in Table 4. | |||
The safety significance approach followed in this TER has not been evaluated for use outside the TER. Therefore, it should not be used on a generic basis unless specifically evaluated for such a use. | |||
5.5.2 Risk Insights - Initiating Event (LOCA) Frequencies Although the NRC staff did not quantify the risk attributable to IVDEs following a LOCA, it sought to understand the frequency of such occurrences to inform its decisionmaking. To do so, the staff sought to determine an order of magnitude estimate for initiating frequency of LOCAs that can challenge LTCC via IVDEs based on the engineering analyses discussed in Section 5.2 of this TER. | |||
The NRC staffs evaluation of the information presented by STP (13) determined that debris generation from a DEGB of a 6 inch or larger pipe on the HL side could result in a fibrous debris bed achieving the threshold limit for challenging LTCC via IVDEs. Section 5.2.1.2 explains how the fiber amount associated with the 6-inch pipe break was determined to be applicable on a generic basis. The 6-inch pipes of interest at STP have an inner diameter of 5.19 inches. | |||
Therefore, the staff sought to determine an order of magnitude estimate for LOCA initiating frequency for breaks 5 inches and larger. Table 7.7 of NUREG-1829, Estimating Loss-of-Coolant Accident (LOCA) Frequencies Through the Elicitation Process, issued April 2008 (55), provides the mean initiating LOCA frequencies for different break categories (each of which corresponds to an effective break size) based on 40-year plant operation. The staff considers the 40-year frequency estimates to be applicable to plants that are operating beyond 40 years. Those values include both piping and non-piping contributions. Table 7.2 of NUREG-1829 provides the ratio of non-piping to piping contributions for each break category. | |||
The non-piping contribution is based on the LOCA initiating frequency estimates at the time of the expert elicitation in NUREG-1829. However, based on information in Sections 6.3.4 and 7.2 of NUREG-1829, the non-piping contribution is expected to remain the same or decrease somewhat. Therefore, the ratio of non-piping to piping contributions for each break category in Table 7.2 were used because they were considered applicable to determining the order of magnitude estimate desired for this evaluation. The resulting initiating frequency for all LOCAs 5-inches or greater is approximately 2 x 10-5 per year as determined by staff calculations. The corresponding value for the arithmetic mean estimates for 40-year plant operation using information from Table 7.11 of NUREG-1829 is approximately 6 x 10-5 per year. Therefore, in both cases, the order of magnitude of relevant breaks is 10-5 per year.ii These estimates are for all piping 5-inches or larger regardless of its location in the RCS. | |||
Since breaks in piping 5 inches or larger on the HL side are consequential to this TER, a refined order of magnitude estimate of the initiating frequency was determined using the expert elicitation data from NUREG-1829 (56). The expert elicitation data were aggregated to determine the initiating frequency of 5 inch or larger breaks in the HL as well as piping connected to the HL such as the surge line. The result is based on the values for 40 years of operation. The overconfidence adjustment derived based on Tables 7.1 and 7.7 in NUREG-1829 was applied. The refined estimate for the initiating frequency for all LOCAs 5 inches and greater on the HL side (i.e., the HL as well as piping connected to the HL) is approximately 1.5 x 10-6 per year (i.e., the order of magnitude is 10-6 per year). This estimate includes consideration of the chemical and volume control system (CVCS) which contributes significantly to the estimate. CVCS lines greater than 5 inches are uncommon and exclusion of the CVCS for the refined estimate results in an initiating frequency for all LOCAs 5 inches and greater on the HL side of approximately 3.6 x 10-7 per year (i.e., the order of magnitude is 10-7 per year). | |||
The size of the break that could generate enough debris to challenge LTCC for a break on the CL side is much larger because smaller breaks do not generate enough debris to block flow to the core. Two cases for very high fiber plants are discussed in terms of generic applicability in Section 5.2.1.1. These demonstrate that only very large breaks, 12 inches and larger, generate significant debris for breaks on the CL side. The staff determined the order of magnitude for breaks in piping 12 inches and larger from Tables 7.7 and 7.2 of NUREG-1829. The resulting initiating frequency for all LOCAs 12 inches and greater in piping is approximately 1 x 10-6 per year. The corresponding value for the arithmetic mean is approximately 6 x 10-6 per year. | |||
ii The estimates in NUREG-1829 are on a per calendar year basis whereas initiating frequencies use a per reactor year basis. The difference between the two is the capacity factor (i.e. the portion of a calendar year that a reactor is at full power). Since the purpose is to determine an order of magnitude estimate, the staff used the calendar year values as if they were on a reactor year basis (i.e. did not account for the capacity factor) because the capacity factor (approximately 0.9 for the entire U.S. fleet) will not impact the order of magnitude estimate and the conclusions drawn therefrom. | |||
Therefore, the order of magnitude of the initiating frequency of the relevant breaks is 10-6 per year. These estimates are for all piping 12 inches or larger, regardless of its location in the RCS. Using the expert elicitation data from NUREG-1829 and the approach outlined in the previous paragraph, a refined estimate of the initiating frequency of breaks in piping 12 inches or larger on the CL side (i.e., the CL as well as piping connected to the CL) is 1.4 x 10-7 (i.e., the order of magnitude is 10-7 per year). | |||
In its comparison with the spectrum of risk thresholds in Section 5.5.1, the NRC staff assumed that an initiating event as defined in this section leads directly to core damage. That is, given a HLB greater than 5 inches or a CLB greater than 12 inches, core damage is assumed. The staff concludes that the margins not included in the assessment of core blockage reduce the risk significantly below that based solely on break frequency and that if core blockage occurs, there is DID available to provide LTCC. | |||
In summary, the estimated order of magnitude of the frequencies of the initiating events (LOCAs) that can generate enough debris to challenge LTCC via IVDEs, along with the sequence of events necessary to result in a high-resistance debris bed at the core inlet discussed in Section 5.2 of this report, supports the staffs conclusions in this TER. | |||
5.5.3 Safety Margins Safety margins associated with the evaluation of IVDEs depend on plant-specific parameters. | |||
The margins associated with the analyses are discussed throughout this document and summarized here. | |||
* This evaluation is performed assuming limiting conditions for debris generation. Many breaks will not generate the amounts of debris assumed. | |||
* The transport assumptions in the evaluation are conservative. All fine fibrous debris that penetrates the strainer is assumed to be transported to the core. All fine particulate debris is assumed to reach the core. Some of this debris would settle or deposit on surfaces. | |||
* The filtering characteristics of the strainer make larger breaks less problematic than would be assumed from their debris generation amounts. For the HLB, a 6-inch break is assumed. Larger breaks may generate significantly more debris, but the amount of debris that passes through the strainer decreases quickly with strainer load. | |||
* The assumed time for switchover to recirculation is conservative. It is based on the smallest RWST water volume and highest ECCS and CSS flows for each plant. The assumed switchover time is 20 minutes for all AFP HLB analyses and all TRACE analyses (21). The fastest time for many plants is about 30 minutes. Times for less limiting but more likely breaks are much longer than 30 minutes. Longer switchover times result in lower decay heat levels which allow AFPs more time to respond to prevent any secondary heatup. Longer switchover times also allow more time for debris to settle. | |||
* Unqualified coatings are assumed to fail as 100 percent fine particulate that is transported to the strainer and core. These coatings would actually fail in a range of sizes and some would not transport to the core. | |||
* The evaluation of particulate debris concentration ignored liquid that would be present in the UP and used a limiting plant value for particulate debris mass. The concentration of particulate will be lower than that calculated for the CLB. | |||
* A 50 percent void fraction is assumed for the core region for the particulate debris concentration calculation. The actual void fraction will be lower and will decrease with time. The concentration of particulate debris for the CLB will be lower than assumed. | |||
* The AFP analysis assumed a uniform bed. This leads to a much lower fibrous debris amount than would actually be required to form a high-resistance bed at the core inlet. | |||
The fiber limits referred to in this analysis are based on a uniform bed. It is likely that a high resistance bed would not form even with significantly more fiber than represented by the limits. | |||
* Chemical effects will probably not occur until after the plants have initiated HLSO. FA testing conducted for WCAP-16793 found that significantly larger amounts of fiber could be accommodated without causing excessive headloss if precipitates are not present. | |||
The staff identified plants that have the potential to incur precipitation before 24 hours. | |||
The PWROG chemical effects testing for WCAP-17788 identified that out of 44 groups, 37 would not incur precipitation before 24 hours. Two plants were identified that could have precipitation before HLSO. Of these, one is a low fiber plant that has already closed GL 2004-02 including IVDEs. The second plant is a single unit for which the staff was unable to confirm that precipitation would occur after HLSO and without more information, the staff considers this unit to be outside the key parameters of this TER. | |||
* In the AFP analysis, the limiting fiber amounts were determined from testing using the limiting flow rate, limiting particulate to fiber ratio, and limiting particulate size distribution. | |||
Changing these parameters resulted in lower head loss values for the same amounts of fiber. | |||
* The core inlet resistance value Kmax was determined using a 60 second ramp rate for increasing the resistance at the time of SSO for Westinghouse and B&W designs. An alternate approach was used for CE designs. This results in a conservative Kmax value, and therefore a conservative fiber limit, as described in Section 5.3.3.5. | |||
* The design of RCS piping assures that it is robust. Implementation of code inspections for Class 1 piping provides some assurance that the piping does not have defects where LOCAs can initiate. Many plants have mitigated dissimilar metal welds to reduce the probability of primary water stress corrosion cracking. For unmitigated welds, the code requires augmented inspections for defects. | |||
* The analysis does not credit leak before break. However, the most likely way for large piping to fail is for a crack to initiate and grow to a critical size, then rupture. All PWRs have systems designed to detect leakage at levels that would occur below the critical crack size. Technical specifications require that plants monitor for leakage and shut down when unidentified leakage indicates potential leakage in the RCS. The intention of the technical specification is to ensure the plant is shut down, leakage identified, and repairs implemented as required, before a LOCA occurs. | |||
* The PCT limit of 800 °F used as an acceptance criterion in WCAP-17788 was based on long-term autoclave testing that showed no significant degradation in cladding properties would occur by spending 30 days at 800 °F. WCAP-16793-NP-A, Appendix A, discusses the basis for the temperature limit. A PCT up to this limit is not expected to challenge fuel integrity. The TH analyses from WCAP-17788 and TRACE show only short-term PCT increases. | |||
* Decay heat in the AFP calculations was modeled similar to the requirements in 10 CFR Part 50, Appendix K (see Section 5.3.3.5). | |||
* The limiting results for Kmax/tblock in the AFP analyses were obtained from minimum recirculation flows, which is a conservative methodology. | |||
* For the limiting AFP analyses, tblock was determined using a conservative ramp rate of 60 seconds for Westinghouse and B&W designs. This results in a conservative tblock value (see Section 5.3.3.5). | |||
* B&W plants exhibited adequate LTCC when the core inlet was blocked in the AFP analysis. The B&W plant response to IVDEs was evaluated assuming a complete blockage at the core inlet beginning at the time of SSO. | |||
* The AFP analyses assumed the limiting FA was operating with a top-peaked power shape, similar to that assumed in small-break LOCA analyses. This maximizes any secondary heatup if the core uncovers during a core inlet blockage. | |||
==6.0 CONCLUSION== | |||
The NRC staff determined that for the operating US PWR fleet, and within the range of assumed plant parameters, debris penetrating the sump strainer is very unlikely to prevent adequate LTCC following LOCAs via IVDEs. This TER considered a wide range of analytical and deterministic information from industry and NRC sources, supplemented by engineering judgement and risk insights. | |||
Based on integrated decisionmaking supported by information from a variety of sources including risk insights, traditional engineering evaluations and insights, operational experience and historical plant performance, engineering judgment, and current regulatory requirements, the NRC staff concludes that the safety significance of the challenge to LTCC following LOCAs via IVDEs is low. The staff believes this analysis is conservative compared to typical operating US PWRs, but it is not an evaluation of any specific plant. It does not address methods for plants to demonstrate compliance with regulations or define licensing basis parameters for the IVDE issue. | |||
The concept of safety significance, and the evaluation methods, findings, and conclusions discussed in this TER are limited in scope to the evaluation of IVDEs with respect to GSI-191 and GL 2004-02. NRC staff does not consider this concept to be applicable to additional issues or topics without further justification. | |||
==7.0 REFERENCES== | |||
: 1. WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid," July 2013 (ADAMS Package Accession No. ML13239A111). | |||
: 2. PWR Owners Group, Submittal of WCAP-17788: "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090)," July 17, 2015 (ADAMS Package Accession No. ML15210A667). | |||
: 3. Title 10 of the Code of Federal Regulations Part 50 "Domestic Licensing of Production and Utilization Facilities". | |||
: 4. Generic Letter 2004-02, "Potential Impact of Debris Blockage on Emergency Recirculation During Design Basis Accidents at Pressurized Water Reactors," September 13, 2004 (ADAMS Accession No. ML042360586). | |||
: 5. WCAP-17360-P, Rev. 0, "Small Scale Unbuffered and Buffered Boric Acid Nucleate Boiling Heat Transfer Tests with Sump Debris in a Vertical 3x3 Rod Bundle," May 31, 2012 (ADAMS Accession Nos. ML12167A435 and ML12167A437) (non-public). | |||
: 6. PWROG-15091-NP, Rev. 0, "Subscale Brine Test Program Report, Analysis committee, PA-ASC-1188," Parts 1 and 2, November 30, 2015 (ADAMS Accession Nos. ML15342A069 and ML15342A070). | |||
: 7. Enclosure, "Revised Content Guide for Generic Letter 2004-02 Supplemental Responses," November 2007 (ADAMS Accession No. ML073110278). | |||
: 8. Nuclear Energy Institute (NEI) 04-07, Vol. 1, "Pressurized Water Reactor Sump Performance Evaluation Methodology," December 2004 (ADAMS Accession No. ML050550138). | |||
: 9. WCAP-16406-P-A, Rev. 1, "Evaluation of Downstream Sump Debris Effects in Support of GSI 191," March 2008 (not publicly available; proprietary information). | |||
: 10. SECY-12-0093 Closure Options For Generic Safety Issue-191, "Assessment Of Debris Accumulation On Pressurized-Water Reactor Sump Performance," July 12, 2012 (ADAMS Accession No. ML121310648). | |||
: 11. Staff Requirements Memorandum SRM-SECY-12-0093, "Closure Options for Generic Safety Issue-191, Assessment of Debris Accumulation on Pressurized Water Reactor Sump Performance," December 14, 2012 (ADAMS Accession No. ML12349A378). | |||
: 12. Final Safety Evaluation for Topical Report WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate Fibrous and Chemical Debris in the Recirculating Fluid," April 8, 2013 (ADAMS Package Accession No. ML13084A161). | |||
: 13. "South Texas Project, Units 1 and 2, Issuance of Amendment Nos. 212 and 198 - Risk-Informed Approach to Resolving Generic Safety Issue 191," July 11, 2017 (ADAMS Package Accession No. ML17019A001). | |||
: 14. WCAP-17788-P, Vol. 5, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090) - Autoclave Chemical Effects Testing for GSI-191 Long-Term Cooling," July 2015 (ADAMS Accession Nos. ML15210A678, ML15210A680, ML15210A681). (non-public). | |||
: 15. WCAP-17788-NP, Vol. 3, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090) - Cold Leg Break (CLB) Evaluation Method for GSI-191 Long-Term Cooling," July 2015. (ADAMS Accession No. ML15210A671). | |||
: 16. WCAP-17788-NP, Vol. 4, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SE-1090) - Thermal-Hydraulic Analysis of Large Hot Leg Break with Simulation of Core Inlet Blockage," July 2015. (ADAMS Accession Nos. ML15210A672 and ML15210A674). | |||
: 17. "Staff Observations of Plant-Specific Testing for Generic Safety Issue 191 During February 26, 2006, Trip to the Alion Hydraulics Laboratory," April 17, 2006 (ADAMS Accession No. ML060750467). | |||
: 18. "Staff Observations of Testing for Generic Safety Issue 191 During April 28, 2006, Trip to the Alion Hydraulics Laboratory," June 27, 2006 (ADAMS Accession No. ML061720514). | |||
: 19. "Seabrook, Unit 1 - Updated Final Response to NRC Generic Letter 2004-02," January 31, 2018 (ADAMS Accession No. ML18031B248). | |||
: 20. "Analysis Report - Post-LOCA PWR Core Inlet Blockage Assessment," February 26, 2007 (ADAMS Accession No. ML070860521). | |||
: 21. "Assessment of Boric Acid Accumulation in a Reactor Core Blocked by Debris," NRC Office of Nuclear Regulatory Research, March 2019. (ADAMS Accession No. ML19086A270). | |||
: 22. Transmittal of WCAP-17057-P/NP, Rev. 1, "GSI-191 Fuel Assembly Test Report for PWROG," PA-SEE-0312, September 2011 (ADAMS Package Accession No. ML112930439). | |||
: 23. EPRI 1011753, "Design Basis Accident Testing of Pressurized Water Reactor Unqualified Original Equipment Manufacturer Coatings," Final Report, September 30, 2005 (ADAMS Accession No. ML071130075) (non-public). | |||
: 24. "Keeler & Long PPG, "Design basis Accident Testing of Coating Samples from Unit 1 Containment, TXU Comanche Peak SES," April 13, 2006 (ADAMS Accession No. ML070230390). | |||
: 25. Attachment 2: Revised Responses to NRC RAIs Related to WCAP-17788 Supporting the Closure of GSI-191 (PA-SEE-1090), September 28, 2018 (ADAMS Accession No. ML18285A019). | |||
: 26. "Vogtle Electric Generating Plant Units 1 & 2, Supplemental Response to NRC Generic Letter 2004-02," April 21, 2017 (ADAMS Package Accession No. ML17116A096) | |||
(non-publicly available). | |||
: 27. Draft Safety Evaluation Report by the Office of Nuclear Reactor Regulation Topical Report WCAP-17788-P, "Comprehensive Analysis and Test Program for GSI-191 Closure," Vol. 5, "Autoclave Chemical Effects Testing for GSI-191 Long Term Cooling,". | |||
January 10, 2019 (ADAMS Accession No. ML19016A310) (non-public). | |||
: 28. Trip Report for May 2013 NRC Staff Visit to the Westinghouse Churchill Laboratory to Observe Autoclave Testing, June 10, 2013 (ADAMS Accession No. ML13159A007). | |||
: 29. Westinghouse Trip Report of November 19, 2013, December 24, 2013 (ADAMS Accession No. ML13337A479) (non-public). | |||
: 30. NUREG/CR-6913, Chemical Effects Head-Loss Research in Support of Generic Safety Issue 191," Argonne National Laboratory, December 2006 (ADAMS Accession No. ML070090553). | |||
: 31. NUREG/CR-6914, Volume 2, Integrated Chemical Effects Test Project: Test #1 Data Report, September 30, 2006 (ADAMS Accession No. ML062560111) (non-public). | |||
: 32. WCAP-16530-NP-A, "Evaluation of Post-Accident Chemical Effects in Containment Sump Fluids to Support GSI-191," March 2008 (ADAMS Accession No. ML081150379). | |||
: 33. Staff Observations of Testing for Generic Safety Issue 191 During July 22 To July 24, 2009 Trip to the Alion Hydraulics Laboratory, September 24, 2009 (ADAMS Accession No. ML092670458). | |||
: 34. "Aluminum Solubility in Boron Containing Solutions as a Function of pH and Temperature, Argonne National Laboratory, September 19, 2008 (ADAMS Accession No. ML091610696). | |||
: 35. Attachment 1, "Responses to NRC RAIs Related to WCAP-17788, Volume 5 Supporting the Closure of GSI-191 (PA-SEE-1090) and Mark-ups to WCAP-17788 PROPPRIETARY Attachment," May 23, 2017 (ADAMS Accession No. ML17293A220) (non-public). | |||
: 36. WCAP-17788-NP, Vol. 1, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090)," July 2015 (ADAMS Accession No. ML15210A669). | |||
: 37. Response to RAIs on WCAP-17788 Vol. 4, B&W Plants Licensing Report, December 31, 2017 (ADAMS Accession No. ML18029A213) (non-public). | |||
: 38. Attachment 1: "Westinghouse Responses to NRC RAIs Related to WCAP-17788, Volumes 1 and 4 Supporting the Closure of GSI-191 (PA-SEE-1090) and Mark-ups to WCAP-17788 Proprietary," December 19, 2017 (ADAMS Accession Nos. ML18029A210 and ML18029A211). (non-public). | |||
: 39. ANP-3583P, Rev. 2, "Response to RAls on WCAP-17788, Volume 4 - CE Plants: | |||
Licensing Report," July 31, 2018 (ADAMS Accession No. ML18285A021) (non-public). | |||
: 40. WCAP-17788-P, Vol. 6, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SE-1090) - Subscale Head Loss Test Program Report," July 2015. | |||
(ADAMS Accession Nos. ML15210A681 and ML15210A682) (non-public). | |||
: 41. WCAP-12945-P-A, "Code Qualification for Best Estimate LOCA Analysis," March 31, 1998 (ADAMS Package Accession No. ML093070051) (non-public). | |||
: 42. Transmittal of WCAP-16009-P-A & WCAP-16009-NP-A, "Realistic Large Break LOCA Evaluation Methodology Using Automated Statistical Treatment of Uncertainty Method (ASTRUM)," January 31, 2005 (ADAMS Package Accession No. ML050910156) (non-public). | |||
: 43. Safety Evaluation for Framatome ANP Topical Report EMF-2103(P) Rev. 1, Realistic Large Break LOCA Methodology for PWRs," Appendix H, "Containment Pressure" (TAC No. MC4259), January 19, 2015 (ADAMS Accession No. ML050190227) (non-public). | |||
: 44. AREVA, Inc. Regarding Final Safety Evaluation for Topical Report EMF-2103(P), | |||
Rev. 3, Realistic Large Break LOCA Methodology for Pressurized Water Reactors,(TAC No. ME2904), June 2016 (ADAMS Package Accession No. ML16286A579). | |||
: 45. Acceptance of Final Safety Evaluation for AREVA NP, Inc Topical Report BAW-10164(P), Rev. 6, RELAP5/MOD2-B&W - An Advanced Computer Program for Light Water Reactor LOCA and Non LOCA Transient Analysis, June 25, 2007. (ADAMS Accession No. ML091410456) (non-public). | |||
: 46. Framatome Technologies, Inc., BAW-10192P-A, Rev. 0, BWNT LOCA: BWNT Loss-of-Coolant Accident Evaluation Model for Once-Through Steam Generator Plants, BAW-10192P-A, Revision 0, June 1998, (ADAMS Accession No. ML093080467) (non-public; proprietary). | |||
: 47. American Nuclear Society, ANS-5 Standards Subcommittee, Proposed ANS Standard, "Decay Energy Release Rates Following Shutdown of Uranium Fueled Thermal Reactors," October 1971. | |||
: 48. American National Standards Institute/American Nuclear Society, ANSI/ANS-5.1-1979, "Decay Heat Power in Light Water Reactors," August 29, 1979. | |||
: 49. "Assessment of Boric Acid Accumulation in a Reactor Core Blocked by Debris," | |||
March 2019 (ADAMS Accession No. ML19086A270) (non-public). | |||
: 50. LIC-504, R. 4 "Integrated Risk-Informed Decision-Making Process for Emergent Issues," May 30, 2014 (ADAMS Accession No. ML14035A143). | |||
: 51. NRC, "Regulatory Analysis Guidelines of the U.S. Nuclear Regulatory Commission," | |||
NUREG/BR-0058, Revision 5, draft for comment, April 30, 2017 (ADAMS Package No. | |||
ML17101A355). | |||
: 52. Regulatory Guide 1.174, Rev. 3, "An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis," January 2018 (ADAMS Accession No. ML17317A256). | |||
: 53. IMC 0609, "Significance Determination Process," October 23, 2018 (ADAMS Accession No. ML18187A187). | |||
: 54. SECY 97-077, "Draft Regulatory Guides, Standard Review Plans and NUREG Document in Support of Risk Informed Regulation for Power Reactors," April 8, 1997 (ADAMS Accession No. ML992920137). | |||
: 55. NUREG-1829, "Estimating Loss-of-Coolant Accident (LOCA) Frequencies Through the Elicitation Process," Main Report, April 2008 (ADAMS Accession Nos. ML082250436 and ML081060300). | |||
: 56. PWR Piping Raw Data for NUREG-1829, June 15, 2011 (ADAMS Accession No. ML080560011). | |||
: 57. WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid," July 2013 (ADAMS Package Accession No. ML13239A111). . | |||
: 58. WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid," July 2013 (ADAMS Package Accession No. ML13239A111). .}} |
Latest revision as of 08:06, 2 February 2020
ML19178A252 | |
Person / Time | |
---|---|
Issue date: | 06/13/2019 |
From: | Victor Cusumano NRC/NRR/DLP/PLPB |
To: | Mirela Gavrilas NRC/NRR/DSS |
Drake J, NRR/DLP, 415-8378 | |
References | |
Download: ML19178A252 (69) | |
Text
June 13, 2019 MEMORANDUM TO: Mirela Gavrilas, Director Division of Safety Systems Office of Nuclear Reactor Regulation FROM: Victor G. Cusumano, Chief /RA/
Technical Specifications Branch Division of Safety Systems Office of Nuclear Reactor Regulation
SUBJECT:
TECHNICAL EVALUATION REPORT OF IN-VESSEL DEBRIS EFFECTS By letter dated July 17, 2015, the Pressurized Water Reactor (PWR) Owners Group submitted a licensing topical report, WCAP-17788, Comprehensive Analysis and Test Program for GSI
[Generic Safety Issue]-191 Closure (Agencywide Documents Access and Management System (ADAMS) Accession No. ML15210A667). WCAP-17788 was intended to support the closure of GSI-191, Assessment of Debris Accumulation on PWR Sump Performance, and Generic Letter (GL) 2004-02, Potential Impact of Debris Blockage on Emergency Recirculation During Design Basis Accidents at Pressurized-Water Reactors, dated September 13, 2004 (ADAMS Accession No. ML042360586), by creating a methodology to define an in-vessel fibrous debris limit to respond to in-vessel questions in GL 2004-02.
The review of WCAP-17788 identified several issues that were not readily resolved. As a result, the staff of the U.S. Nuclear Regulatory Commission (NRC) reevaluated the overall significance of the effects of debris on in-vessel blockage considering improvements in understanding gained in recent years. This is in accordance with Office of Nuclear Reactor Regulation efforts to ensure that agency resources are expended on issues commensurate with their safety significance. As part of that evaluation of safety significance, the enclosed technical evaluation report (TER) evaluated the effects of debris on in-vessel blockage against risk thresholds consistent with the Commissions health objectives. The NRC staff determined that the issue is generally of low safety significance.
CONTACTS: Ashley Smith Stephen Smith Paul Klein Ben Parks 301-415-3201 301-415-3190 301-415-3040 301-415-0979 Matt Yoder Shilp Vasavada Andrea Russell 301-415-4017 301-415-1228 301-415-8553
M.Gavrilas In part, GL 2004-02 requested licensees to evaluate in-vessel effects for their plants. The NRC staff is evaluating compliance with Title 10 of the Code of Federal Regulations 50.46, Acceptance Criteria for Emergency Core Cooling Systems for Light-Water Nuclear Power Reactors, also requested in GL 2004-02, separately from the enclosed TER. The information contained in the TER, however, may be useful to staff and licensees while evaluating compliance with that regulation.
Enclosure:
Technical Evaluation Report
M.Gavrilas
SUBJECT:
TECHNICAL EVALUATION REPORT OF IN-VESSEL DEBRIS EFFECTS Dated: June 13, 2019 DISTRIBUTION:
VCusumano, NRR ASmith, NRR SSmith, NRR NON-PUBLIC SVasavada, NRR RLukes, NRR MYoder, NRR SBloom, NRR ARussell, NRR LPerkins, NRR PKlein, NRR JDrake, NRR MReisi-Fard, NRR JBorromeo, NRR JDonoghue, NRR CHoxie, RES MGavrilas, NRR BBenney, NRR TZaki, RES GCasto, NRR SBajorek, RES JStaudenmeier, RES BParks, NRR ADAMS ACCESSION NO. ML19073A044 *concurred via e-mail NRR-106 OFFICE NRR/DSS/SRXB NRR/DSS/STSB NRR/DMLR/MCCB NRR/DSS/SNPB NAME ASmith* SSmith* PKlein* BParks*
DATE 5/30/2019 5/29/2019 5/31/2019 6/4/2019 OFFICE NRR/DSS/STSB NRR/DRA/RILIT: NRR/DMLR/MCCB NRR/DLP/LA TL (Acting)
NAME ARussell* SVasavada* MYoder* DHarrison*
DATE 5/29/2019 6/3/2019 5/31/2019 6/5/2019 OFFICE QTE NRR/DSS/SNPB: NRR/DRA/APLB:BC NRR/DSS/SRXB BC (Acting) :BC (Acting)
NAME JDougherty* RLukes* MReisi-Fard* JBorromeo*
DATE 3/21/19 6/10/2019 6/7/2019 6/5/2019 OFFICE NRR/DMLR/MCCB: RES/DSA/CRAB: NRR/DSS/STSB:BC BC BC NAME SBloom* CHoxie* VCusumano DATE 6/5/2019 5/19/2019 6/13/2019
- OFFICIAL RECORD COPY -
TECHNICAL EVALUATION REPORT OF IN-VESSEL DEBRIS EFFECTS Enclosure
i Contents 1.0 EXECUTIVE
SUMMARY
.............................................................................................. 2.0 LIST OF ACRONYMS ..................................................................................................
3.0 INTRODUCTION
4.0 BACKGROUND
............................................................................................................ 4.1 Overview of WCAP-17788 .........................................................................................
5.0 TECHNICAL EVALUATION
......................................................................................... 5.1 System Description .................................................................................................. 5.1.1 Key Parameters ................................................................................................ 5.2 Sequence of Events ................................................................................................ 5.2.1 Availability of Emergency Core Cooling System Primary Flowpath ................. 5.2.1.1 Cold-Leg Break ......................................................................................... 5.2.1.1.1 Debris Collection at the Core Inlet ......................................................... 5.2.1.1.2 Boric Acid Precipitation .......................................................................... 5.2.1.1.3 Debris Suspended in the Core ............................................................... 5.2.1.1.4 Local Collection of Debris ...................................................................... 5.2.1.1.5 Concentration of Particulate Debris ....................................................... 5.2.1.1.6 Conclusions for the Cold-Leg Break ...................................................... 5.2.1.2 Hot-Leg Break ........................................................................................... 5.2.1.2.1 Fiber Penetration ................................................................................... 5.2.1.2.2 Boric Acid Precipitation .......................................................................... 5.2.1.2.3 Conclusions for the Hot-Leg Break ........................................................ 5.2.1.3 Upper Plenum Injection Plants .................................................................. 5.2.1.3.1 Conclusions for Upper Plenum Injection Plants ..................................... 5.2.1.4 Debris Generation and Transport .............................................................. 5.2.1.5 Chemical Effects ....................................................................................... 5.2.1.5.1 Testing ................................................................................................... 5.2.1.5.2 Test ResultsPrecipitation Timing ........................................................ 5.2.1.5.3 Staff Analysis of Precipitate Timing ....................................................... 5.2.1.5.4 Chemical Deposits on Fuel .................................................................... 5.2.1.6 ConclusionAvailability of Primary Flowpath ........................................... 5.3 Defense in Depth .....................................................................................................
ii 5.3.1 Containment ..................................................................................................... 5.3.2 Operator Actions .............................................................................................. 5.3.3 Alternate Flowpaths .......................................................................................... 5.3.3.1 Background ............................................................................................... 5.3.3.2 Simplifying Assumptions ........................................................................... 5.3.3.3 Description of Analyses ............................................................................. 5.3.3.4 Analytic Methods ....................................................................................... 5.3.3.5 Conservative Assumptions ........................................................................ 5.3.3.6 Results ...................................................................................................... 5.4 TRACE Sensitivity Studies ...................................................................................... 5.4.1 CLB .................................................................................................................. 5.4.1.1 Westinghouse Upflow Plant ...................................................................... 5.4.1.1.1 Blockage at the Lower Core Plate (99 percent) ..................................... 5.4.1.1.2 Blockage at the Lower Core Plate (99.5 and 99.9 percent) ................... 5.4.1.1.3 Blockage at the Fuel Inlet Nozzles (99 and 100 Percent) ...................... 5.4.1.1.4 Blockage at the Fuel Inlet Nozzle with Delayed HLSO (100 percent) .... 5.4.1.1.5 Core Power Sensitivity ........................................................................... 5.4.1.1.6 Axial Power Shape Sensitivities ............................................................. 5.4.1.1.7 Large Split Break at Top of CL ............................................................... 5.4.1.2 Combustion Engineering Plant .................................................................. 5.4.1.3 TRACE Study Applicability to the PWR Fleet ........................................... 5.4.2 HLB .................................................................................................................. 5.4.2.1 Westinghouse Upflow Plant ...................................................................... 5.4.2.2 Combustion Engineering Plant .................................................................. 5.4.3 Applicability to the Pressurized Water Reactor Fleet ....................................... 5.5 Integrated Decisionmaking ...................................................................................... 5.5.1 Determination of Safety Significance ............................................................... 5.5.2 Risk Insights - Initiating Event (LOCA) Frequencies ....................................... 5.5.3 Safety Margins .................................................................................................
6.0 CONCLUSION
7.0 REFERENCES
1.0 EXECUTIVE
SUMMARY
Generic Safety Issue (GSI)-191, Assessment of Debris Accumulation on PWR [pressurized water reactor] Sump Performance, is a longstanding issue and a brief background of the issue is included in this technical evaluation report (TER). Initial efforts by industry and the U.S.
Nuclear Regulatory Commission (NRC) evaluated sump strainer blockage. More recent testing and analysis focused on the potential for fine debris to pass through the sump strainer and block the reactor vessel (RV) core inlet (in-vessel downstream effects (IVDEs)). The NRC staffs safety evaluation (SE) of the initial Pressurized Water Reactor Owners Group (PWROG) fuel assembly (FA) test program in WCAP-16793-NP-A, Revision 2, Evaluation of Long Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid, issued July 2013 (1) resulted in a core fiber limit that permitted approximately one third of operating PWRs to successfully demonstrate long-term core cooling (LTCC).
A second PWROG in-vessel testing and analysis program, described in WCAP-17788, Comprehensive Analysis and Test Program for GSI-191 Closure, submitted July 2015 (2),
used a design specific approach and considered phenomena timing to justify higher plant specific in vessel fiber amounts. The NRC staff reviewed WCAP-17788 but did not complete its review or determine that WCAP-17788 could be approved for use by licensees. However, during its review of WCAP-17788, the NRC staff identified a significant amount of evidence indicating that potential IVDEs have low safety significance.
The evaluation documented in this TER combines the evidence in WCAP-17788 with the NRC staffs independent analyses and evaluations to provide a more definitive conclusion regarding the safety significance of IVDEs. The staff used risk insights to differentiate the type of evaluation performed based on the likelihood of a break size. For more likely (smaller) breaks, mechanistic analyses for some of the highest fiber plants showed successful mitigation of IVDEs. For less likely (larger) breaks, the NRC staff established that the safety significance is low by considering the low initiating event frequencies and thermal hydraulic (TH) analyses which show that even in the case of substantial core inlet blockage, alternate flow paths (AFPs) through the core barrel-baffle region should be effective for all operating plant designs. All evidence considered indicated that IVDEs are unlikely to compromise LTCC following a loss-of-coolant accident (LOCA).
The staff evaluation considered differences in plant designs, and the differing system responses associated with various break locations. The staff considered cold leg (CL) and hot leg (HL) breaks separately because of differences in plant response and issues (e.g., boric acid precipitation (BAP)) that are more significant for certain types of breaks. Plants with upper plenum injection (UPI) were also considered separately because of their unique design features.
The timing of chemical precipitation was considered in relation to the timing of post-LOCA actions. FA tests have shown that higher fiber quantities can be tolerated if no precipitates are present, while actions such as switching to HL recirculation would mitigate the impact of any precipitates by bypassing any blockage and potentially disrupting bed formation at the core inlet.
This TER applied a novel approach to conclude that the safety significance of IVDEs is low since post LOCA debris is unlikely to result in a loss of LTCC. It is not intended to establish compliance with existing regulations (i.e., Title 10 of the Code of Federal Regulations (10 CFR) 50.46, Acceptance Criteria for Emergency Core Cooling Systems for Light Water Nuclear Power Reactors) (3). The NRC staff is evaluating compliance in an effort separate from this TER. The information contained in the TER, however, may be useful to staff and licensees while evaluating compliance with that regulation.
2.0 LIST OF ACRONYMS ADAMS Agencywide Documents Access and Management System AFP alternate flow path ANL Argonne National Laboratory ANS American Nuclear Society ANSI American National Standards Institute B&W Babcock & Wilcox BAP boric acid precipitation BB barrel/baffle BN bottom nozzle BWNT Babcock & Wilcox Nuclear Technologies CDF Core Damage Frequency CE Combustion Engineering CFR Code of Federal Regulations CL cold leg(s)
CLB cold-leg break(s)
CSS containment spray system CVCS chemical and volume control system DC downcomer DEGB double-ended guillotine break DID defense-in-depth ECCS emergency core cooling system EM evaluation model EPRI Electric Power Research Institute F Fahrenheit FA fuel assembly g/FA grams per fuel assembly GL Generic Letter GSI Generic Safety Issue HPI high pressure injection HL hot leg HLB hot-leg break HLSO hot-leg switchover IBOB initial set of autoclave tests with debris not in mesh bags IVDE in-vessel downstream effects LOCA loss-of-coolant accident LOCADM Loss-of-Coolant Accident Deposition Model LP lower plenum LTCC long-term core cooling NRC Nuclear Regulatory Commission NRR Office of Nuclear Reactor Regulation PCT peak cladding temperature psi pounds per square inch PTFE polytetrafluoroethylene PWR pressurized water reactor PWROG Pressurized Water Reactor Owners Group RAI request for additional information RCS reactor coolant system
RES Office of Nuclear Regulatory Research RG Regulatory Guide RLBLOCA realistic large-break LOCA RV reactor vessel RWST refueling water storage tank SE safety evaluation SG steam generator(s)
SI safety injection SSO sump switchover STB sodium tetraborate STP South Texas Project TER technical evaluation report TH thermal-hydraulic(s)
TSP trisodium phosphate UP upper plenum UPI upper plenum injection WCAP Westinghouse Technical Report Number Preface (formerly Westinghouse Commercial Atomic Power)
ZOI zone of influence
3.0 INTRODUCTION
Each operating PWR licensee is required to demonstrate that LTCC can be maintained as required by 10 CFR 50.46, considering that debris may have an impact on system function.
This provides assurance that the emergency core cooling system (ECCS) strainers will function adequately, that the ECCS and containment spray system (CSS) pump net positive suction head margins are maintained, and that downstream components do not become blocked or damaged by debris that may pass through the strainer.
The NRC staff issued Generic Letter (GL) 2004-02, Potential Impact of Debris Blockage on Emergency Recirculation During Design Basis Accidents at Pressurized Water Reactors, dated September 13, 2004 (4) to request PWR licensees to evaluate the effects of debris on LTCC as stated in 10 CFR 50.46(b)(5). Specifically, 10 CFR 50.46(b)(5) requires that the calculated core temperature shall be maintained at an acceptably low value and decay heat shall be removed for the extended period of time required by the long-lived radioactivity remaining in the core, following the initial system transient that results from a primary system coolant pipe break.
Information in GL 2004-02 summarizes applicable operating experience and previous NRC generic communications, including GSI-191. Considered together, GL 2004-02, GSI-191, and the operating experience referenced within these documents, describe the impact that debris can have on LTCC.
The industry response to GL 2004-02 included: (1) substantial improvements to plant safety via ECCS recirculation sump design modifications and enhancements to containment cleanliness and (2) increases in phenomenological understanding of the various issues associated with LTCC when considering the presence of debris. Examples of the former include the installation of significantly larger ECCS suction strainers and reductions of potential debris sources in PWR containments. Examples of the latter include a large body of experimental testing, advances in investigating the risk significance through probabilistic risk analysis, and investigation of downstream effects using computer codes for TH analysis. However, IVDEs remain unresolved
for approximately two-thirds of the PWR fleet. This report evaluates the overall safety significance of IVDEs.
The NRC staff conducted this technical evaluation in accordance with Office of Nuclear Reactor Regulation (NRR) efforts to ensure that in expending agency resources, the staff gives priority to issues in accordance with their safety significance (see Section 5.5). Given the ongoing technical issues with portions of the TH analysis in WCAP-17788, the staff reevaluated the overall significance of the effects of debris on in-vessel blockage, considering improvements in understanding gained in recent years. Alternative means of demonstrating ECCS performance and compliance with 10 CFR 50.46, as requested in GL 2004-02, are being developed separately, and may be informed by this evaluation.
The NRC staff used an integrated decisionmaking process to make its conclusions regarding IVDEs. This TER relies on engineering judgment, and qualitative and quantitative evidence, including risk insights. The staff approach is provided within the following sections of the TER:
- Section 4.0, Background, provides an overview of the GSI-191 history.
- Section 5.0, Technical Evaluation, describes the scenario that requires evaluation of IVDEs at a high level.
- Section 5.1, System Description, describes the initial plant response to a large LOCA.
This section also includes the key parameters that affect the reliability of the primary flow path and/or DID risk insights.
- Section 5.2, Sequence of Events, describes phenomena that apply to all LOCA scenarios. It also describes how hot-leg breaks (HLBs) and cold-leg breaks (CLBs) affect the system response for typical PWR designs and UPI designs. This section evaluates chemical effects and subscale FA testing and associated analyses. Finally, this section provides the basis for the staff conclusion that the primary coolant flow path (through the fuel inlet nozzles) will provide adequate LTCC.
- Section 5.3, Defense in Depth (DID), discusses plant options to provide LTCC if debris completely blocks the core inlet. The DID topics include containment, operator actions, and AFPs.
- Section 5.4, TRACE Sensitivity Studies, provides a summary of supplemental NRC calculations.
- Section 5.5, Integrated Decisionmaking, defines the criterion NRC staff used to determine safety significance. Section 5.5 also establishes an expected frequency for initiating events that could challenge LTCC due to IVDEs. Finally, Section 5.5 considers margins that provide confidence that the scenarios that are considered to successfully maintain LTCC through the design flow path will be successful.
In addition to the testing and analysis performed by the PWROG in support of WCAP-17788, the NRC staff used additional information and engineering judgment to inform the TER conclusions, as summarized below.
- In WCAP-17788, the PWROG referred to heated rod testing (5) and brine testing (6).
The staff reviewed these test programs to determine the extent to which the information
from them could be used in the TER. The test programs demonstrated phenomenological behavior and are referred to solely in this respect.
- The NRC staff used a combination of information from WCAP-17788 and independent NRC staff TRACE TH analyses to reach the conclusions in this TER. The WCAP-17788 analyses were performed using computer codes approved for some aspects of LOCA analysis, but not specifically for application to LTCC. The novelty of the application led to significant questions during the WCAP-17788 review. These issues and the staff approach for addressing them are discussed further in Section 5.3.3. The analyses were not typical of licensing basis LOCA calculations in that they were not plant specific and lacked rigorous qualification for some key phenomena in the models. The staff determined that the results are representative of the classes of plants being modeled.
The staff used the analyses to determine acceptable HLB fibrous debris loads, demonstrate typical flow patterns within the RV, and demonstrate AFP viability. The staff considers the use of the WCAP-17788 analysis results, in combination with the independent NRC staff TRACE results, sufficient to support the TER conclusions.
This TER is an evaluation of the general safety significance of IVDEs. The staff believes this analysis is significantly conservative compared to typical operating US PWRs, but it is not an evaluation of any specific plant. It does not address methods for plants to demonstrate compliance with regulations or define licensing basis parameters for the IVDE issue.
4.0 BACKGROUND
As discussed above, GL 2004-02 required plants to evaluate the effects of debris on LTCC as defined in 10 CFR 50.46.
In the NRC document, Revised Content Guide for Generic Letter 2004-02 Supplemental Responses, issued November 2007 (7), the NRC staff identified several areas for licensees to evaluate to ensure that plants have the ability to mitigate LOCAs. All plants performed evaluations to address the areas delineated in the content guide. The technical areas of the content guide fall into two general categories, strainer and downstream effects. The downstream effects are further split into in-vessel effects and effects on other equipment.
Industry developed guidance to address evaluation of the strainer and downstream effects on equipment (Nuclear Energy Institute 04-07, Volume 1, Pressurized Water Reactor Sump Performance Evaluation Methodology, issued December 31, 2004 (8), and WCAP-16406, Revision 1, Evaluation of Downstream Sump Debris Effects in Support of GSI 191, issued March 2008 (9)). The NRC endorsed this guidance with modifications.
All licensees addressed aspects of the effects of debris on LOCA mitigation. For example, all operating plants installed larger strainers and implemented actions to limit the introduction of debris into containment, and to ensure that debris is removed from containment prior to restarting units. The upgraded strainers installed in the plants have larger surface areas to accommodate debris that may arrive and have smaller screen openings when compared to earlier strainer designs. These smaller openings decrease the size distribution of debris that may pass downstream and enter the reactor core or affect other equipment.
Most PWRs provided sufficient information to demonstrate that their strainers will function adequately under conditions caused by the most challenging LOCAs for their plant. A few are still completing evaluations, and some are supplementing deterministic evaluations with
risk-informed analyses. All licensees have evaluated ex-vessel downstream effects and used NRC approved WCAP-16406-P-A (9), to evaluate the effects of debris on pumps, valves, spray nozzles, orifices, and instrumentation. to SECY-12-0093, Closure Options for Generic Safety Issue-191, Assessment of Debris Accumulation on Pressurized-Water Reactor Sump Performance, dated July 9, 2012, (10) provides a history of GSI-191. In the staff requirements memorandum (SRM) (11) associated with SECY-12-0093, the Commission directed the NRC staff to use risk insights to close GL 2004-02. The SRM also stated that based on plant actions there is adequate DID to allow the plants time to demonstrate compliance with the regulations. Industry continued testing and analysis to demonstrate that in-vessel debris loads are acceptable and to show acceptable strainer performance. In accordance with the SRM, approximately six licensees elected to use risk-information to show acceptable system performance.
Industry also developed guidance (WCAP-16793-NP-A) (1) for the evaluation of the effects of debris in the RV on LTCC. The NRC approved this guidance in an SE (12). The debris limits approved in the guidance were based on bounding assumptions that cover all plant designs.
WCAP-16793 was able to be used by 19 PWR units to conclude that in-vessel debris would not inhibit LTCC. The NRC staff is aware of five additional low fiber units that could use the WCAP-16793 guidance to demonstrate acceptable in-vessel performance but have chosen not to in order to gain larger margin to debris limits. The NRC staff recognizes that the WCAP-16793 limits are conservative because the methods used to develop them were greatly simplified to reduce resource expenditure and to assure that test conditions bounded the PWR fleet. Two additional units performed plant-specific TH analyses to demonstrate that LTCC would be assured, even if a greater amount of debris than approved in WCAP-16793-NP-A was transported to the core (13).
The licensees for the remaining units have calculated, or assumed, that the amounts of debris that will pass through the strainer and enter the reactor coolant system (RCS) are greater than the approved WCAP-16793 limit. The PWROG developed and submitted WCAP-17788 (2) to demonstrate on a plant-specific basis that greater amounts of debris in the core would not adversely affect LTCC. WCAP-17788 incorporated results from FA testing, chemical effects testing, TH analysis, and additional mathematical modeling. It also referred to other work that increased the knowledge of the effects of debris on LTCC, particularly in the RV.
WCAP-17788 evaluated chemical effects, HLBs, and CLBs separately. Instead of using a bounding analysis as in WCAP-16793-NP-A, the PWROG used design-specific information to show that larger amounts of debris could be accommodated without inhibiting LTCC. The PWROG differentiated the analyses between four reactor designs and accounted for differences in fuel design. The method also allowed plants to consider plant-specific design characteristics such as ECCS flow rates, strainer size, chemical effects, and the number of FAs in the core.
The analyses made bounding assumptions for some of the inputs like decay heat curves, timing of switchover to recirculation, FA flow velocity, and debris mixture arriving at the core. The analyses relied on a relatively complex TH analysis to demonstrate that LTCC could be maintained when the core inlet flow was impeded by debris following switchover to sump recirculation. The analyses also assumed that debris eventually completely blocked the core inlet.
4.1 Overview of WCAP-17788 The NRC staff did not complete its review of WCAP-17788, and thus, it is not approved for use.
However, the NRC staff relied on the analyses, evaluations, and experimental data submitted in WCAP-17788 to support its evaluation of the safety significance of IVDEs. WCAP-17788 proposed a methodology to use plant-specific analyses to determine in-vessel debris amounts for each plant based on plant-specific characteristics (e.g., plant-specific decay heat, strainer characteristics, debris transported to the strainer, flow rates, time of sump recirculation, activation of AFPs, formation of chemical precipitates, and implementation of measures to prevent BAP). Once the mass of fiber entering the RV is established, it is compared to an acceptance criterion calculated for each plant and fuel design to determine if the calculated fiber load would inhibit LTCC.
The NRC staff reviewed WCAP-17788 and developed numerous requests for additional information (RAIs), especially about the TH analysis. The PWROG used computer codes used in ECCS evaluation models (EMs) that had been found acceptable for short-term system response analysis and attempted to determine a bounding peak cladding temperature (PCT) to demonstrate LTCC adequacy. The staff found that the codes used for the analyses, while previously approved for short-term LOCA analyses, had not been accepted by the NRC staff to evaluate the phenomena associated with LTCC including predicting a PCT to demonstrate adequate LTCC. In some cases, the results indicated nonphysical behavior. To allow plants to resolve the issues individually despite these uncertainties, the NRC staff drafted limitations on the use of WCAP-17788. These limitations would have required significant plant-specific analyses.
5.0 TECHNICAL EVALUATION
The scenario of relevance for evaluating IVDEs is initiated by a primary coolant system pipe rupture or break that leads to a LOCA. Because reactor coolant is discharged through the pipe break, debris is generated within containment. Although the ECCS is initially aligned to a reservoir (e.g., borated water storage tank or refueling water storage tank (RWST)), once the liquid in the reservoir is depleted, the ECCS suction must be realigned for recirculation by switching the source from the depleted reservoir to the containment sump (i.e., sump switchover (SSO)), which collects primary system liquid that has spilled out of the pipe break or collected from containment spray.
Under recirculation cooling, the coolant that spills out of the break and returns to the containment sump can pick up debris dislodged by the initial pipe break, as well as other debris that may be in containment. The system includes a strainer that is intended to collect most of the debris, but some debris will pass though the strainer, and be pumped into the RCS. Once within the reactor, the debris will transport through the RV. Design features like FA debris filters may collect some of the debris, while other debris may settle in more stagnant regions of the system. If significant amounts of debris collect at the debris filters, the flow resistance at the core inlet may increase, challenging the adequacy of LTCC.
The fluid initially present in the RCS is borated water. Injected emergency coolant is also borated water. Coolant flows out of the break and containment spray interacts with plant materials as it is transported through the containment to the sump, creating a pool with additional dissolved species. Chemical effects include the formation of precipitates and other
chemical products that could be filtered by a fibrous debris bed and further impair the flow of coolant needed for core cooling. An additional concern in the core region for CLBs is the potential for boric acid to concentrate due to boiling associated with the decay heat of the fuel, potentially resulting in BAP. A key mitigating measure is the use of HL switchover (HLSO) (or equivalent) after a period of time to flush the core by injecting coolant into the HL, rather than the CL. HLSO (or equivalent) is required by all PWR Emergency Operating Procedures with its timing dependent on plant-specific evaluations.
WCAP-16793 stated that the following criteria must be met to ensure that debris in the core will not adversely affect LTCC:
- Adequate flow for decay heat removal is supplied to the core
- BAP is prevented The staff agrees that if these are met, LTCC will be successful.
The method used by the staff to evaluate this issue includes the premise that it is unlikely a LOCA large enough to generate and transport enough debris to the core to prevent LTCC will occur. The evaluation considered the low probability of the initiating event in combination with the low likelihood of the formation of a debris bed across the entire core. Sections 5.2.1 and 5.5.2 discuss these parameters.
The NRC staff also has significant knowledge of how the system will respond to varying amounts of debris (or resistance to flow) at the core inlet, including complete blockage at the fuel inlets. The knowledge base for these conditions consists of test and analytical information.
In some cases, the test results were used as inputs to TH analyses. Section 5.2.1 discusses this information.
Table 1 below lists some of the major phenomena associated with the evaluation of IVDEs. The table is designed to demonstrate the differences or margins between the assumptions used in the safety significance evaluation and the expected realistic behavior. To the extent possible, the phenomena are listed sequentially as a LOCA scenario would progress in the plant. Many of the phenomena overlap in time.
Table 1: LOCA Phenomena Sequence of Attribute Assumption Most Probable Margin Events Initial LOCA Size 6-inch <6-inch Less debris is likely Recovery More debris settles Recirculation before reaching sump, Switchover 20 minutes >30 minutes less debris, lower decay Time heat Unqualified Time Some fail as chips, less Coating 100% Fine <100%
dependent debris transports Failure Switchover Some fines and Early Debris 100% Fines trapped, including Recirculation
<100% transports Transport Transport debris recirculating from CS and break Fiber <15%, decreases Depends on plant 15%
Penetration with debris amount design Core Inlet More debris required Uniform Non-Uniform Fiber Bed to block core Conservative Higher flow, flow, particulate FA Debris different Debris limits from amount, Early Testing particulate testing are conservative particulate Recirculation characteristics sizing to Long-TH Analysis Assume Slow increase in Additional Term for Debris instantaneous core inlet conservatism in debris Limit blockage resistance limits Core inlet unblocked at Chemical fiber limit w/o chemical Effects 6 Hours 24 Hours effects, lower decay Timing heat when AFP is sole Long-Term flowpath Effects Hot Leg Consider actual plant 24 Hours Plant specific, most Switchover timing, bypasses debris Maximum less than 10 hours1.157407e-4 days <br />0.00278 hours <br />1.653439e-5 weeks <br />3.805e-6 months <br /> Timing bed
5.1 System Description This description applies to all LOCA locations. There are two separate categories for LOCA break locations depending on whether the break is upstream or downstream of the core (CL side or HL side). The break locations are called HLBs and CLBs. Most plant responses are similar for both categories. Because of differences in plant response to breaks on the hot and cold side of the RCS, HLBs and CLBs are evaluated separately. The system responses to HLBs and CLBs are primarily described in Section 5.2. However, aspects of the LOCA response are discussed throughout this document.
The initial plant response for LOCAs is similar regardless of whether they occur on the HL or CL side of the RCS. Following a large LOCA, the CSS is actuated to suppress containment building pressure, and the ECCS is actuated to replenish liquid in the core and remove decay heat. Some breaks, depending on the break size and the plant setpoint, will not initiate CSS.
Initially, the water source is from stored locations (e.g., the RWST). The borated water from these stored locations refloods the core within a few minutes, decreasing the cladding temperatures to within a few degrees of saturation. Water that is pumped into the RV is discharged through the break into containment, where it collects on the containment building basement floor and in the ECCS sump(s). The time required for the stored water to be depleted depends on the size of the break, the volume of fluid in the tank, and the ECCS and CSS pump capacities. This coolant injection phase can end as early as 20 minutes following the initiating event, if maximum ECCS and CSS flow and minimal stored water volumes are assumed.
After the stored water supply is exhausted, the CSS and ECCS are realigned to draw coolant from the containment sump. The coolant discharged from the RCS and CSS is recirculated into the RCS to provide for continued LTCC without the need for additional cooling water. Most plant responses are similar for HLBs and CLBs, however, Westinghouse 2-loop plants have a significantly different ECCS design known as UPI. Section 5.2.1 describes the responses of the different plant types to breaks in different locations.
Debris-laden coolant may begin to enter the RCS once the ECCS suction source is switched from the RWST to the containment sump. The amount of debris that enters the RCS depends on the filtering efficiency of the strainer, debris generation amounts, ECCS configuration, break location, amount of flow through the CSS, and the amount of time coolant is being drawn from the sump. Some debris that enters the RCS will reach the RV.
During CL recirculation, the flow transports debris through the downcomer (DC) into the lower plenum (LP) and into the core through the core inlet support region. After recirculation is switched to the HL, flow transports debris into the UP and into the top of the core. Depending on the break location and the injection location, some debris may flow out of the break.
Debris may collect at the bottom nozzle (BN) of the FA, causing resistance to flow. In some cases, it may collect at spacer grids higher in the FA. Ultimately, if enough debris is collected on the BNs, flow through the core inlet could be impeded, compromising LTCC.
All U.S. PWRs have AFPs in the RV that provide a path for coolant to reach the core if the core inlet becomes completely blocked with debris. Examples of AFPs include the barrel baffle (BB) region, pressure relief holes, upper head spray nozzles, and flow over the top of the steam generator (SG) tubes. The types of AFPs and resultant amount of flow area they provide is plant-specific.
5.1.1 Key Parameters The NRC staff reviewed the effects of debris on LTCC with respect to debris in the RCS for the current operating fleet of PWRs in the U.S. The review is based on inputs in WCAP-17788, information provided by licensees, and other information available. It is considered representative of US PWRs, if not necessarily bounding. The following are the key parameters considered by the staff as they evaluated the general safety significance of the issues.
Consideration of these parameters may be useful as licensees and staff evaluate plant-specific compliance.
(1) Plants have a method of diluting the coolant in the core by implementing HLSO or equivalent action. This addresses the potential for BAP and concentration of other debris types in the reactor and reduces the potential for chemical effects to result in complete blockage due to a debris bed at the core inlet.
(a) To allow the [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx] CLB limit, plants other than UPI initiate HLSO before chemicals may form as per Volume 5 of WCAP-17788 (14).
See Section 5.2.1.5.
(b) To prevent excessive concentration of particulate debris in the core, HLSO is initiated within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. See Section 5.2.1.1.3.
(c) To prevent BAP, plants initiate HLSO in accordance with their licensing basis.
See Section 5.2.1.1.2.
(2) Chemical effects timing (tchem) relative to the earliest time that complete core blockage can occur while not compromising LTCC (tblock). See Sections 5.2.1.5 and 5.3.3.3.
(3) Plant-specific post-LOCA environment is consistent with WCAP-17788 Test Group. See Section 5.2.1.5.
(4) Earliest time to swap to recirculation. See Section 5.3.3.2.
(5) Particulate debris amounts reaching the core pertaining to CLB. See Section 5.2.1.1.5.
(6) Fibrous debris amounts reaching the core pertaining to CLB and HLB. See Sections 5.2.1.1, 5.2.1.2, and 5.3.3.6.
(7) Minimum ECCS flow. See Section 5.3.3.5.
(8) FA debris capture characteristics as evaluated in WCAP-17788.
(9) AFP resistance values. See Section 5.3.3.2.
(10) Available driving head for the CLB. See Section 5.2.1.1.1.
(11) RCS liquid mass. See Section 5.2.1.1.5.
(12) LOCADM results. See Section 5.2.1.5.4.
(13) LOCA frequency. See Section 5.5.2.
These key parameters are considered essential considerations for the NRC staffs safety significance determination. Some of these parameters, or others not listed here, may be important to individual licensees LTCC adequacy evaluations.
5.2 Sequence of Events This section describes phenomena that apply to all LOCA scenarios. It also describes how HLBs and CLBs affect the system response for typical PWR designs and UPI designs. Finally, this section provides the basis for the staff conclusion that it is unlikely the primary coolant flowpath (through the fuel inlet nozzles) will become blocked to an extent that would compromise the adequacy of LTCC.
The break size and location govern the characteristics of the initial system response, as well as the amount of debris generated. These must be considered in the IVDE evaluation.
The IVDEs associated with larger breaks are assumed to be more challenging to mitigate than smaller breaks. A complete, double-ended guillotine break (DEGB) would have the highest energy discharge and the greatest potential to generate the most debris. In addition, the coolant pressure and inventory are lost most quickly following the largest break, and the source of pumped injection flow is hence depleted the most quickly, leading to the earliest SSO time.
These dynamics lead to a scenario in which recirculating coolant would have the greatest concentration of debris, and the debris-laden coolant would be delivered to the core at the earliest time, while the core decay heat levels remain high. The resulting scenario would require the greatest coolant flow rate, yet the high concentration of debris would have the highest likelihood of causing a blockage that could inhibit adequate core cooling.
Smaller breaks, by comparison, would tend to generate less debris and deplete the source of injectable coolant more slowly. As a result, the recirculation would begin later, and would feature a lower concentration of debris. Longer times to recirculation also provide additional time for debris to settle in the post-LOCA sump pool.
Based on the evaluation described above, the NRC staff determined that the following assumptions were reasonable to use when evaluating the safety significance of IVDEs:
- Break size and SSO time are inversely correlated; larger breaks would be associated with faster switchover to recirculation, and smaller breaks would allow for SSO when the decay heat has dissipated more.
- Break size and debris generation are directly correlated; larger breaks discharge coolant at a faster rate resulting in more damaging jets and greater damage to materials in containment.
5.2.1 Availability of Emergency Core Cooling System Primary Flowpath The analyses in WCAP-17788, Volume 3 (15) and Volume 4 (16), as well as the earlier work documented in WCAP-16793-NP-A, conservatively assumed that debris would collect uniformly at the core inlet because that results in the greatest headloss for a given amount of debris.
However, the debris bed is realistically expected to collect non-uniformly for reasons described below. As a result, the amount of debris required to completely block the core inlet would be greater than that assumed in the analyses.
If a debris bed does not form at the core inlet or other locations that affect cooling, current LTCC analyses remain applicable. However, based on testing (WCAP-16793-NP-A (1) and WCAP-17788 (2)), it was determined that a fiber bed could form at FA inlets. This testing was
conducted with a single FA or portion of an FA and assumed uniform flow into all areas of the core.
WCAP-17788 established debris limits of [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. The HLB debris limit is based on subscale FA testing and the TH analyses performed for WCAP-17788. The CLB limit applies to all RCS designs. The HLB fiber amount referenced in this section is based on the combined RCS and fuel design with the smallest debris limit. These debris limits assume that the debris bed forms uniformly as it did during reduced-scale FA testing performed as part of the WCAP-17788 analyses.
The PWROG TH analysis and the NRC TRACE evaluations (See Section 5.4) predict significant variations in flow rate and direction at the core inlet, which will prevent the deposition of a uniform debris bed. This makes complete blockage of the core inlet less likely. The flow in the center of the core will be up (from the LP to the core), while the flow at the periphery will be down, (from the core to the LP). The patterns are caused by higher decay heat in the central core and relatively low decay heat in the periphery. Contributing to this flow pattern is the downward flow entering the LP from the DC. These flow patterns will cause the deposition of fiber at the center of the core inlet first. As debris builds in the areas with upward flow, some flow redistribution to the periphery may occur.
Other influences on flow will tend to prevent uniform flow to the core. For example, the lower core plate, core support plate, and other hardware within the LP disturb the flow patterns at the core inlet. These contribute to non-uniform arrival of flow and debris at the core inlet. This phenomenon was observed during brine testing (PWROG-15091, Subscale Brine Test Program Report, dated November 30, 2015 (6)) in which it was difficult to form a uniform bed at the FA inlet.
The NRC staff observed similar non-uniform debris bed formation during several strainer tests, which provides evidence that similar behavior will occur at the core inlet (17) (18) (19). The velocities approaching the disks in most strainers are non-uniform because the plenum that collects the flow that passes through the perforated disks is hydraulically closer to some surfaces than others. The velocities closer to the plenums are higher. The velocity profile determines the degree of non-uniformity. Greater disruptive forces are present for the in-vessel case. Some of the flow is down, or away from the core inlet. This prevents deposition of debris in those areas. The flow velocity at the core inlet can vary considerably depending on the ECCS design and the break location. CLB flow velocities are low, while HLB velocities can be much higher if all pumps are injecting. CLB average core inlet velocities are similar to strainer approach velocities. HLB velocities are generally higher.
The staff considered that chemical precipitates in the debris bed could cause the bed to deposit more uniformly over the core inlet by increasing the differential pressure across the bed and causing flow to be redirected. However, as discussed in Section 5.2.1.5, precipitation is expected to occur later in the transient. By the time precipitates may form, the debris will be depleted from the pool so that further transport to the core inlet is significantly reduced.
Testing for WCAP-17788 indicates that the debris deposited at the core inlet would not cause a significant resistance to flow until the fiber limit is approached. This is consistent with testing from WCAP-16793 that showed there was a threshold fiber amount below which significant headloss would not occur. Therefore, significant redistribution of the debris will not occur early in the scenario. The design of the PWR core also permits crossflow between FAs which allows
some equalization of the differential pressure in the areas where debris is deposited under the low-velocity conditions following a LOCA.
When considering a full core relative to the subscale testing, it is expected to take more debris than the limits in WCAP-17788 to prevent coolant from reaching the core. In the event the core inlet accumulates enough debris to stop flow, AFPs are available to provide adequate core cooling. Section 5.3.3 discusses these AFPs. The debris limits assume that chemical effects have not occurred. Section 5.2.1.5 discusses chemical effects. In addition, once a plant initiates HL recirculation or a similar BAP prevention action, adequate LTCC is maintained.
HL recirculation is also likely to dislodge portions of the debris that may collect at the core inlet.
The NRC staff considers the assumption that all debris passing through the strainer reaches the core inlet to be conservative. For breaks large enough to initiate the CSS, the amount of fiber reaching the core inlet would be reduced by the amount of fiber diverted to the CSS. For smaller breaks, the CSS may not initiate; however, the overall amount of debris generated would be less.
Evaluations conducted by the PWROG and the NRC demonstrated that even a very small open flow area at the core inlet will allow adequate coolant to enter the core. One evaluation conducted for WCAP-16793-NP-A (1), Appendix B, showed that up to 99.4 percent of the inlet could be blocked and flow to ensure core cooling would still be adequate. Calculations performed by the NRC Office of Nuclear Regulatory Research (RES) confirmed this evaluation (20). Later, more detailed TH evaluations performed by RES showed that 99 percent of the flow area could be blocked for HLBs and CLBs while ensuring adequate flow and prevention of BAP (21). This 99 percent flow area was converted to a core inlet resistance and applied uniformly over the core inlet area in the RES model.
An additional source of information regarding debris bed formation at the core inlet is testing conducted for WCAP-17360-P, Small Scale Unbuffered and Buffered Boric Acid Nucleate Boiling Heat Transfer Tests with Sump Debris in a Vertical 3x3 Rod Bundle, dated May 31, 2012 (5). This WCAP discusses 3x3 heated rod bundle testing that was conducted with fibrous and particulate debris. The BN and first spacer grid provided the most likely places for debris to be captured. Capture of fibrous debris at spacer grids was observed during testing for WCAP-16793. During the heated rod testing, there was no debris capture at the spacer grids except for a small amount of debris captured between the outer surface of the grid and the test column. Although the testing did not include a prototypical geometry at the approach to the fuel, the test results suggest that turbulence may help to inhibit debris bed formation. The heated rod testing provided qualitative evidence that turbulence will tend to prevent debris bed formation, especially at spacer grids.
As part of a DID analysis, the NRC staff considered how the LOCA response would evolve if a high-resistance debris bed were to form at the core inlet. Section 5.3.3 discusses the availability of AFPs that bypass the normal core inlet.
5.2.1.1 Cold-Leg Break Following a CLB, for all PWR designs except UPI, the ECCS is injected into the CL and flows down the DC into the LP and up into the core. Only the flow necessary to make up for inventory in the core region lost to boil-off enters the core. The rest of the flow goes out of the break in the CL and spills back to the sump. The amount of debris reaching the core inlet for the CLB is
much less than for the HLB because most of the debris flows out the break. For the CLB, the force driving the coolant into the core is the manometric balance between: (1) the level in the CL (up to the break elevation) and DC, and (2) the level in the core. As the water boils out of the core it is replaced because of the gravity head in the DC. There are also flow losses from liquid and vapor flows. The amount of coolant required to replace the boil-off is higher at the beginning of the event and is reduced as decay heat decreases. Flow to the CSS does not enter the core and is returned to the sump with any entrained debris.
For a CLB, when injection is switched to the HL, the discharge of the ECCS pumps is aligned to the UP (with pressure limited by the height of the SG tubes). This flow will reverse the flow at the core inlet, flush concentrated debris and boric acid from the core, and disrupt any debris bed at the core inlet.
The NRC staff determined that it is unlikely for a debris bed to form across the entire core inlet following a CLB. Circulating currents will prevent or limit debris deposition at the core periphery.
Maintaining a small open area at the core inlet will assure adequate flow to cool the core. The concepts outlined here are the same as those discussed in Section 5.2.1 and are not repeated in detail.
The PWROG submitted an evaluation of the effects of debris on a CLB in WCAP-17788 and updated the evaluation with supplemental information. This establishes a basis for a debris limit for the CLB case. In the analysis, the AFP credited in the HLB was assumed to be unavailable, and all flow to the core was assumed to flow through the core inlet. The CLB debris limit was set at [xxxxxxx].
Boiling in the core is vigorous at the beginning of recirculation, which results in sufficient liquid carryover to limit the buildup of boric acid in the RV and adequate mixing in the RV to maintain homogeneous boron concentration. The WCAP-17788 TH analysis shows upflow in the center of the core and downflow in the periphery. TRACE studies predict similar flow patterns at the core inlet (see Section 5.4). The circulating currents result in the transport of boron from the core region to the LP. The boron concentration in the core may increase until the density of the fluid in the core is high enough to result in density driven convection between the core and LP.
As boiling decreases, convection can become the dominant mixing mechanism. If the boron concentration in the RV becomes too high, boron precipitation could occur which would affect the ability to cool the fuel. The TRACE studies also predict that mixing between the core and LP will occur. The LP provides a volume for dilution of boric acid and other debris that may be suspended in the core.
Debris will begin to arrive at the core inlet shortly after recirculation begins and most debris will reach the core inlet within a few hours. Most will arrive relatively early because the strainer will become more efficient at filtering over time as debris collects on it. Debris will also be removed because of collection in the core or settling elsewhere throughout the sump or RCS.
The following topics are addressed to evaluate the fiber limit of [xxxxxxx].
- debris collection at the core inlet
- resulting in increased pressure losses
- preventing transport of high-concentration boron solution from the core to the LP
- debris suspended in the liquid phase
- increasing pressure losses through the inner RV because of changes in fluid properties and two-phase flow characteristics and changes in heat transfer processes
- causing early BAP because of reduced liquid inventory, reduced liquid carryover, or changes to the precipitation mode or precipitation location
- debris deposition on spacer grids in the core increasing frictional losses or reducing heat transfer Note that UPI plants are treated differently because of the location of the ECCS injection.
Section 5.2.1.3 discusses UPI plants.
A greater amount of debris generation is required to inhibit flow for a CLB than for a HLB. This is based on the distribution of flow after it passes through the strainer. For a HLB, all ECCS is pumped through the core. For a CLB, only the volume of water required to replace that boiled out of the core flows to the core inlet. The remainder of the flow goes out the break, and with it any debris that is entrained in the flow. The debris that flows out of the break is returned to the sump pool and has an opportunity to enter the ECCS again. However, testing has shown that the filtration efficiency of the strainer for fiber increases quickly as fiber is captured on the strainer surface.
As an example, South Texas Project (STP) performed studies to determine how much debris could arrive at the core inlet under various conditions. STP has among the highest fiber amounts in containment compared to the other U.S. PWRs because it uses only fibrous insulation in its containment. The studies were based on strainer penetration test results and modeling of the flow as it circulated through the strainer, the ECCS, and the CSS. The model found that for a design basis case with a single failure (one train of ECCS and CSS running),
376 grams of fiber would arrive at the core inlet. Without CSS (only low-head safety injection and high-head safety injection running), the value was 539 grams. For a single low-head safety injection pump running using the most conservative penetration results, the debris amount arriving at the core was 1,119 grams. This amount of debris is assumed to be evenly distributed among 193 FAs. The FA loading for the single low-head safety injection case is very low at 5.8 g/FA. The NRC staff noted that it is unlikely that only a single low-head safety injection pump would be operating during recirculation, but even if that is the case, the debris loading is well below the level of concern. Pumps may be secured to reduce flow through the strainers, but this is generally done at a time when most of the debris has been removed from the sump fluid by filtering. As can be seen above, the more realistic cases have lower debris amounts.
STP also performed beyond design basis sensitivity studies to determine the most challenging flow configuration for CLB in-vessel debris. STP established a strainer debris limit of 192 pounds of fine fiber based on strainer headloss testing. LOCAs that resulted in greater than 192 pounds of fine fiber at the strainer were assumed to contribute to core damage. As shown in the analysis for the license amendment request (13), 192 pounds of fine fiber could result from a 12.8-inch DEGB. However, most breaks generating this much fiber were significantly larger. A few of the largest breaks at STP could result in fine fiber amounts of up to 830 pounds.
The beyond design basis sensitivity studies were based on fine fiber debris loads at the strainer of 192 pounds (design basis limit) and 800 pounds. The 800-pound value is representative of the amount of fine fiber that can arrive at the strainer from the largest LOCA breaks at STP.
The sensitivity studies demonstrated that even with the largest strainer fiber loading and overly conservative flow conditions, the CLB limit was not reached for STP. It was determined that lower strainer flow rates resulted in higher core debris amounts. For example, the highest debris amount was calculated assuming one-half high-head safety injection flow and no other pumps in service. The result was 2,400 grams of fiber at the core. For STP, this results in 12.4 g/FA. It was further determined that the highest core loading would occur with the largest number of strainers in service with the lowest possible flow through each.
Vogtle Electric Generating Plant performed a similar evaluation for the CLB, also using plant-specific penetration testing and modeling. Vogtle is also a very high-fiber plant. Instead of using a plant-specific debris amount in the testing, a theoretical strainer debris bed thickness of one inch was attained for each test by batching in small amounts of fiber. Testing demonstrated that additional debris did not penetrate the strainer, even before the one-inch threshold was reached. Most of the fiber penetrated the strainer early in each test when the debris load was low. Vogtle calculated that 13 g/FA of fiber would arrive at the core inlet for the CLB under the most limiting design basis conditions for pump and strainer operating combinations. As a comparison, for the Vogtle HLB case, 91 g/FA of fiber was calculated to reach the core. Note that STP did not calculate a HLB fiber amount, but conservatively assumed that the core was completely blocked relatively early in the event.
Plants have different design configurations that will affect the amount of fiber that may reach the core following a CLB. However, the STP study and the Vogtle results demonstrate that small breaks do not generate enough debris to inhibit LTCC following a CLB. These examples provide evidence that even the largest CLBs in high-fiber plants are unlikely to result in greater than [xxxxxxx] fibrous debris at the core inlet. Also, because debris deposition at the core inlet will be non-uniform, it is very unlikely that debris at the core inlet will inhibit LTCC following a CLB. In conclusion, even the largest DEGB is unlikely to generate enough debris to approach the [xxxxxxx] CLB limit presented in WCAP-17788 and a large CLB is predicted to occur very infrequently (see Section 5.5.2).
5.2.1.1.1 Debris Collection at the Core Inlet The fluid velocity profile across the core inlet varies depending on the relative power of the fuel.
The flow patterns at the core inlet will deposit debris close to the center of the core but not at the periphery. Boiling and circulating flows will prevent a uniform debris bed from forming across the entire core inlet for the debris amounts determined by testing. Therefore, mixing between the core and LP are unlikely to be inhibited during the LOCA recovery (see Section 5.2.1). Even with an increase in density in the core region, which was not modeled by industry analyses, the recirculating currents will continue as decay heat decreases and boron concentrates in the core region. This conclusion is supported by TRACE simulations (21) that included the effects of boric acid concentration in the core. Section 5.4 discusses TRACE calculations.
The TH analysis in WCAP-17788 becomes unreliable once decay heat decreases to the point where boron concentration in the core increases because of limitations of the code in accommodating buoyancy-driven flow. The TH analysis shows that the steam quality exiting the
RV remains below 0.8 before the switchover to sump recirculation and that this ensures that boron solutes will not concentrate in the core up to that point.
The driving force for water entering the core is the manometric balance between the DC region of the RV and the core region. The head from the DC must overcome liquid and vapor flow losses on the core side as well as any headloss caused by debris and density difference resulting from concentration of boric acid or other debris in the core region.
The SE for WCAP-16793-NP-A found that at less than 15 g/FA, flow to the core would continue for the HLB condition, even if chemical precipitates were present. For the CLB, the staff found that LTCC would be assured for plants that meet the HLB criterion of 15 g/FA, but this finding did not address the effects of BAP. The staff finding also assumed that the maximum fibrous debris load arriving at the core inlet for the CLB would be roughly half of the HLB amount.
For this TER, it is assumed that chemical effects, discussed in Section 5.2.1.5, are delayed until after the licensee implements HLSO (or equivalent). WCAP-16793-NP-A testing found that under CLB flows, the deposition of fibrous and particulate debris at the FA inlet resulted in very low headloss. These headlosses were found to be much lower than those occurring when chemical precipitates were included. The CLB testing for WCAP-16793-NP-A (see WCAP-17057-P, Revision 1, GSI-191 Fuel Assembly Test Report for PWROG, issued September 2011 (22)) included several tests with a fiber loading of [xxxxxxx] and varying amounts of particulate. The tests included chemical precipitates. During the testing, headlosses were allowed to stabilize and were recorded before chemical debris was added to the test. The highest non-chemical headloss recorded during the testing was [xxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. The PWROG stated that the available driving head for all plants is greater than this value. [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. Information provided by the PWROG indicates that the driving head calculation was based on a reactor design that has 14-foot FAs while all other designs have 12-foot FAs. The DC for the example plant was stated to be similar to the designs that have 12-foot FAs. This indicates that the calculation has about 2 feet of margin for most plants (approximately 1 psi).
The PWROG also performed small scale boiling rod bundle tests that included fibrous and particulate debris. These tests were performed and documented in WCAP-17360 (5). The testing used a vertical 3x3 rod bundle that had dimensions and materials like those of PWR fuel.
The rods were heated to achieve boiling conditions similar to those expected after a large CLB.
During the rod bundle tests no debris bed formed at the fuel inlet. The debris was observed to pass through the BN and settle on top of the nozzle or pass into the heated core region.
WCAP-17360 determined that the boiling conditions downstream of the FA inlet produce instabilities that inhibit debris bed formation. A small amount of fiber was observed to collect at the outer edge of the spacer grids (between the grids and the test column). The rod bundle testing was not conducted with a prototypical FA or LP, so it cannot be definitively relied on as applicable to a full core. However, it does provide evidence that turbulence from boiling will prevent significant amounts of debris from collecting at the spacer grids.
The rod bundle tests demonstrated that density-driven flow occurs in the boiling condition through observation of temperature stratification in pure water tests and mixing of fluid, as determined by relative temperature between the simulated core and LP, for tests that included boric acid. Additional evidence of mixing was obtained by sampling the water in the simulated
LP. Maintaining the LP as a mixing volume reduces the concentration of boric acid and other debris in the core making BAP less likely and delaying its onset.
Previous studies performed by the PWROG for WCAP-16793 and RES showed that for the CLB, a single open FA provides adequate coolant to maintain LTCC.
The following events are each unlikely as discussed above:
(1) Large CLB LOCA (2) Transport of fibrous debris more than the analyzed limit (based on a uniform debris bed)
(3) Formation of a uniform debris bed Considering the above, it is very unlikely that a high resistance debris bed will form at the core inlet such that flow into the core will be prevented. Therefore, the potential for a loss of LTCC because of debris in the core is of low safety significance for the CLB. The initiation of HL injection or an equivalent action is assumed before chemical effects may occur. HL injection bypasses the debris bed at the core inlet before headlosses resulting from chemical effects exceed the available driving head and flushes concentrated debris and boric acid from the core.
5.2.1.1.2 Boric Acid Precipitation The PWROG performed SKBOR studies to estimate the effects of debris buildup at the core inlet on BAP. The simulations reflected conditions typical of design-basis safety analyses. A case was performed as a baseline assuming no core blockage and 50 percent mixing with the LP. Sensitivity cases assumed 100 percent mixing availability with the LP but assumed percentages of core inlet blockage varying from 0 to 75 percent. The study found that the cases with 0, 25, and 50 percent blockages took longer to reach the boric acid solubility limit than the baseline. The 75 percent blockage case reached the BAP limit faster than the baseline.
Therefore, it was concluded that greater blockage leads to delays in the exchange flow between the core and LP.
The study demonstrated that up to 50 percent of the core inlet can be blocked with debris before the flow exchange between the core and LP will be inhibited such that the licensing basis assumption of crediting 50 percent of the LP volume becomes questionable. The SKBOR studies are conservative because of the application of conditions used in design basis safety analyses. Based on the SKBOR studies, WCAP-17788 states that a relatively low steam quality (less than 80 percent) allows a significant amount of the boron to be removed from the core via entrainment and that this will delay the onset of BAP compared to most of the SKBOR cases in the study. The PWROG performed an SKBOR study that demonstrated a steam quality of 0.95 would result in significant boron concentration in the core, but that the concentration would never exceed the precipitation limit. Higher steam qualities resulted in eventual precipitation.
These studies assumed that 50 percent of the LP was available for dilution of the boron.
The findings from the SKBOR studies are consistent with the trends found in TRACE studies of BAP completed by RES (21). The TRACE studies are discussed in Section 5.4. TRACE found that BAP would be delayed significantly longer than in the SKBOR studies conducted by the PWROG.
The PWROG conducted a brine test program described in PWROG-15091-P, Subscale Brine Test Program Report, dated November 30, 2015 (6). The testing was conducted to learn more about density-driven mass transport between the core and the LP with and without debris present at the FA inlet. The test program demonstrated the fluid exchange and showed that highly resistive debris beds cannot form at the core inlet at the established CLB in-vessel fiber debris limit. The program did not consider chemical effects because it was assumed that HL recirculation would be initiated before chemical precipitation. Testing included conditions expected to occur following a CLB and considered both Westinghouse and AREVA core inlet geometries.
Fibrous debris ranges for the testing represented 2.5 to 22.5 g/FA. Some of the tests included particulate, but it was ultimately determined that the amount of particulate included did not significantly affect the test results or differential pressure across the debris bed for the tested conditions. The results from the brine testing were consistent with the subscale headloss testing that found that [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxx].
The brine testing demonstrated that the exchange flow rate (between the simulated core and LP volumes) was influenced by the amount of fiber included in the test. Larger fiber amounts blocked more open areas in the BN. Under the test conditions, the relative timing of the debris injection, the flow velocity into the FA, and the increase in density of fluid in the core region affected the ability of a debris bed to form. If a debris bed formed, the exchange between the modeled core and LP regions stopped but was reestablished when the density of the fluid in the core increased or when the flow from the LP to the core decreased and the debris bed was disrupted. The brine testing provides evidence that exchange flow can take place, even if a relatively low resistance debris bed forms at the core inlet.
The NRC staff noted that the brine test results and the results of FA testing for WCAP-16793-NP-A had different headloss values for similar fibrous debris loads. The PWROG stated that the results from the brine testing and other FA tests had different headlosses because the test methods were different. It was noted that higher in-bed particulate loads can result in higher headlosses and that the tests being compared might not have similar particulate loads because the WCAP-16793-NP-A tests used a recirculating flowpath while the brine testing used a once-through system. Some of the particulate in the brine testing might have passed through the fuel inlet and avoided capture in the debris bed. The PWROG also noted that both test programs support the assertion that a [xxxxxxx] limit will not result in a significant headloss at the core inlet.
The staff concluded that the headlosses that may occur at debris loads below the fiber limit are relatively low before the arrival of chemical precipitates and that both test programs demonstrated this.
Based on the SKBOR studies and the brine testing discussed in this section, and the TRACE results presented in Section 5.4, the NRC staff determined that debris effects do not invalidate the current licensing bases or HLSO times for operating PWRs. The analyses indicate that BAP may be of lower risk than was previously thought. However, changes to a plants licensing basis that draw on these insights would have to be justified on a plant-specific basis. In addition, maintaining HLSO (or equivalent) is important to ensure continued LTCC by flushing concentrated debris from the core and to decrease the probability that chemical precipitates cause a significant increase in headloss.
5.2.1.1.3 Debris Suspended in the Core The effects of debris in the core for the HLB are expected to be bounded by those that will occur during the CLB. This section includes some discussion of HLB behavior, but the behavior following a HLB should be similar to that following a CLB, except for the amount of debris present.
Debris that passes through the core inlet or enters through an AFP will reach the core and UP.
Debris in the core can affect fluid and heat transfer properties of the fluid. Because the CLB initially has the core in a boiling pot mode, any debris that reaches the heated core is assumed to stay there. For the HLB, much of the debris is flushed from the core making the CLB limiting.
For the CLB, debris can concentrate in the core, similar to the way boric acid can. Most fibrous debris will reach the core in the first few hours of the event, and most particulate debris will arrive much later because most transportable particulate debris is failed unqualified coatings.
The amount of fibrous debris reaching the core is much less than the amount of particulate debris that may reach the core. For the HLB, the amount of fiber that reaches the core may be greater than for the CLB. However, before significant core blockage, some of the fibrous debris will be flushed from the core. The amount of fibrous debris that may reach the core is an order of magnitude less than the potential particulate debris.
Electric Power Research Institute (EPRI) testing, described in EPRI 1011753, Design Basis Accident Testing of Pressurized Water Reactor Unqualified Original Equipment Manufacturer Coatings, dated September 30, 2005 (23), demonstrated that most unqualified coatings will fail after 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Early in the event, boiling maintains turbulence in the core and keeps debris well mixed. After a few hours, depending on plant-specific procedures, plants initiate actions, like HLSO (or equivalent), to flush concentrated boric acid from the core. This action will also flush debris from the core. Because most of the debris that may enter the core following a CLB cannot arrive until after HLSO (or equivalent), the amount of debris in the core is relatively small, minimizing the effects of debris on fluid and heat transfer properties. Section 5.2.1.1.5 discusses in more detail coatings failures, particulate debris, coatings failure timing, and debris characteristics.
Rod bundle testing demonstrated that debris in the heated region remained well mixed with the fluid. [xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx]. Some deposition of debris at the outer edge of the spacer grids was observed. Similar behavior is expected for the HLB. The TH analyses discussed above demonstrate that circulating currents are present following a CLB. The currents assist with keeping debris and boric acid well mixed in the core.
Debris suspended in the core may have an influence on fluid properties and displace liquid in the core. WCAP-17788 assumed a 5 percent mass limit on the debris in the core. The increase in density from this amount of debris was determined to have a small effect on headloss. The effects on fluid viscosity and surface tension were also determined to be small. The rod bundle testing was referenced to show that bubbly flow observed in the testing was not significantly affected by the addition of debris.
Debris suspended in the core may also affect fluid heat transfer properties. Rod bundle testing validated that no major degradation of heat transfer occurred. Open literature, provided as
references in WCAP-17788, indicates that limited amounts of particulate debris suspended in fluid will improve heat transfer.
Debris in the core could affect BAP by displacing liquid and reducing the liquid volume available for mixing. Based on the small amount of liquid displaced, any changes to the rate of boron buildup in the core will be negligible. The 3x3 rod bundle testing also provided evidence that precipitation modes, and boric acid transport and mixing phenomena, are not negatively affected by debris in the core.
5.2.1.1.4 Local Collection of Debris The rod bundle testing demonstrated that significant collection of fiber at the spacer grids is unlikely. However, should debris become trapped in the spacer grids, the effects would be limited to local areas. WCAP-16793-NP-A, Revision 2 (1) evaluated complete blockage around a fuel rod at a spacer grid and demonstrated that axial conduction would allow the fuel cladding to remain at acceptably low temperatures. In addition, the 3x3 rod bundle testing did not identify any significant collection of debris at the spacer grids, even with a relatively high fibrous debris load equivalent to 150 g/FA. The open lattice fuel design for PWRs permits crossflow between assemblies, which will allow flow around localized blockages. This section applies to HLBs and CLBs.
5.2.1.1.5 Concentration of Particulate Debris The NRC staff evaluated the different categories of coatings and other particulate debris sources to assess the potential effects of particulate debris. The three coating categories evaluated are qualified coatings, degraded qualified coatings, and unqualified coatings. The evaluation of coatings depends on failure timing to demonstrate that an excessive amount of particulate debris will not concentrate in the core following a CLB. For a HLB, the coatings and other particulate debris are not expected to concentrate because there is significant excess flow through the core that will flush debris as it exits the reactor.
Similar to boric acid, debris may concentrate in the core during a CLB because water boils away and particulate debris remains behind. Because there is not significant empirical data for particulate filtration by sump strainers, it is assumed all particulate debris passing through the strainer arrives at the core. In addition, filtering at the core inlet is not assumed.
All qualified, unqualified, and degraded coatings within the ZOI are assumed to fail immediately as fine particulate that is transported to the core. Based on the review of individual plant submittals in response to GL 2004-02, the approximate range for ZOI coatings amounts is 100 to 2,000 pounds. Qualified coatings outside the ZOI are assumed to remain intact and attached to the substrate. They do not contribute to the debris source term at any time during the LOCA recovery.
Unqualified coatings are assumed to fail as particulate debris. This category of coating is generally the largest contributor to the fine particulate debris source term. Some plants consider degraded qualified coatings to be unqualified. Based on the review of individual plant submittals in response to GL 2004-02, the amount of unqualified coating debris ranges from less than 100 pounds to 20,000 pounds. Based on the EPRI testing discussed below, it is estimated that less than 10 percent of the unqualified coatings will fail before 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
Therefore, it is assumed that 10 percent of the unqualified coatings are available for transport to
the core before the initiation of HLSO (or equivalent), which occurs within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> after the initiating LOCA. Once HLSO (or equivalent) is initiated, the core will be flushed of concentrated boric acid and particulate debris.
Degraded qualified epoxy coatings outside the ZOI are assumed to fail as chips. The chips are assumed to settle or be captured by the strainer and not be available to transport to the core.
Degraded qualified inorganic zinc outside the ZOI is assumed to fail in particulate form. This particulate is assumed to be available to transport to the core, but the degradation of this coating category into particulate is estimated to take longer than 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> so it would not be transported to the core before HLSO. Many plants did not differentiate between degraded qualified coatings and unqualified coatings in their analyses. For these plants, the unqualified coating term is conservative with respect to the amount of particulate debris that may pass downstream of the strainer. The basis for the assumed debris characteristics of the degraded qualified coatings is testing performed by Keeler & Long. (24)
Unqualified coatings make the most significant contribution to amounts of particulate debris in the core. To evaluate the potential effects of particulate debris in the heated core, the NRC staff referred to an EPRI report that discusses the behavior of unqualified coatings when exposed to post-LOCA conditions (23).
In addition to coatings, particulate debris can originate from latent debris and particulate type insulations. According to NRC guidance, the assumption for the mass of latent particulate is generally 170 pounds (8). Guidance also recommends plants to confirm their latent debris amounts. Plants identified a range of latent particulate, with the largest amount assumed to be 500 pounds.
Most plants do not have a significant particulate insulation source term. However, one licensee determined that 2,200 pounds of calcium silicate insulation could be generated by a limiting break. This plant has the largest particulate source term when calculated, as discussed below.
To estimate the amount of particulate debris that might transport and collect in the core during a CLB, the NRC staff used information provided by licensees in their responses to GL 2004-02.
The staff summed 100 percent of the ZOI generated coatings, 100 percent of the latent particulate, 100 percent of the particulate insulation debris calculated to reach the strainer, and 10 percent of the total unqualified coatings. A few plants were estimated to have maximum values of particulate that could reach the core between 2,500 and 3,000 pounds. These values are considered realistic for the worst breaks, and therefore conservative for other breaks. The largest calculated particulate amount calculated by the staff was 2,910 pounds.
The staff assumed that 2,910 pounds of particulate would be mixed with the coolant in the core and LP regions. Mixing between the regions is discussed in the CLB discussion in Section 5.2.1.1. The amount of liquid in the core and LP is assumed to be 60,480 pounds based on 50 percent voiding in the core and 100 percent of the LP volume available for the Westinghouse 3-loop reactor design. If the entire mass of particulate is mixed in the LP and core region, the mass ratio is 4.8 percent particulate to liquid. A 3-loop Westinghouse design was used to determine the fluid mass because it has smaller volumes for the LP and active core regions than the 4-loop Westinghouse reactor design. The 3-loop and 4-loop Westinghouse designs are considered as representative of the PWR fleet. The 2-loop Westinghouse designs have smaller RCS volumes but use injection into the UP, which reduces the concern with concentration of debris in the core. The staff verified that the maximum particulate mass for the
2-loop plants is less than 1,000 pounds so the mass ratio for these plants is limited to about 2 percent considering the smaller volumes for these plants.
For design-basis BAP calculations, 50 percent of the UP volume, up to the bottom of the HL, is typically credited. This volume was ignored for the particulate debris calculation and provides margin in the calculation.
The debris amounts in the example above are realistic maximum debris generation values for the limiting plant in the fleet. The amounts of debris generated were calculated for the worst-case breaks using accepted methods that are considered conservative. Other breaks, especially smaller, more likely breaks, would result in less debris. Also, a low liquid mass was used in the calculation by choosing a plant design with low volume core and LP regions (except for UPI plants discussed above) and assuming a high void fraction for the core region.
Therefore, the calculated mass ratio is considered a conservatively high value for the fleet.
Almost all breaks would generate significantly less particulate debris. In addition, the amount of particulate assumed to reach the core is based on 100 percent transport. It is likely that some would not reach the strainer or would settle in the LP, be filtered by the strainer, or be filtered at the core inlet. The latent particulate is a relatively small portion of the source term for the plants with a large amount of particulate, but a significant amount of this source would settle because of its larger size.
WCAP-17788 evaluated a limit of 5 percent by mass particulate to liquid in the core. The final evaluation of this issue was included in the Westinghouse response to Request for Additional Information (RAI) 1.3 by letter dated September 28, 2018 (25). The NRC staff found the evaluation to reasonably address the potential issues that could result from particulate suspended in the core.
Once HLSO (or equivalent) is initiated, any debris that has collected in the core will be diluted by mixing the coolant in the reactor back into the sump. Therefore, it is important that plants perform HLSO within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> to prevent excessive unqualified coatings from concentrating in the core.
5.2.1.1.6 Conclusions for the Cold-Leg Break Based on the discussion above, the staff concludes that the [xxxxxxx] fiber limit for the CLB is acceptable and that flow to the core will be maintained with this debris load if chemical precipitates do not deposit or form on the debris bed. The probability of any PWR accumulating
[xxxxxxx] at the core inlet is very low because of the low volume of flow, and therefore the low amount of debris that approaches the core inlet during a CLB. If more than [xxxxxxx] is transported to the core inlet, the formation of a uniform bed is unlikely, and flow to the core will likely continue. The actual debris amount that might result in preventing adequate flow to the core is not defined because testing is based on a uniform debris bed. If the core inlet becomes completely blocked, AFPs will allow flow to reach the core. There is adequate DID to assure LTCC following a CLB for the current operating fleet of PWRs.
Current plant licensing bases for BAP do not need to be revisited because of the effects of debris on LTCC. Analyses performed in WCAP-17788 and TRACE (21) show that boric acid will not concentrate to the point of precipitation in areas important to support cooling.
The collection of debris in the heated core will not significantly affect LTCC. The amount of debris that enters the core following a CLB is limited by the amount of debris available before the initiation of HL injection (or equivalent) and the low amount of flow entering the core. A large amount of particulate debris may be available, but this debris will not become transportable until after HLSO is initiated. This prevents the concentration of significant amounts of debris in the core. Filtering by the strainer also reduces the amount of debris that may enter the core.
Local collection of debris (at spacer grids) will not adversely affect cooling of the fuel as demonstrated by calculations for WCAP-16793-NP-A.
Flow from HL injection (or equivalent) is expected to be initiated prior to BAP. Complete blockage of the core inlet is not expected for plants that initiate HL injection (or equivalent) before the formation of chemical precipitates.
5.2.1.2 Hot-Leg Break Following a HLB LOCA, for all PWR designs except UPI, the ECCS is injected into the CL and flows down the DC, into the LP, and up through the core. All the injected flow passes through the core. All debris that bypasses the strainer and is pumped by the ECCS will arrive at the core inlet and may collect there causing an increased resistance to flow through this path. The driving head available for this scenario is the ECCS pump discharge as limited by the height of the SG tubes. If the core inlet resistance increases significantly because of debris, AFPs will become activated to allow coolant to reach the core and remove decay heat, helping to maintain LTCC. Flow to the CSS does not enter the core and debris in that stream returns to the sump.
After a few hours, PWRs change the injection path to prevent the buildup of boric acid in the core following a CLB. This action is taken following all large LOCAs because the operators may not be able to determine if the break is on the hot or cold side of the RCS. The injection is switched to the HL or a combination of the HL and CL. For a HLB, this injects coolant into the UP where it can drain into the core to provide cooling. Excess flow spills out of the break in the HL and returns to the sump.
To determine the approximate size of a break that could result in significant debris at the core inlet, the NRC staff used information from several sources. First, a debris limit of [xxxxxxxxx]
was assumed. This is taken from the WCAP-17788 analysis for the most conservative fuel and plant design combination and applies only to the HLB scenario. Debris limits for the HLB were calculated to be between [xxxxxxxxxxxxxxx] depending on the fuel and reactor design combination. Some combinations could have accommodated more, but [xxxxxxx] represents the maximum debris amount tested for WCAP-17788. [xxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx xxxxx].
Table 2 of Section 5.3.3.6 lists the HLB fibrous debris limits for the various RCS design and fuel type combinations. Section 5.2.1.1 discusses CLBs.
Although the staff did not accept the WCAP-17788 analysis that determined the fiber limits, the staff considered that the limits were established at limiting flow, particulate size, and particulate to fiber ratios. The staff also compared the fibrous debris limits from WCAP-17788 to test
results from WCAP-16793. The fiber limits from WCAP-17788 testing are lower than those in tests that WCAP-16793 indicate as acceptable. Even with significantly higher fiber amounts included, test results from WCAP-16793 indicate that the WCAP-17788 limits are conservative.
The NRC staff assumed that information from the STP submittal could be used to estimate the HLB size that would generate and transport enough fibrous debris to reach the acceptance limit.
Plants that have less fibrous insulation installed in the containment are of lesser concern and are bounded by plants with large amounts. The choice of STP as the example plant is reasonable because of the large amount of fibrous insulation, the large number of breaks evaluated around the break size of interest, and the use of the largest debris amount from the subset of breaks available at the given size. STP provided debris generation and transport information for all welds that could fail and result in a LOCA as part of a risk-informed license amendment request to resolve GSI-191 (13). The staff considers STP representative of the most challenging plants with respect to the generation and transport of fibrous debris to the strainer.
To determine the maximum amount of fiber that can arrive at the strainer and still meet the most conservative fiber limit, the staff considered a plant with the minimum number of FAs (157) and an 85 percent filter efficiency at the strainers. For these conditions, the maximum fiber allowed at the core inlet is 3,202.8 grams or 7.1 pounds (20.4 grams of fiber multiplied by 157 FAs.)
Using an 85 percent filter efficiency for the strainer, this implies that 47 pounds of fiber can arrive at the strainer. (UPI plants have fewer FAs but are evaluated in a different way. See Section 5.2.1.3.)
The smallest break at STP that results in 47 pounds of fibrous debris at the strainer is a 6-inch DEGB (5.19-inch inner diameter). The STP debris loading includes 28.5 pounds of latent fiber.
Therefore, for this case, 16.5 pounds of break generated fiber are transported to the strainer.
Some 6-inch breaks generate more fibrous debris, and some generate less. The range of fibrous debris loads at the strainer for all 6-inch breaks is 29 to 63 pounds.
The next smaller break in the STP analysis is on a 4-inch pipe. The limiting 4-inch break transported about 38 pounds of fiber to the strainer. All other 4-inch breaks resulted in less fiber at the strainers.
A fiber pass through of 15 percent (85 percent filter efficiency) is realistic and biased conservative based on the data that licensees have provided to the NRC. Section 5.2.1.2.1 discusses strainer penetration in further detail. The value is based on testing that used 100 percent fine fiber inserted upstream of the strainer and used realistic or conservative flow rates and other test methods. The data provided strainer filter efficiency for a range of debris loads. The NRC staff does not have a large database of penetration testing that is useful for calculating penetration values at relatively low fiber amounts. Most early penetration testing was performed to determine the maximum amount of fiber that can penetrate the strainer considering a plants maximum debris load. There is some uncertainty associated with the 15 percent penetration value because of the many different strainer configurations (size, number of strainers, flow velocity, and other factors) throughout the PWR fleet. The evaluation of the smallest break size that could generate enough debris to cause in-vessel issues is centered on a relatively small fiber amount. As the amount of fiber at the strainer increases, filtration efficiency increases.
For this evaluation, the pipe break that could result in a problematic debris amount for the most limiting HLB conditions is assumed to be a DEGB of a 6-inch pipe on the hot side of the RCS.
Such breaks are predicted to occur infrequently (see Section 5.5.2). Smaller, more likely breaks, will be less challenging to the fuel than analyses conducted by the PWROG and NRC indicate. Smaller breaks result in lower ECCS injection flow rates.
Also, as discussed above CSS may not be initiated for some plants for small breaks. The lower injection flows delay the switchover to recirculation. The WCAP-17788 analyses assumed an early switchover based on the largest LOCAs. Early switchover results in a higher decay heat assumption that is more challenging to fuel temperatures. Delaying the switchover to recirculation would result in a decrease in decay heat and require less flow to maintain LTCC.
This provides margin in the analyses.
Subscale FA testing performed for WCAP-17788 found that headloss across debris at the core inlet will vary depending on the size and amount of particulate debris arriving at the core inlet.
The fibrous debris limit was established using a conservative size distribution and amount for particulate debris. Small amounts of particulate result in an uncompact bed with low headloss.
Excessive amounts of particulate result in an unstable bed and bed breakthrough at the fiber amounts tested. Flow also affected debris bed formation, with higher flow rates resulting in bed breakthrough. The fiber limits determined by testing were based on the most limiting conditions for the parameters studied. Alternate conditions will result in higher fiber loads required to reach the same headloss.
The analysis discussed in this section is considered realistic or conservative for the following reasons:
- It used debris loads representative of a very high fiber plant
- It used a realistic to conservative strainer filtering value
- It used the most limiting fiber limit for 2 FA types and 4 plant design categories
- It used the smallest number of FAs for applicable plant designs
- Other margins discussed in this section and throughout the document apply 5.2.1.2.1 Fiber Penetration There is a significant database that can be used to determine plant-specific total fiber penetration amounts or total amounts per specific strainer area. Early testing did not gather periodic data to determine the penetration behavior for fiber, but simply collected a total amount at the end of the test. More recent testing collected fiber downstream of the strainer at frequent intervals. Information from this testing has been used to determine how fiber penetration changes over time or with strainer debris load.
Penetration data indicates that the percentage of debris that passes downstream of the strainer at any time is related to the debris load on the strainer. Penetration percentage values taken only at the end of the test are not meaningful except to the specific condition tested or to determine a maximum fiber penetration value for a specific plant. The NRC staff has a limited database of strainer penetration testing that can be used to determine time or load dependent
strainer penetration amounts. Figure 1 below (from the Vogtle test program) illustrates that fiber penetration decreases as debris amount on the sump strainer increases (26).
Figure 1 - Fiber Penetration as a Function of Debris Load The staff used information from six recent strainer penetration test programs that have more refined information based on frequent downstream debris collection. Each program performed several tests to gain understanding into the factors that affect penetration. The staff determined that fiber penetration amount decreases quickly with fiber buildup on the strainer as shown in the figure above. Relatively smaller breaks may be almost as challenging for downstream effects as larger breaks because the debris initially arriving at the strainer increases the filtration efficiency.
The evaluation for the 6-inch HLB in this TER considered the results from these six penetration test programs because they were the only relevant information available to the staff. The staff used a value for penetration that was conservative with respect to the available data for the debris load considered for the case. Some plants may have plant-specific strainer configurations that would allow greater penetration at the specific debris load considered.
However, two of the units with recent testing have very large strainer areas (one is the largest strainer installed in any U.S. PWR) making it likely that the data would be conservative with respect to the fleet. Other considerations like approach velocity and strainer perforation size
can influence penetration values. The amount of fiber that reaches the core also depends on plant configuration, including the number of strainers, and how the operating ECCS and CSS pumps are aligned to the strainers. For example, any debris that is returned to the sump via the CSS or flows out of the break, returns to the sump and may be removed by filtering on the strainer. If CSS does not initiate or a plant has lower ECCS flow, more debris is delivered to the core before the strainer efficiency increases.
The amounts of debris transported to the core for the CLB are less than for the HLB because most of the ECCS flow spills out the break in the CL. WCAP-17788 assumed that less than
[xxxxxxxx] of fibrous debris would arrive in the heated core and at the core inlet combined for the HLB. The Vogtle study discussed above is for a very high fiber plant. This study indicates that 90 g/FA could reach the core under limiting conditions for Vogtle. Sections 5.2.1.1.3, 5.2.1.1.4, and 5.2.1.1.5 in the CLB section discuss debris suspended in the core.
Considering available data and variations in plant strainer parameters, the penetration value used by the staff is conservative for most or all plants in the U.S. PWR fleet.
5.2.1.2.2 Boric Acid Precipitation BAP has historically been evaluated with respect to CLBs (Section 5.2.1.1.2 above discusses the potential for BAP to occur following a CLB). However, if debris collects at the core inlet following a HLB, to the extent that flow into the core is significantly reduced, conditions similar to those in a CLB could occur. The staff considered the conditions following a HLB and determined that the potential for BAP was bounded by the CLB conditions. For a HLB, the coolant delivered to the core before a significant amount of debris arrives is much greater than that needed to replace coolant lost to boiling. This excess flow flushes boric acid and other debris from the core more than during a CLB. Flushing of the core will continue unless a significantly resistant debris bed forms at the core inlet. Even if a resistant debris bed forms and begins to inhibit flow through the core inlet, coolant is redirected to the core through AFPs. The flow delivered to the core exceeds that required to replace boil off and the fluid in the core is well mixed by convective currents and boiling. Therefore, boric acid is removed from the core such that the concentration required for precipitation will not be approached.
5.2.1.2.3 Conclusions for the Hot-Leg Break The combination of the low frequency of a break that could generate and transport enough debris to challenge LTCC, and the conclusion that a debris bed that could block core inlet flow is unlikely to form implies that it is very unlikely that debris would adversely affect LTCC.
Based on the discussion above, the staff concludes that the fiber limits from WCAP-17788 (also listed in Table 2 in Section 5.3.3.6) are acceptable metrics to assist in evaluating the acceptability of HLB debris amounts. Flow to the core will be maintained with these debris loads if chemical precipitates do not deposit or form on the debris bed. If a limited amount of fiber greater than the design-specific limit is transported to the core inlet, a uniform bed is unlikely to form and flow to the core will continue. The actual debris amount that might result in preventing adequate flow to the core is not defined. If the core inlet becomes completely blocked, AFPs will allow flow to
reach the core as described in Section 5.3.3. There is adequate DID to assure enough flow to maintain LTCC following a HLB for the current operating fleet of PWRs that demonstrate acceptable fiber amounts at the core inlet.
Considering the above, the NRC staff concludes that it is very unlikely that debris in the RV, generated by a HLB, would inhibit LTCC for the current operating PWR fleet. In addition, the BAP evaluation for the CLB is limiting with respect to the HLB, so BAP is not a concern for the HLB.
5.2.1.3 Upper Plenum Injection Plants The UPI plants have an ECCS configuration unlike other PWR designs. The operating fleet has five UPI units (R.E. Ginna Nuclear Power Plant; Prairie Island Nuclear Generating Plant, Units 1 and 2; and Point Beach Nuclear Plant, Units 1 and 2). The ECCS initially injects directly to the UP, above the core via an independent nozzle. All fluid and entrained debris are available to enter the top of the core. The initial injection into the UP is similar to HL injection for a typical PWR. Therefore, a HLB for the UPI design is like the CLB for other plants and most of the flow goes out the break in the HL. The core inlet for ECCS injection and initial recirculation is at the top of the core instead of at the bottom of the core.
The top of the core is not likely to become blocked with debris because of the turbulence caused by boiling and steam exiting the core. The steam exiting the core will impact structures above the core exit and add to the turbulence. For a HLB, most of the coolant spills out the break. For a CLB, the coolant flows through the core and out the break. Therefore, debris enters the core region with the flow. The effects of turbulence on the formation of debris beds were discussed above for typical PWR designs. Turbulence is much greater for the UPI design because coolant enters the reactor at the top of the core where the velocity of steam exiting the core is relatively high. Because the top of the core is unlikely to become blocked, cooling is maintained during injection into the UP.
UPI plants use something similar to HLSO to prevent BAP following a HLB. Instead of switching injection from the CL to the HL, UPI plants switch from UPI to CL injection, or a combination of UPI and CL injection. The switch to CL injection occurs several hours after the initiating LOCA.
If the break is in the hot side, the switch to CL injection will flush concentrated debris and boric acid from the core. If the break is in the cold side, coolant injected into the CL will flow into the core via the DC and LP and only the coolant required to make up for boil off enters the core. By the time a UPI plant switches to CL injection, debris in the sump will be depleted so it is unlikely that a debris bed will form at the core inlet. Because of the unique design of UPI plants, it is not critical that the switch to CL injection be initiated before chemical precipitates may form. This is because a filtering bed is unlikely to form at the top of the core where injection occurs.
However, it is important that CL injection is initiated to flush concentrated debris and boron from the core.
Once injection is switched to the CL, debris remaining in the sump may be pumped to the core inlet. The injection into the CL can provide pressure up to the shut off head of the ECCS pumps (as limited by the height of the SG tubes) if the break is in the HL. For this condition, WCAP-16793-NP-A and the associated staff SE accepted a fibrous debris load of up to 15 g/FA at the core inlet, even when chemical effects are included in the debris headloss.
The switchover to CL recirculation takes place several hours after the LOCA; therefore, it is possible for chemical precipitates to form and filter on any debris bed. By this time, filtering on the strainer, settling, or deposition in the core will have removed most of the fibrous debris from the sump. A small amount of fiber may be suspended in the pool and transported to the strainer. Most of this will be captured on the debris bed on the strainer, but a small amount may penetrate and transport to the fuel inlet nozzles. As discussed in Section 5.2.1, a continuous bed is not expected to form at the core inlet, especially with the limited amount of fibrous debris arriving at the bottom of the core for this scenario. The debris amounts likely to arrive at the core inlet this late in the scenario are much less than 15 g/FA because the source term in the pool has been reduced to almost zero by the time CL injection is initiated.
After switchover to CL injection, for a CLB, coolant is delivered to the core only by the head resulting from the manometric balance between the core and the DC. This is similar to the CLB with CL injection in other PWR designs, but it occurs for UPI designs only after the switchover to CL recirculation. For UPI plant CLBs, after switchover to CL injection, only the coolant required to make up for boil off enters the core. The remainder of the coolant with any entrained debris flows out the break. Because the switch to CL injection is relatively late in the LOCA scenario, chemical precipitation may occur. For this condition, WCAP-16793-NP-A did not conclude that a debris limit of 15 g/FA was justified. After chemical effects are included in the debris headloss, testing shows that the headlosses for the CLB flow rates can increase by 1 psi above non-chemical headlosses for tests that included fiber of [xxxxxxx]. This headloss could encroach on the available driving head for CL recirculation for the UPI plants. However, because the break is in the CL and most of the debris will have been deposited on the strainer before the initiation of CL injection, the amount of debris reaching the core inlet is much less than for the HLB.
The largest source of fiber may be that from erosion of larger pieces that have settled in the pool. Most of this erosion occurs relatively early in the event and decreases with time. Erosion is a relatively small source of fiber and the eroded fines are likely filtered by the sump strainer/debris bed. Most of the fluid injected into the CL flows out of the break and back to the sump where it is filtered at the strainer before being recirculated back to the core. Because the amount of fiber in the pool has been depleted by the time CL recirculation is initiated and the majority of any debris that penetrates the strainer flows out the break, the quantity of fiber reaching the core inlet for a CLB is much lower than for the HLB considered in WCAP-16793-NP-A. The amount of fiber reaching the core inlet is unlikely to have a measurable effect on headloss or flow at the core inlet.
5.2.1.3.1 Conclusions for Upper Plenum Injection Plants Because of the unique design of the ECCS for UPI plants, a debris bed will not inhibit flow into the core. The initial injection is into the UP where turbulence will prevent a bed from forming.
By the time a UPI plant switches to CL injection, the amount of fiber that may be transported from the pool to the strainer is very small and the filtration efficiency of the strainer is high
because of the debris bed deposited on it. Therefore, the amount of fiber reaching the core inlet (FA BNs) will not affect LTCC.
5.2.1.4 Debris Generation and Transport This section discusses margin related to debris generation and transport. It provides details, especially for the aspects of the analyses that are more complex. The margins are summarized in Section 5.5.3.
The most likely breaks are smaller breaks that are less challenging to mitigate from several perspectives. Reasons for this are discussed below with respect to specific parameters that are considered in the evaluation. In addition to those, smaller breaks result in lower required ECCS makeup flow. The injected sources are available for a longer period and decay heat is lower when the system is switched to recirculation. Lower decay heat allows for lower flows to make up for coolant lost to boil off and greater time for operators to respond to debris blockage.
Debris generation values used by licensees in their analyses are for the most challenging breaks for strainer performance. The methodology used to determine debris generation amounts contains conservatism, but also uncertainty. The staff determined that the methodology is generally conservative, with few exceptions. However, the use of the debris generation values from the most challenging breaks clearly results in a substantially conservative bias because of the very low frequency associated with these breaks. The smaller, more likely breaks will result in much lower debris generation and reduced challenges to LTCC.
The transport assumptions used in this evaluation and by licensees are based on approved guidance that is considered conservative. The assumptions about transport and debris characteristics are chosen to maximize the amount of debris arriving at the strainer. In addition, the debris amounts represent the bounding breaks for each debris type.
The debris characteristics assumed for the transport evaluation are chosen to increase the probability that debris will be transported to the strainer. Coating particulate is assumed to be 10-micron particles that are suspended in the fluid. Testing has shown that the coatings will fail in a range of sizes and that some of the particulate will be large enough that it will settle in the sump pool or LP of the reactor. Fine fiber is assumed to be individual strands that remain suspended in the coolant. Clumps of fiber in larger size ranges are assumed to be the smallest size within the range, making them more likely to be predicted to transport.
Fluid velocities and turbulence used in transport calculations are based on flows and coolant entering the pool that are developed from the largest breaks and maximized turbulence from fluid falling into the pool. Smaller breaks would result in lower flow rates, and reduced turbulence from water falling into the pool, and they might not result in flow from the CSS if CSS is not initiated. The lower flow rates and reduced turbulence would result in increased debris settling. In addition, lower flow rates result in delayed switchover to recirculation because the RWST is depleted more slowly. The delayed switchover allows more time for debris to settle in a relatively stagnant pool. A smaller break can also result in lower flow rates to the RCS and a higher CSS to ECCS flow ratio. The higher flow ratio results in the return of more fluid to the sump, which allows any entrained debris to pass through the strainer again with the potential for filtering.
The transport evaluation assumes that particulate is not filtered but continues to recirculate until it reaches the core and remains there. Testing has demonstrated that some particulate is deposited on fiber beds relatively early. As the debris bed becomes denser, the amount of particulate filtered increases. Some particulate would also settle or deposit on surfaces. Early in the scenario, the conservatism associated with the assumption that particulate debris is not filtered may be small.
Credit is not given for the capture of fine debris on interceptors, walls, or other structures in containment or within the RV. These structures would capture some fine debris.
It is assumed that all fibrous debris is captured at the core inlet for both the CLB and HLB cases. Some of the fiber will pass through the core inlet and remain in the heated core region or exit the core with excess flow (for the HLB). The amount of fiber assumed trapped at the core inlet is conservative.
Testing conducted for a new reactor design demonstrated that some fiber will settle in the LP and not be available for transport to the core inlet. Settling of fiber is not credited.
Therefore, the amounts of fiber and particulate used for this evaluation are conservative.
5.2.1.5 Chemical Effects Volume 5 of WCAP-17788 contains the in-vessel chemical effects evaluation (14). The following discussion summarizes the PWROG in-vessel chemical effects testing and the NRC staff evaluation. For greater detail, the draft NRC staff SE for Volume 5 is non-publicly available in ADAMS (27). The following discussion provides a high-level summary of the PWROG chemical effects testing and the NRC staff evaluation. Section 5.2.1.2 of this report gives an overview of the WCAP-17788 methodology, including how chemical effects were considered for a HLB.
5.2.1.5.1 Testing PWROG survey data were used to create post-LOCA plant materials and sump chemistry inputs for the test plan. The plant materials input consisted of metals (including aluminum and galvanized steel), insulation materials, fire barrier material, and concrete. The fibrous and particulate insulation materials included calcium silicate, E-glass (fiberglass), silica powder, mineral wool, microporous insulation, and aluminum silicate. Fire barrier material included Interam'. The sump chemistry inputs included the projected post-LOCA temperature profile, sump pool volume, pool alkalinity (pH), and buffer type. The temperature profiles were based on either a post-LOCA temperature profile provided by a plant or were estimated using temperature data provided by the plants with their survey responses.
Testing was performed with independently controlled 60-liter capacity stainless steel autoclaves (pressure vessels). Two resistive heater bands were mounted on the outside of each vessel to provide the heating for the elevated test temperatures. One of these heating bands was located near the bottom of the autoclave to promote convection and more uniform temperatures. The test set-up included means to introduce pressurized air and to vent air. A stainless-steel sampling line was located at approximately the mid-height of the autoclave vessel to avoid sampling either easily settled materials or floating materials in the fluid samples. The autoclave assemblies also included pressure gauges and pressure relief valves. Thermocouples provided temperature measurement of the test fluid. A computerized data system used feedback from the thermocouples to automatically control the temperature of the autoclaves to a
pre-determined profile specific to the test group. Metallic test coupons were suspended from the top of the autoclave using polytetrafluoroethylene (PTFE) cord material to electrically isolate the coupon from the stainless-steel autoclave. Depending on the type of autoclave test, fibrous and particulate debris were either placed in a bag to constrain the loose debris (i.e., in-bag tests) or were freely dispersed within the autoclave test fluid (i.e., initial battery out of bag (IBOB) and out-of-bag tests).
The NRC staff visited the Westinghouse laboratory in Churchill, Pennsylvania in May 2013 and November 2013 to witness testing and to discuss the test methodology and preliminary results.
The visits were intended to improve staff understanding of the tests to better support a review of the WCAP-17788 submittal. The NRC staff trip report dated June 10, 2013 (28) is available in ADAMS. The NRC staff trip report dated December 24, 2013 (29) is non-publicly available in ADAMS.
The initial autoclave water chemistry included representative amounts of boric acid for all tests.
Autoclave test pH depended on the plant-specific conditions. For tests that used a sodium hydroxide (NaOH) buffer, pH adjustment occurred before insertion of the test materials into the autoclaves since NaOH is inducted into the containment spray immediately following a LOCA.
Therefore, the pH for plant groups with NaOH was constant throughout each test except for any small changes caused by debris reactions. For plants-specific tests using sodium tetraborate (STB) or trisodium phosphate (TSP) buffers, the sodium carbonate and the buffer were added after the coupons had been exposed to the lower of either the maximum test temperature or 200 degrees Fahrenheit (°F) for 30 minutes. This sequencing is intended to expose the test materials to a more representative initial acidic plant pH before buffer dissolution. In the event of a LOCA, the sump pool pH for plants with STB buffering increases as STB is either dissolved from retention baskets or is introduced from melting ice (in the case of plants with ice condensers). For plants with TSP buffering, the pool pH following a LOCA increases as TSP dissolves from retention baskets. Following the addition of buffer and carbonate, the autoclaves are pressurized with air (approximately 30-50 psig) and controlled to a temperature profile specific to the test group.
Fluid samples were taken from the autoclave tests at various times and were immediately placed in an oven at a predetermined temperature lower than that of the autoclave. The exact oven temperature was dependent on the autoclave fluid temperature at the time of sampling.
The temperature difference between the autoclave temperature and the oven temperature was intended to simulate the temperature drop of sump fluid passing through a heat exchanger before entering the RV. After a minimum oven hold time of 1 hour1.157407e-5 days <br />2.777778e-4 hours <br />1.653439e-6 weeks <br />3.805e-7 months <br />, the samples were analyzed by performing vacuum filtration (to determine filtration times for each sample) and by turbidity measurement.
Vacuum filtration was used to determine the specific timing of the chemical precipitate product.
As a screening criterion, the PWROG established a filtration time (drain time) threshold value of 100 seconds for 50 milliliters of fluid to indicate the possible presence of chemical precipitates in the test fluid. The vacuum filter membranes used were a 1.0-micron pore size PTFE fiber filter designed for high-temperature filtration.
The PWROG hot filtration test objective was to detect the presence of any chemical products that could cause high pressure drops across a fiber bed in the RV during the initial 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a LOCA. Since drain times can vary for a variety of reasons not related to chemical precipitates, the PWROG selected a 100 second drain time as the screening threshold for chemical effects based on earlier testing performed by Westinghouse and experience with the
first six WCAP-17788 test groups. The 100 second threshold value resulted in identification of chemical precipitates and generated false positives, when drain times exceeded 100 seconds as a result of non-chemical debris. Tests with drain times exceeding 100 seconds and uncertainty about the cause of the longer drain times were then sorted out by further experimentation (e.g., bench tests). In Section 7.1 of Volume 5 of WCAP-17788, the PWROG later evaluated and confirmed the appropriateness of a 100 second drain time threshold value.
The NRC staff also evaluated the filtration test 100 second drain time threshold for chemical effects screening. The staff reviewed the filtration test results and found the 100 second drain (filtration) time acceptable as a screening threshold as discussed in the draft SE. Based on the information provided by the PWROG, the staff found the temperature profiles and oven cooling before hot filtering that were used in the test program to be acceptable.
The Volume 5 test results showed that aluminum precipitates were the cause for high drain times related to chemical effects. Finding aluminum-based precipitates to be the dominant contributor to post-LOCA chemical effects in PWR environments is consistent with previous testing results, including NUREG/CR-6913, Chemical Effects Head-Loss Research in Support of Generic Safety Issue 191, issued December 2006 (30); NUREG/CR-6914, Volume 2, Integrated Chemical Effects Test Project, issued September 2006 (31); and WCAP-16530-NP-A, Evaluation of Post-Accident Chemical Effects in Containment Sump Fluids to Support GSI-191, issued March 2008 (32). This is also consistent with NRC staff observations from many vertical loop and strainer chemical effect headloss tests performed at five different strainer vendor facilities, including tests with addition of pre-mixed WCAP-16530-NP-A precipitates and tests in which dissolved chemicals were added and precipitation occurred in the test loop (33).
The NRC staff finds the aluminum precipitation boundary identified in Section 7 of Volume 5 of WCAP-17788 to be an acceptable prediction tool for chemical precipitation within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> of a LOCA. The staff compared Argonne National Laboratory (ANL) and WCAP-17788, Volume 5, solubility predictions at various pH and temperatures. Although the ANL solubility equations (34) are more conservative at all temperatures of interest when the pH is less than 8.5, the ANL equations over-predict solubility for the high pH tests that can be representative of plants that add NaOH to the containment spray following a LOCA. The WCAP-17788 precipitation boundary predicted all cases of aluminum precipitation in Volume 5 testing, except for Test Group 41, which was not representative of a plant condition. In contrast to the one test group exception, however, there were 39 test samples where the aluminum precipitation boundary predicted chemical effects but there was no evidence of precipitation from subsequent filtration testing. In addition, there were 38 test samples where the precipitation boundary predicted precipitation and it was detected during filtration testing. These numbers indicate that the precipitation boundary function proposed in Volume 5 is conservative within the 24-hour post-LOCA time frame of interest.
Although the NRC finds the aluminum precipitation boundary function identified in WCAP-17788, Volume 5, Section 7, to be conservative, the staff also recognizes there is inherent uncertainty associated with a predictive solubility tool. The PWROG addressed the uncertainty in the solubility boundary by proposing that plants retain margin to the boundary, as shown in Table RAI 5.13-1 (35). PWROG guidance in Section 7.6 indicates that plants that have changed material quantities from the test group can demonstrate that the change is acceptable by either (1) showing that the actual plant conditions reduce precipitation risk relative to the test, or (2) demonstrating that there is still sufficient margin to precipitation if the deviation from the test condition results in an increase in the risk of precipitation. The NRC staff finds the
PWROG guidance related to the margin to precipitation values acceptable since it accounts for potential errors in temperature, pH, and inductively-coupled plasma mass spectrometer measurements in a conservative manner.
Based on the considerations summarized above and discussed in more detail in the draft SE, the NRC staff finds the overall technical approach in Volume 5 of WCAP-17788-P to be acceptable for determining short term (within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following a LOCA) chemical effects in the RV since this approach used representative test conditions, accounted for uncertainties, and had an overall test approach that was effective in detecting chemical precipitation.
5.2.1.5.2 Test ResultsPrecipitation Timing The overall WCAP-17788 methodology relies on the time to chemical precipitation being longer than the minimum time at which AFPs into the core can accommodate complete core inlet blockage. WCAP-17788 expresses success as tchem > tblock. Section 5.3.3 further describes tblock. Volume 1 of WCAP-17788 states that tblock is less than 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> for all plant designs (36).
At that point, flow through AFPs into the core is sufficient for LTCC according to the WCAP-17788 analysis. The autoclave and filtration testing performed for Volume 5 showed that a 6-hour precipitation timing is conservative for almost all PWRs.
Figure 2 shows the timing of chemical precipitation based on the WCAP-17788, Volume 5, PWR test group data. One test group with plant representative conditions had precipitation earlier than 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />.
Figure 2: Timing of Chemical Precipitation for WCAP-17788 Test Groups
For the remaining test groups, the NRC staff concluded that one test group may have experienced precipitation as early as 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />, although the PWROG concluded that precipitation did not occur until 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> for this group based on the mass of aluminum on the filter paper.
Five other test groups showed evidence of precipitation in the 24-hour samples. One of these five test groups with precipitation at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> only performed an out of bag test with plant representative amounts of debris. Earlier in bag tests with extra aluminum margin had early chemical effects. The plant-representative out of bag test contained about 2/3 of the aluminum coupon surface area compared to the earlier in bag tests with early chemical effects.
No precipitation was detected up through the 24-hour sampling time for 37 of 44 test groups.
5.2.1.5.3 Staff Analysis of Precipitate Timing The NRC staff used the results from the WCAP-17788, Volume 5, chemical effects testing to consider precipitation timing and other post-LOCA operator actions that could impact the integrity of a fibrous debris bed at the core inlet. The NRC staff compared chemical precipitation timing to the time when safety injection flow paths are modified (e.g., switching to HL injection or the equivalent) to preclude BAP. Switching safety injection to the HL, injecting simultaneously into the HL and CL, or an equivalent modified flow path (for some plant designs) before chemical precipitation is expected to reduce the probability of a high resistance debris bed forming at the core inlet. The NRC staff observed during the brine testing (6) that a reduction in flow allowed the buoyancy-driven circulation to disrupt the debris bed. This suggests that HLSO could disrupt a debris bed at the core inlet.
The NRC staff evaluation focused the timing evaluation on the seven test groups (representing nine units) that confirmed chemical precipitation during the WCAP-17788, Volume 5, testing.
One test group (two units) with confirmed chemical precipitates before 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> represents a two-unit, low fiber plant that has closed out in-vessel concerns under the conservative fiber limits established by the NRC staff in the SE for WCAP-16793. Therefore, assuming chemical precipitates arrive at the core inlet at this plant earlier than 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />, there is insufficient fiber to form a bed capable of blocking flow at the core inlet. This plant calculated a post-LOCA in-vessel fiber quantity of approximately 10 g/FA, significantly less than the 15 g/FA acceptance criterion the NRC staff approved in the WCAP-16793-NP-A SE. The staff approved the 15 g/FA limit since FA testing showed that the pressure drop from this quantity of fiber, with particulate and chemical precipitates, is acceptable. The single FA tests were designed to promote uniform flow to the bottom of the FA to ensure that a fiber bed was formed. This plant design also has UPI following a LOCA. Fiber bed formation at the top of a UPI plant core following a LOCA is not expected because of turbulence from boiling (see Section 5.2.1.3).
Other than the units discussed above, the results from the next earliest test group indicated that precipitation may have occurred as early as 8 hours9.259259e-5 days <br />0.00222 hours <br />1.322751e-5 weeks <br />3.044e-6 months <br />. The staff notes that the PWROG considered the confirmed precipitation time for this plant to be 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. This plant switches to HL injection no later than 6.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> following a LOCA.
One additional test group (one unit) with confirmed chemical effects at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> using plant-representative conditions, had early chemical effects during initial in bag testing (with greater than representative materials, including aluminum surface area). The final out of bag test for this group was performed with representative amounts of calcium silicate, fiberglass,
and aluminum. This plant switches to simultaneous injection in the HL between 5.5 and 6.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br /> after a LOCA. Without a plant representative in bag test, the staff is uncertain if precipitation would occur before HLSO since previous tests for this group showed in bag testing produced earlier chemical effects than out of bag tests.
Four other test groups confirmed chemical precipitation at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Three of the four groups have a maximum time before safety injection flow into the HL (or other flow injection into the RV) to prevent boron precipitation equal to 6.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />. The final unit that confirmed precipitation at 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> has a HL injection time of 14 hours1.62037e-4 days <br />0.00389 hours <br />2.314815e-5 weeks <br />5.327e-6 months <br /> following a LOCA.
The remaining 37 test groups did not detect precipitation within the 24-hour test period. Each of these plants is expected to flush the core within 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br /> following the assumed LOCA, thereby increasing the probability that any bed on the core inlet would be disrupted before chemical precipitation.
5.2.1.5.4 Chemical Deposits on Fuel The staff previously evaluated the effects of chemical deposits on fuel cladding as part of its SE of WCAP-16793-NP-A. That SE explains why the NRC staff approved the use of the Loss-of-Coolant Accident Deposition Model (LOCADM) to evaluate deposits on fuel following a LOCA. The NRC staff reviewed LOCADM analyses for the 21 units that have closed GL 2004-02 and none of these units have results that challenge the 800 °F or 50-mil thick deposit acceptance criteria. This includes units with high fiber loading. LOCADM is a tool developed by the PWROG that can be used to validate that chemical deposits will not prevent adequate LTCC.
5.2.1.6 ConclusionAvailability of Primary Flowpath The staff concludes that following a LOCA, debris is not expected to completely block the normal ECCS flowpath through the core and prevent adequate LTCC. This is supported by testing and analysis that, taken as a whole, make this an implausible event. This conclusion considers the timing of events, understanding of the TH behavior of the system, and safety margins in the analyses. Should the primary flowpath through the core inlet become blocked, the staff considered DID offered by AFPs and other considerations. Section 5.3 describes these in more detail.
5.3 Defense in Depth In the preceding sections, the NRC staff developed a basis to conclude that a core-wide inlet blockage arising from IVDEs following a postulated LOCA is an unlikely scenario. If, however, the core inlet became completely blocked with a high resistance debris bed, the NRC staff identified multiple mitigation features to provide LTCC, or to assure that the potential inhibition of LTCC does not lead to unacceptable public health and safety consequences. Since the NRC
staff determined that a core-wide inlet blockage is unlikely, these are considered DID. The sections below evaluate each DID mechanism identified by the staff.
5.3.1 Containment Blockage at the core inlet does not affect the ability of the CSS to depressurize and remove heat from the containment. Therefore, the failure of containment is not expected to result from an in-vessel blockage and the containment barrier is assumed to remain intact to provide DID for the IVDEs under consideration in this TER. The staff recognizes that the debris being evaluated for in-vessel blockage can also block the ECCS strainers that supply water to the CSS. However, this evaluation assumes acceptable strainer performance has been demonstrated. The staff evaluates strainer performance separately.
The staff considered that a loss of LTCC could result in core damage that might affect strainers located close to the RV if the vessel is breached due to molten fuel reaching the lower head.
The staff determined that few plants have with this configuration, and some of these have remote strainers in addition to the ones near the RV. The staff also determined that even if the core inlet is blocked, coolant delivered to the LP will reduce the potential for molten fuel to reach the lower head and for vessel breach. Therefore, it is unlikely that IVDEs would cause strainer failure and CSS will remain functional.
5.3.2 Operator Actions All PWRs use HLSO or a similar action to prevent the buildup of boric acid in the core. In response to the issue of the effects of debris on LTCC, PWRs have initiated plant-specific actions to monitor for, and respond to, loss of flow to the core. HLSO is one of the actions that the plants use to inject coolant via a flowpath that is not blocked with debris (injection to the top of the core). Use of HL injection will also tend to disrupt any bed formed at the core inlet by reversing flow through the core. This will be more effective for CLBs, although reducing or stopping flow to the core inlet was shown to remove debris beds from the BN for low-resistance beds.
In addition, industry developed actions that include monitoring and responding to blockage at the core inlet. These measures include monitoring core exit thermocouples, reactor levels, and containment radiation levels from the control room. If core blockage is detected, the plant operators have multiple options to mitigate the condition. Mitigative measures include refilling the RWST and injecting via different flowpaths, reducing RCS pressure, restarting reactor coolant pumps, and initiating early HL injection. Therefore, in the unlikely event that a LOCA occurs, a high-resistance debris bed forms at the core inlet, and AFPs are not successful in providing flow to the core, operator action options exist to maintain LTCC.
5.3.3 Alternate Flowpaths If the core inlet becomes completely blocked, each plant has AFPs to provide coolant to the core for LTCC purposes. These AFPs include design features such as flow passages in the BB region, upper head spray nozzles, and pressure relief holes between the BB region and the core
periphery. A series of TH analyses described in Volume 4 of WCAP-17788 (16) and associated RAI responses (37) (38) (39) evaluates the efficacy of the AFPs to provide LTCC.
The material described in this section relates to debris accumulating at the core inlet (specifically, the BN or spacer grid) after a HLB. Another consideration is that debris present in the coolant not captured at the core inlet, passes through the AFPs, and continues to the heated core region. This can happen in two ways: (1) resistance at the inlet builds to the point where AFPs are activated, directing flow through the AFP to the core or (2) operator action occurs to initiate HLSO, in which a fraction of the ECCS flow is transferred from the CL to the HL to mitigate possible BAP. The combined amount of fibrous debris that can accumulate at the core inlet and within the core without affecting LTCC is assumed to be [xxxxxxxx], which is discussed in Volume 1, Section 6.4, of WCAP-17788. The staff evaluated in-core fibrous and particulate debris as described in Section 5.2.1.1.3 of this TER. The remainder of this section considers the effect of debris blocking the core inlet.
5.3.3.1 Background Once a plant depletes its injectable inventory, its ECCS is aligned for recirculation via SSO. At that point, fibrous and particulate debris may enter the RCS. Such debris may collect at the core inlet as the recirculating coolant travels from the LP into the core. Over time, as debris continues to collect, the hydraulic resistance at the core inlet will increase. The presence of chemical effects, such as precipitates, may also contribute to increased inlet flow resistance. If the flow resistance continues to increase, a greater fraction of liquid will enter the core through the AFPs as opposed to the core inlet. If the core inlet becomes completely blocked, the AFPs become the sole source of liquid entering the core to replace that which is boiled off or spilled out of the break.
Many variables can influence the scenario described above, including the following:
- the depletion time of the source of injectable emergency coolant,
- the amount of debris entering the RCS,
- the fraction of that debris that can be transported to the core inlet,
- the amount by which the hydraulic resistance at the core inlet increases over time,
- the ability of, and time at which, transported debris may cause a complete core inlet blockage The scenario is further complicated by plant-specific considerations, including design features such as flow connections between the BB and core regions or ECCS flow capacities, and the particular operating characteristics (e.g., core power distribution) at the time of the initiating event.
5.3.3.2 Simplifying Assumptions To distill this complex scenario into a tractable problem, the PWROG made several simplifying assumptions in evaluating the efficacy of the AFPs. Some of these assumptions are summarized below:
- break location - The AFP analyses simulated a HLB because it results in the largest fraction of recirculating ECCS flow traveling to the core inlet, where debris could then collect to form a bed.
- SSO time - The AFP analyses assumed that SSO would occur 20 minutes following the initiating event. This is intended to represent a bounding minimum that would cause most debris to be transported into the RCS early, while the decay heat remains high.
- core operating characteristics - The AFP analyses assumed the limiting FA was operating with a top-peaked power shape, similar to that assumed in small-break LOCA analyses, to maximize a second heatup if the core uncovers during a core inlet blockage.
- reactor design differences - Four separate sets of analyses make up the AFP analyses, each set intended to represent a different class of operating PWR reactors. These included: (1) an upflow-configured, 4-loop Westinghouse plant, (2) a downflow configured, 3-loop Westinghouse plant, (3) a Combustion Engineering (CE) plant, and (4) a Babcock & Wilcox (B&W) plant.
- analytic methods - The AFP analyses were performed using EMs that were derived from NRC-approved EMs used to analyze short-term ECCS performance.
- AFP treatment - Each class of plant has different AFPs that were modeled. The AFP resistances were hand calculated using physical plant geometry and, via scaling analyses, were adjusted to ensure that core power was considered as well (see below).
The treatment of rated thermal power level also warrants additional discussion. In WCAP-17788, the thermal power levels selected for each category were not necessarily bounding for each plant in that category. Instead, the PWROG intended to account for variations in core power in the way that the AFP resistances were modeled. The PWROGs response to RAI 4.2 (38) provided a scaling analysis to demonstrate that the core power was considered by scaling the AFP resistance for the modeled plant. Although the NRC staff did not review this RAI response in sufficient detail to make a regulatory finding, it provides information to indicate that the analyses included a means to account for variations in core power level without modeling an explicitly bounding, high-power level.
Although these simplifying assumptions enabled a reasonably consistent set of analyses among the four design classes, it should be noted that the CE analyses were performed using slightly different approaches, in some cases, because the minimum ECCS flow per FA is lower than for the other plants. As appropriate, the sections below note these differences. Furthermore, because the CE plant analyses used a different, and generally less limiting, set of assumptions, the NRC staff performed supplemental analyses for the CE plants using the TRACE computer code, as discussed in Section 5.4.
5.3.3.3 Description of Analyses The goal of the WCAP-17788 analyses in Volume 4 (16) was to determine if AFPs were adequate to maintain LTCC after switchover to recirculation and formation of a debris bed blocking the core inlet. While 10 CFR 50.46(b)(5) requires that the core be maintained at an acceptably low value to remove decay heat from the core, the PWROG opted for an
acceptance criterion of 800 °F as a more quantitative, acceptably low value to show that LTCC is not compromised.
In this TER, however, the NRC staff evaluated the AFP analyses differently. The NRC staff considered the AFP analyses, insofar as the analyses provided evidence that the AFPs would ensure LTCC if a complete core inlet blockage occurred. It should be noted that, while the AFP analyses were performed using methods based on NRC-approved models, the analytic results were not considered to provide an exact prediction of the PCT, or to provide a bounding prediction of behavior for the entire PWR fleet. As discussed below, some parameters, such as the assumed plant power level, do not bound all plants in the category analyzed. However, the NRC staff considers the analyses to establish the concept that the AFPs can provide an additional means to cool the core thus preventing core damage beyond that which occurs during the blowdown and recovery phases of the event. The NRC staff determined that this conclusion is appropriate, given the low likelihood of complete core blockage following the already unlikely initiating event.
The AFP analyses described in WCAP-17788 include three variablesi that characterize the effects of debris blockages at the core:
(1) Kmax is the maximum resistance due to fibrous and particulate debris bed formation that can be tolerated at the core inlet. This value is compared to results of subscale headloss testing to establish an upper bound of the amount of fibrous debris that can be tolerated at the core inlet. A Kmax value corresponding to the headloss testing results for each FA was applied beginning at the SSO time, although the CE analyses modeled Kmax as a piecewise linear ramp beginning 800 seconds after the SSO time. This difference was justified in the response to RAI 4.26 (39). In actuality, Kmax is a function of ECCS flow rate. The limiting cases are derived from the lowest ECCS flow rates.
(2) tblock is the earliest time that complete core blockage can occur while not compromising LTCC. While Kmax was applied at SSO time (except for the CE analyses identified above), the core inlet flow resistance was increased to a prohibitively high value to prevent flow through the core inlet and establish the AFPs as the primary source of core cooling. For the CE plants, the analysis for tblock was separate from the Kmax analysis.
(3) tchem is the time at which chemical precipitates form, as determined by autoclave testing.
The methodology assumes that after tchem, the resistance through the debris bed is greatly increased because of chemical precipitate formation. Due to this increased resistance, the WCAP-17788 methodology requires the timing of complete core blockage, tblock, to be less than the timing of chemical precipitates, tchem. If tchem is less than tblock, then the method in WCAP-17788 is not valid and further analysis is needed.
To simplify the calculations and provide margin, the Kmax values were determined at limiting conditions as discussed above. The Kmax values are converted to debris limits as described in Volume 6 of WCAP-17788 (40) and discussed further in Section 5.2.1.2.
i In a separate series of analyses, the PWROG also evaluated the parameters ksplit and msplit, which related to the core inlet flow resistances and flow rates required to divert flow into the AFPs, but these are not evaluated in detail in this TER. They merely assure that, at any given ECCS flow rate, some portion of flow will divert to the AFPs as the core becomes blocked.
The Kmax and tblock values are determined iteratively. That is, for Kmax, the value was increased until the PCT limit was approached when the resistance was applied at SSO (except for the CE analysis which is discussed above). For tblock, the time of application of a near infinite resistance was decreased until the PCT limit was approached.
5.3.3.4 Analytic Methods The PWROG used three separate computer codes to perform the AFP analyses:
(1) The Westinghouse upflow and downflow analyses were performed using WCOBRA/TRAC. The NRC staff approved this code for use as a part of the Westinghouse Code Qualification Document (41) and Automated Statistical Treatment of Uncertainty Method (ASTRUM (42)) ECCS EMs.
(2) The CE analyses were performed using S-RELAP5. The NRC staff approved S-RELAP5 for use, among other analytic methods, as a part of the Realistic Large-Break LOCA (RLBLOCA) EM documented in EMF-2103(P)(A), Revision 1, Realistic Large Break LOCA Methodology for Pressurized Water Reactors, issued January 2015 (43),
and Revision 3, issued June 2016 (44).
(3) The B&W analyses were performed using RELAP5/Mod2-B&W (R5M2BW) (45). The NRC staff approved R5M2BW for use, among other analytic methods, as a part of the Babcock & Wilcox Nuclear Technologies (BWNT) LOCA EM (46).
WCOBRA/TRAC and S-RELAP5 are approved as elements of realistic ECCS EMs, in which the computer code is applied with a statistical methodology that accounts for uncertainties, while R5M2BW is approved as an element of an EM conforming to 10 CFR Part 50, Appendix K, ECCS Evaluation Models, in which the EM relies on a series of bounding, conservative assumptions to account for uncertainties in the results. Regardless of the computer code however, all the AFP evaluations were performed in a deterministic sense, with reliance on conservative assumptions to offset the effects that uncertainties could have on the predicted results. No attempt was made to quantify uncertainties using the statistical features of the Code Qualification Document, ASTRUM, or RLBLOCA.
5.3.3.5 Conservative Assumptions As described above, these analyses relied on a set of generally conservative modeling assumptions to offset some uncertainties that were not explicitly addressed. While this evaluation is not intended to review these conservatisms exhaustively, examples include the following:
- Decay heat was modeled similar to the requirements in 10 CFR Part 50, Appendix K.
The Westinghouse upflow and downflow analyses, as well as the B&W analyses, used the 1971 American Nuclear Society (ANS) Proposed Standard, Decay Energy Release Rates Following Shutdown of Uranium Fueled Thermal Reactors (47) with a 1.2 multiplier. The CE analyses for tblock used the 1979 standard American National Standards Institute (ANSI)/ANS-5.1-1979, Decay Heat Power in Light Water Reactors (48), for decay heat with a 1.1 multiplier, while the CE Kmax analyses used the 1979 ANS standard for decay heat with a 1.2 multiplier.
- Owing to variability in conceivable ECCS recirculation flow rates, the PWROG studied a range of flow rates on a per-FA basis. The limiting results for the Kmax/tblock analyses were obtained from minimum recirculation flows, which are the basis for this evaluation.
- For the limiting analyses summarized in Section 5.3.3.6, Kmax was applied at the time of SSO, with the exception of the CE Kmax analysis, which applied a piecewise linear ramp beginning at 2,000 seconds (i.e., 800 seconds after SSO time, as described in the response to RAI 4.26 (39)). The response to RAI 4.26 also showed a reasonable amount of conservatism in the CE Kmax modeling approach.
- For the Westinghouse upflow and downflow analyses, as well as the CE analyses, a complete blockage was modeled instantaneously at tblock. This is conservative because a debris bed would take time to build at the core inlet, especially to the point where inlet blockage occurs. Modeling these blockages with a more gradual ramp would result in earlier establishment of the AFPs and could even prevent a secondary heatup.
However, instantaneous blockages were modeled to obviate the need to justify a particular time-dependent ramp of the core inlet flow resistance.
- Because the B&W plants have low resistance AFPs, resulting in greater coolability, they were evaluated assuming a complete blockage at the core inlet beginning at the time of SSO.
5.3.3.6 Results In aggregate, the limiting analyses for each plant indicated that, if the core inlet blocked with a finite amount of fibrous and particulate debris at SSO, followed by a complete blockage at some later time, a secondary cladding heatup would occur. In the Westinghouse analyses for Kmax, the heatup occurred coincident with the application of the Kmax blockage at the time of SSO. In the Westinghouse tblock analyses, a heatup of similar magnitude occurred with a complete core blockage at time tblock. In both cases, Kmax and tblock were modeled to ensure that PCT did not exceed 800 °F. In the CE analyses, there was not a significant heatup in the Kmax analysis, but there was a heatup to about 545 °F after tblock. Meanwhile, the B&W analyses indicated that, because of the ample flow communication between the BB and core region for that category of plants, LTCC would be assured with no secondary heatup, even if the core inlet is completely blocked at the time of SSO. Altogether, the analytic results indicate that the AFPs are an available source of DID, in the unlikely event that the entire core becomes blocked with debris.
The limiting results from the analyses in WCAP-17788 are summarized in Table 2.
In WCAP-17788, Volume 4, and in response to NRC staff RAIs, the PWROG also performed sensitivity studies, which demonstrated the conservative nature of the analyses. For example, for the Westinghouse upflow analyses, the PWROG also simulated a linear ramp in core inlet flow resistance in two cases. These cases illustrated more benign system behavior compared to the Kmax analyses, and the PWROG stated that these linear ramp cases were more realistic regarding the timing at which debris is expected to arrive at the core inlet. As discussed elsewhere in this TER, the NRC staff agrees with this assessment because the debris would
require transport to the core from the sump, such that an instantaneous blockage at the time of SSO is unlikely.
Table 2: Debris Limits and the Associated tblock Plant Design Fuel Type Debris Limit (g/FA) tblock (min) PCT (°F)
Westinghouse [xxxx]
Westinghouse Upflow 143 780 Framatome [xxxx]
Westinghouse Westinghouse [xxxx]
260 790 Downflow Framatome [xxxx]
Westinghouse [xxx]
Westinghouse [xxx]
B&W 20 +
Framatome [xxxxx]
- These represent maximum values based on the debris load tested.
+The B&W analysis did not indicate a secondary cladding heatup.
As discussed in Section 5.3.3.3, the PWROG applied a PCT acceptance criterion of 800 °F or less to demonstrate adequate LTCC. The NRC staff did not find that the results demonstrated, with high probability, that LTCC would be ensured. However, the NRC staff accepts the analyses as a demonstration that AFPs provide a measure of DID. The specific results are quantified above to illustrate the variability among analyses, as well as to correlate the results to in-plant debris limits, specific tblock values, and to provide a comparison to the TRACE supplemental analyses (21) discussed in Section 5.4.
These analyses assume that in-vessel debris is limited to a value that would correspond to a core inlet flow resistance that was modeled as Kmax for each plant type. The PWROG performed subscale headloss testing, as described in Section 6 of Volume 1 of WCAP-17788, to correlate specific fiber limits to the Kmax values modeled in the AFP analyses.
Tables 6-3 and 6-5 of WCAP-17788, Volume 1, provide these fiber limits, which are summarized above in Table 2.
5.4 TRACE Sensitivity Studies RES conducted TRACE analyses (21) to explore the effects of debris on LTCC. These analyses apply to both HLB and CLB conditions. Details of the various studies performed are shown in Table 3. None of these are confirmatory calculations in the traditional sense in that they were not intended to validate or verify specific industry results. The first two columns include studies performed in the 2015/2016 timeframe. They were completed before WCAP-17788 was submitted to the NRC for review. The second two columns are a result of the current effort to aid in determining the safety significance of IVDEs.
As shown in Table 3, the TRACE analyses were performed for two specific plant models that are representative of or conservative compared to the operating PWR fleet. Modeling was not performed using plant-specific conditions but used realistic inputs and assumptions to provide insights regarding LTCC behavior considering the effects of debris. The analyses also provided
insights into how changes in important assumptions affect behavior. The analyses show that flow within the core and LP regions will promote non-uniform debris distribution at the core inlet.
They also show that liquid entrainment will limit the concentration of boric acid within the core region such that BAP is not a concern if licensees implement CL injection as required in their licensing bases. The TRACE work provides a separate and independent source of knowledge regarding the effect of IVDEs on LTCC.
Table 3: TRACE Sensitivity Studies CLB BAP HLB BAP/AFP HLB AFP CLB AFP Purpose Evaluate effects of Evaluate effects Evaluate PWROG Evaluate debris buildup at of debris buildup conservatisms PWROG core inlet on BAP at core inlet on conservatisms BAP and LTCC Model DEG, 4-loop W upflow, mid-peaked axial, DEG, CE, top-peaked axial, ANS 79 ANS 79 Cases/ 99.9% blocked - 99.5-100% -Decay heat -Base case LCP blocked core only Sensitivities inlet -Bypass 100% blocked nozzle -Blockage timing
-HLSO time, decay -Rate of blockage heat, axial skew -HPI flow 5.4.1 CLB For the CLB, the TRACE analyses address both the effects of blockage at the core inlet and the potential for BAP.
The TRACE runs conducted by RES (21) for the CLB indicate that it is unlikely for BAP to occur, or for LTCC to be inhibited because of a loss of flow, even if a debris bed forms at the core inlet.
TRACE runs were performed for a Westinghouse 4-loop upflow design PWR and a CE design PWR. The majority of the CLB analyses were conducted for the Westinghouse design and the discussions in this section concentrate on those analyses. The staff concluded that for the CLB, the mechanical design of the reactor would not have a large influence on the analysis.
In the Westinghouse simulations, TRACE was modified to account for higher densities and viscosities that occur because of the concentration of boric acid within the core region following a CLB. The TRACE runs confirmed the temperature driven circulation patterns that were demonstrated in industry modeling. The simulations assumed that recirculation started at 1,200 seconds to ensure the use of high decay heat values. Several cases were run. A
reference case that assumed no blockage was performed for comparison to the cases that simulated blockage. The reference case and most other cases assumed a mid-core peak and nominal decay heat based on the 1979 ANS draft Standard ANS-5.1/N18.5, Decay Energy Release Rates Following Shutdown of Uranium Fueled Thermal Reactors (48). The reference case found that up until about 10,000 seconds, the boric acid concentration in the core region was limited because of entrainment of liquid in the steam flowing to the SGs. As decay heat (and entrainment) decreased, the concentration of boric acid in the core increased. The concentration of boric acid in the core and LP was uniform. Mixing with the LP was caused by the upflow in the central core and downflow in the periphery. The difference in boric acid concentration between the two volumes was not significant enough to cause buoyancy-driven circulation. However, the boric acid concentration in the LP remained slightly lower than that in the core. It was assumed that HL injection was initiated at 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />, which eliminated the increase of boric acid concentration in the core. The baseline simulation was continued to 7.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />, at which time the core was covered, and boric acid concentrations returned to low levels.
5.4.1.1 Westinghouse Upflow Plant The following list includes the reference case assumptions for the Westinghouse Upflow Plant TRACE CLB analysis.
- DEGB of the CL
- No debris blockage
- SSO at 20 minutes
- Blockage applied from 1,200 seconds to attain full resistance at 5,000 seconds, linear ramp
- HLSO at 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br />
- Nominal decay heat
- End of simulation at 7.5 hours5.787037e-5 days <br />0.00139 hours <br />8.267196e-6 weeks <br />1.9025e-6 months <br />
- Flat axial power peaked at mid-core
- 212 °F recirculation temperature Additional TRACE cases simulated progressively increasing levels of debris blockage. The blockage was simulated at the lower core plate, below the area where blockage would occur in the plant. This is a conservative scenario because simulating the blockage at this location prevents flow through the AFPs credited in the PWROG analyses for the HLB from WCAP-17788. Other sensitivity studies, discussed later in this section, applied blockage at the core inlet which allows flow to the AFPs. The blockage was applied by closing valves to the assumed blocked area. RES checked the resistance (k/A2) value for the valves against the expected debris resistance value to ensure the modeling was valid. The blockage was applied over time starting at 1,200 seconds and was ramped to the full blockage for the case being run
over the next 3,800 seconds, or just longer than an hour, which is considered reasonable. The full blockage value for the case was assumed to occur at 5,000 seconds.
Each of the sensitivity studies for the Westinghouse upflow plant are summarized in the following sections.
5.4.1.1.1 Blockage at the Lower Core Plate (99 percent)
In the case where 99 percent of the core area was blocked, the core remained covered and no temperature excursions occurred. Boric acid concentration was elevated but remained well below the concentration where BAP could occur. Localized BAP was observed only in the UH in areas where trapped fluid boiled away leaving the boric acid solids behind. There was no BAP predicted at the core inlet, BB regions, or areas that could impede coolant flow to the core.
The simulation of HLSO effectively ended the transient, even when the resistance at the lower core plate was assumed to remain in place.
5.4.1.1.2 Blockage at the Lower Core Plate (99.5 and 99.9 percent)
The case for 99.5 percent blockage had slightly elevated boron concentration compared to the 99 percent case, but LTCC was maintained. Cases where core inlet blockage exceeded 99.9 percent resulted in cladding heat up as blockage at the core inlet prevented adequate coolant from entering the core. In this simulation, boric acid concentration in the core was found to be below the precipitation limit, even when the PCT excursion occurred. This finding implies that fuel failure caused by loss of coolant flow will occur before failures caused by BAP. The boric acid concentration did not approach the solubility limit in any of the blockage sensitivity studies.
5.4.1.1.3 Blockage at the Fuel Inlet Nozzles (99 and 100 Percent)
The TRACE simulations investigated the effect of debris blockage location by simulating blockage at the fuel inlet nozzles instead of the lower core plate. Blockage at this location prevents (or reduces) coolant flow entering the core directly but allows flow to the BB region. A 99 percent blockage was applied at the fuel inlet nozzles. Results of this simulation showed that LTCC continued and boric acid concentration remained well below precipitation limits. Flow from the BB region into the core near the top nozzles mixed with the coolant in the core and diluted boric acid in the core.
For the 100 percent blockage case, the only flowpath to the core was through the BB region.
Core coolability was maintained and boric acid concentration was maintained well below the BAP limit. Flow through the BB region and into the core exceeded boil off from decay heat.
Because of the low resistance assumed at the core inlet in WCAP-17788 CLB cases, flow through the BB region was not credited to ensure LTCC. The identification of the BB flowpath by the TRACE simulations indicates that the diverse flowpaths credited for the HLB can also provide the required flow for the CLB. The flow is through the same pathways credited for the HLB analysis, so all reactor designs evaluated have flowpaths available to allow LTCC for a CLB if the core inlet becomes completely blocked.
5.4.1.1.4 Blockage at the Fuel Inlet Nozzle with Delayed HLSO (100 percent)
A case was run to evaluate the effects of a much later switchover to HL injection. Instead of assuming 6 hours6.944444e-5 days <br />0.00167 hours <br />9.920635e-6 weeks <br />2.283e-6 months <br /> for the initiation of HL injection, the case was run for 36 hours4.166667e-4 days <br />0.01 hours <br />5.952381e-5 weeks <br />1.3698e-5 months <br />. The case assumed 100 percent blockage at the FA inlet. The boric acid concentration increased over time but did not exceed the solubility limit. The concentration plateaued at around 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />.
The boric acid in the core remained well mixed. Some precipitation was predicted in the UP after 30 hours3.472222e-4 days <br />0.00833 hours <br />4.960317e-5 weeks <br />1.1415e-5 months <br />. This likely resulted in maintaining the core concentration below the solubility limit.
5.4.1.1.5 Core Power Sensitivity A sensitivity case using Appendix K decay heat assumptions was run by increasing the reference case decay heat by 30 percent. A 99 percent blockage was assumed at the lower core plate. The higher power case resulted in a more highly voided core and UP, but no heatup occurred. Compared to the similar case at the reference decay heat value, the boric acid concentrations were significantly lower and the rate of increase in concentration remained low for a longer time. This is caused by entrainment of the boric acid in the liquid exiting the core.
5.4.1.1.6 Axial Power Shape Sensitivities Top and bottom skewed axial power shape cases were run. Both cases assumed 99 percent blockage at the lower core plate. There were differences in the very early PCT results during blowdown and reflood, but the LTCC results were very similar.
An interesting result from the decay heat sensitivity is that higher decay heat results in lower boric acid concentrations. Greater liquid entrainment resulted in more boric acid removal from the core. This is counterintuitive unless entrainment, which is not credited for design basis calculations, is considered. The TRACE simulations also provide more realistic results than the SKBOR methodology discussed in Section 5.2.1.1.2 in that they account for the physical effects of liquid carryover in the steam and for viscosity and density changes resulting from concentrated boric acid.
5.4.1.1.7 Large Split Break at Top of CL A sensitivity case was run to examine the effects of a split break at the top of the CL. The results were like those of the DEGB cases.
For all Westinghouse CLB cases, there is liquid carryover at the beginning of the event so that the buildup of boron in the core region is limited. Later in the event, boron may begin to concentrate in the core. However, the increased density of boric acid in the core region will promote buoyancy driven convection and mixing with the LP. This convection, convection caused by heating of the fluid, and the agitation caused by boiling results in mixing within the core and between the core and LP. Early in the LOCA recovery, heat-driven convection and boiling will drive the mixing. Later, convection resulting from density increases from boron concentration may become more dominant.
5.4.1.2 Combustion Engineering Plant The CE cases were run mainly to gain understanding of the behavior of the CE design following an HLB. The CE cases did not model concentration of boric acid in the core. The single CLB case run for the CE design used a decay heat multiplier of 1.1 times the ANS 1979 decay heat curve and a top peaked axial power shape. The case simulated full core blockage at 20,000 seconds. No cladding heatup was predicted. This case shows that the AFPs can provide cooling for the CE design even if a full core blockage occurs. The conclusion is valid for the conditions simulated but is dependent on several variables including the time blockage occurs.
5.4.1.3 TRACE Study Applicability to the PWR Fleet The Westinghouse and CE HLB TRACE studies are discussed in Section 5.4.2.
Applicability of the BAP TRACE study to the PWR fleet considered the following:
- The TRACE runs modeled only two reactor designs (Westinghouse 4-loop upflow and CE), and most of the CLB cases were run for the Westinghouse upflow design.
- The boron concentration aspects of the model were validated to Primrkreislauf-Versuchsanlage (commonly known as PKL) data with good correlation, but additional validation of modeling for boron concentration and density was not performed because of a lack of available data. See Section 4 of the TRACE report (49).
Other parts of the model were more rigorously validated.
- The simulations did not use design-basis assumptions but performed sensitivity studies to understand how assumptions can affect the results. On the other hand, the studies considered core inlet blockage, which has not been considered in design-basis calculations.
- The results are not considered design basis type calculations, but they indicate that BAP is much less likely to occur than previously postulated and provide independent information regarding the PWROG calculations for core cooling and BAP.
The TRACE results also identified that circulatory currents are present within the RV. These currents will prevent uniform debris bed formation at the core inlet. The PWROG simulations showed that these currents were present early in the LOCA recovery period but, unlike the TRACE studies, did not attempt to demonstrate that they would be present later in the recovery because the concentration of boric acid in the core was not modeled.
RES concluded that the qualitative results of the TRACE analysis should be applicable to all plants that have bypass holes that will not become blocked by debris. UPI plants are significantly different so are not comparable. Variations in key plant parameters will affect the quantitative results. Westinghouse and B&W designs have core bypass flow paths that are qualitatively similar to those of the CE plants. B&W and some Westinghouse designs have AFPs that connect the bypass region to the core at lower elevations so that coolant can enter the core without filling up the entire bypass flow volume. The BAP results from the Westinghouse plant CLB TRACE analysis are applicable to the CE and B&W plants since there are significant mixing volumes above the core and in the HLs for all designs. The CLB debris
limits are the same for all fuel and reactor designs and the response of the system is similar enough for the results to be considered applicable.
The NRC staff considers the Westinghouse plant selected for boric acid concentration studies conservative relative to the PWR fleet due to its upflow configuration in the core bypass region.
The upflow design reduces the mixing volume available for dilution of boric acid. The design also delays flow from the AFP to the core if a 100 percent blockage at the core inlet nozzles occurs because the BB must fill to the top before flow to the core is established.
The Westinghouse CLB TRACE analysis found that core uncovery occurs before BAP takes place. Because core uncovery is the central issue during long-term cooling, additional calculations were made for the representative CE plant. These calculations showed that even with a conservative blockage at the fuel inlet nozzles, uncovery is not expected to occur for this design, similar to the Westinghouse plant findings. The calculations and sensitivity studies performed in the TRACE report indicate that even with a highly conservative blockage at the core inlet nozzles, neither core uncovery nor BAP is expected for the CLB case.
In the unlikely event that the fuel inlets become completely blocked, TRACE simulations show that the BB will fill with coolant and coolant will flow into the core via AFPs. TRACE demonstrates that the core remains covered. The AFP through the BB region that was credited for the HLB (see Section 5.3.3) is also viable for the CLB, although industry analyses for the CLB do not credit it.
5.4.2 HLB RES performed a series of HLB sensitivity studies using the TRACE code for both the Westinghouse upflow and CE plant categories as described below.
5.4.2.1 Westinghouse Upflow Plant RES performed Westinghouse upflow sensitivity studies to determine if a break in the HL could result in high boric acid concentrations in the core and RV. The plant selected for simulations was a Westinghouse four loop PWR with a core power of 3,411 megawatts thermal. The plant model (representative of the Seabrook Station) is an upflow plant, meaning that core bypass flow enters the BB region near the core inlet and exits near the fuel top nozzle elevation. No communication was modeled between the BB region and the active core. This configuration is assumed to be conservative for investigations involving core inlet blockage with debris. Coolant cannot reach the core unless it flows completely though the BB region and back into the core near the core exit, unless it flows through the high-resistance path of the debris.
For the HLB calculations, the blockage was simulated to occur at the bottom fuel nozzles using a resistance ramped from 1,200 to 5,000 seconds. The total blockage ranged from 99 percent to 100 percent. At 100 percent blockage, no flow could go directly into the bottom of the core, and any coolant reaching the core was forced to flow through the BB region and enter at the top of the core. While the BB region filled, sufficient inventory remained in the core region so that the rods remained covered and cladding temperatures remained at saturation temperature.
After the BB region refilled, sufficient coolant entered at the top of the core to maintain core coverage and prevent cladding heat up.
At 100 percent blockage, the core boric acid concentration was only slightly greater than that obtained in the calculations with 99.5 percent blockage or less. Flow through the BB region was sufficient to remove decay heat and keep the two-phase level at the HL elevation.
Both HL calculations (99.5 and 100 percent blockages) maintain the two-phase level to the level of the HL, and as a result the carryover fraction is nearly 1.0. This is because when the two-phase level is at the level of the HL, coolant can easily flow into the HL and to the break.
Essentially, the vessel is full and a mass balance exists, and the total ECCS flow is matched by the total flow out of the vessel. Excess coolant simply flows to the break. Thus, the HL calculations do not indicate the minimum amount of carryover necessary to prevent boron precipitation.
5.4.2.2 Combustion Engineering Plant This section compares a TRACE analysis for CE to the PWROG analysis that used S-RELAP5.
Objectives include: (1) determining how the dependence of the long-term core depression and heatup varies with core blockage ramp times, (2) determining whether there is a ramp rate that would preclude a core uncovery and heatup altogether, and (3) understanding the sensitivity of the results to assumptions about key parameters and boundary conditions.
The TRACE plant model used for these calculations is similar to the St. Lucie Plant. The analyzed event consists of a double-ended guillotine HLB with only high-pressure injection (HPI) available and a containment pressure of 1 atmosphere during long-term cooling. The core inlet is assumed to be completely blocked by debris at some point in time and core cooling water must flow through the core bypass flow path and empty into the top of the core. Sensitivity calculations include varying key input parameters and assumptions including decay heat, bypass flow resistance, HPI flow rate, core blockage time, and blockage rate.
The analyses showed that there is no significant fuel heatup in the long term as the result of core blockage. Sensitivity calculations showed that decay heat, bypass flow resistance, HPI flow rate, core blockage time, and blockage rate could be changed significantly with little sensitivity to the core level after the blockage. The results are qualitatively similar to those of the PWROG analysis.
5.4.3 Applicability to the Pressurized Water Reactor Fleet TRACE sensitivity studies were not performed for the Westinghouse downflow or B&W plant categories. The qualitative results of this analysis should be applicable to all plants that have bypass holes large enough to not be blocked by debris after the switchover to sump recirculation. WCAP-17788 tested openings down to 1/4-inch to demonstrate that these sized openings would not become blocked with debris. Openings less than 1/4-inch were not credited in the WCAP-17788 analyses. Modeling of different plant designs would reveal quantitative differences that depend on the details of the key plant parameters such as specific plant dimensions, core power, bypass resistance, and ECCS flow rate. However, based on the industry TH analyses and the TRACE studies performed for the Westinghouse upflow and CE plants and the similarities in PWR designs, the NRC staff does not believe that additional TRACE studies are necessary.
5.5 Integrated Decisionmaking 5.5.1 Determination of Safety Significance The NRC staff prepared this TER using integrated decisionmaking which includes an assessment of traditional engineering evaluations and probabilistic information. Integrated decisionmaking considers information from a variety of sources including risk insights, traditional engineering evaluations and insights, operational experience and historical plant performance, engineering judgment, and current regulatory requirements. In addition, evaluation of DID and safety margins are an important part of the integrated decisionmaking process.
While a quantitative analysis of the risk associated with IVDEs following a hypothetical LOCA at each U.S. PWR is impractical, the NRC staff combined risk insights with its engineering evaluation, including the low initiating frequency of LOCAs that can challenge LTCC via IVDEs to determine the safety significance of IVDEs. This section discusses the NRC programs that were used to evaluate the safety significance of the issue.
The NRC staff considered guidance that is presented in: LIC-504, Revision 4, Integrated Risk-Informed Decision-Making Process for Emergent Issues (50); NUREG/BR-0058, Regulatory Analysis Guidelines of the U.S. Nuclear Regulatory Commission (51); RG 1.174, Revision 3, An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis, January 2018 (52); and the Significance Determination Process (SDP) (IMC 0609) (53). Table 4 provides a summary of the risk metrics associated with those programs. The guidelines for assessing risk in the established programs discussed in Table 4 were derived from the Commissions Safety Goal Quantitative Health Objectives (QHOs) and the values were selected to be consistent with the Safety Goal QHOs (see, for example, SECY 97-077, (54)). All these processes use integrated decisionmaking and include considerations of DID and safety margin in addition to risk information to assess appropriate regulatory response relevant to the process (e.g., reactive inspection, license amendment request approval, etc.). As described in NRC guidance, each of these processes are tailored to the context of their regulatory use. For this reason, the extent to which they use the risk assessment in conjunction with the other elements of integrated decisionmaking varies.
Further, the risk assessments used for these processes vary in scope, level of detail, and technical rigor.
For example, RG 1.174 is used to evaluate plant licensing changes. In this process, the change in risk and its comparison against risk acceptance guidelines is one of five key principles that are considered by staff in its evaluation. NRC guidance for performing regulatory analysis for substantial safety enhancements includes risk thresholds for identification of such enhancements (51). The thresholds are starting points and the evaluation includes consideration of available safety margins and DID, which together with the thresholds influence staff decisions. LIC-504 is NRC staff guidance for evaluating emergent issues and refers to the same five key principles defined in RG 1.174 as important inputs to be used in addition to risk values calculated in the process. The SDP uses PRA analyses to evaluate the change in risk resulting from a degraded condition associated with a licensee performance deficiency and relates the significance to established thresholds. The SDP process also includes consideration of qualitative factors as part of the assessment and consideration of safety margins and DID is prominent in cases where the impact of the performance deficiency cannot be readily captured
using PRA analyses. The values in Table 4 are simplified representations of the thresholds from these processes.
Table 4: Risk Metric Lower Risk Risk Metric ( CDF/Rx Year) Greater Risk Reference 10-7 or Lower 10 10-6 10 10-5 10 10-4 10-4 or Higher RG 1.174 Acceptable Evaluate Unacceptable SDP Green White Yellow Red LIC-504 No Action Evaluate (Shut Down if CDF > 10-3)
NUREG/BR-No Action Evaluate Take Action 0058 As stated above, the NRC staff did not quantify the risk attributable to IVDEs following a LOCA.
However, as discussed in Section 5.5.2 the staff determined that the order of magnitude of the occurrence frequency of LOCAs that can challenge LTCC via IVDEs would be 1 x 10-6 per year or less. Assuming that an initiating event (LOCA) that can challenge LTCC via IVDEs leads directly to core damage will result in an order of magnitude of 1 x 10-6 per year or less for core damage frequency. This is consistent with the low end of the quantified risk spectrum for existing NRC programs and corresponding guidance referenced in Section 5.5.1. Additionally, given the high probability of the containment remaining intact following a LOCA including IVDEs, an increase in the frequency of large early releases attributable to IVDEs is not expected (see Section 5.3.1).
Therefore, based on its evaluation summarized in this TER, including the low initiating frequency of LOCAs consequential enough to challenge LTCC via IVDEs (see Section 5.5.2),
DID, and the available safety margins, the staff concludes that a quantitative assessment would indicate that the core damage frequency attributable to the types of IVDEs considered in this TER are of low safety significance. Using integrated decisionmaking, the evidence considered by the staff indicates that the safety significance associated with IVDEs is expected to be in the acceptable range for several existing NRC processes shown in Table 4.
The safety significance approach followed in this TER has not been evaluated for use outside the TER. Therefore, it should not be used on a generic basis unless specifically evaluated for such a use.
5.5.2 Risk Insights - Initiating Event (LOCA) Frequencies Although the NRC staff did not quantify the risk attributable to IVDEs following a LOCA, it sought to understand the frequency of such occurrences to inform its decisionmaking. To do so, the staff sought to determine an order of magnitude estimate for initiating frequency of LOCAs that can challenge LTCC via IVDEs based on the engineering analyses discussed in Section 5.2 of this TER.
The NRC staffs evaluation of the information presented by STP (13) determined that debris generation from a DEGB of a 6 inch or larger pipe on the HL side could result in a fibrous debris bed achieving the threshold limit for challenging LTCC via IVDEs. Section 5.2.1.2 explains how the fiber amount associated with the 6-inch pipe break was determined to be applicable on a generic basis. The 6-inch pipes of interest at STP have an inner diameter of 5.19 inches.
Therefore, the staff sought to determine an order of magnitude estimate for LOCA initiating frequency for breaks 5 inches and larger. Table 7.7 of NUREG-1829, Estimating Loss-of-Coolant Accident (LOCA) Frequencies Through the Elicitation Process, issued April 2008 (55), provides the mean initiating LOCA frequencies for different break categories (each of which corresponds to an effective break size) based on 40-year plant operation. The staff considers the 40-year frequency estimates to be applicable to plants that are operating beyond 40 years. Those values include both piping and non-piping contributions. Table 7.2 of NUREG-1829 provides the ratio of non-piping to piping contributions for each break category.
The non-piping contribution is based on the LOCA initiating frequency estimates at the time of the expert elicitation in NUREG-1829. However, based on information in Sections 6.3.4 and 7.2 of NUREG-1829, the non-piping contribution is expected to remain the same or decrease somewhat. Therefore, the ratio of non-piping to piping contributions for each break category in Table 7.2 were used because they were considered applicable to determining the order of magnitude estimate desired for this evaluation. The resulting initiating frequency for all LOCAs 5-inches or greater is approximately 2 x 10-5 per year as determined by staff calculations. The corresponding value for the arithmetic mean estimates for 40-year plant operation using information from Table 7.11 of NUREG-1829 is approximately 6 x 10-5 per year. Therefore, in both cases, the order of magnitude of relevant breaks is 10-5 per year.ii These estimates are for all piping 5-inches or larger regardless of its location in the RCS.
Since breaks in piping 5 inches or larger on the HL side are consequential to this TER, a refined order of magnitude estimate of the initiating frequency was determined using the expert elicitation data from NUREG-1829 (56). The expert elicitation data were aggregated to determine the initiating frequency of 5 inch or larger breaks in the HL as well as piping connected to the HL such as the surge line. The result is based on the values for 40 years of operation. The overconfidence adjustment derived based on Tables 7.1 and 7.7 in NUREG-1829 was applied. The refined estimate for the initiating frequency for all LOCAs 5 inches and greater on the HL side (i.e., the HL as well as piping connected to the HL) is approximately 1.5 x 10-6 per year (i.e., the order of magnitude is 10-6 per year). This estimate includes consideration of the chemical and volume control system (CVCS) which contributes significantly to the estimate. CVCS lines greater than 5 inches are uncommon and exclusion of the CVCS for the refined estimate results in an initiating frequency for all LOCAs 5 inches and greater on the HL side of approximately 3.6 x 10-7 per year (i.e., the order of magnitude is 10-7 per year).
The size of the break that could generate enough debris to challenge LTCC for a break on the CL side is much larger because smaller breaks do not generate enough debris to block flow to the core. Two cases for very high fiber plants are discussed in terms of generic applicability in Section 5.2.1.1. These demonstrate that only very large breaks, 12 inches and larger, generate significant debris for breaks on the CL side. The staff determined the order of magnitude for breaks in piping 12 inches and larger from Tables 7.7 and 7.2 of NUREG-1829. The resulting initiating frequency for all LOCAs 12 inches and greater in piping is approximately 1 x 10-6 per year. The corresponding value for the arithmetic mean is approximately 6 x 10-6 per year.
ii The estimates in NUREG-1829 are on a per calendar year basis whereas initiating frequencies use a per reactor year basis. The difference between the two is the capacity factor (i.e. the portion of a calendar year that a reactor is at full power). Since the purpose is to determine an order of magnitude estimate, the staff used the calendar year values as if they were on a reactor year basis (i.e. did not account for the capacity factor) because the capacity factor (approximately 0.9 for the entire U.S. fleet) will not impact the order of magnitude estimate and the conclusions drawn therefrom.
Therefore, the order of magnitude of the initiating frequency of the relevant breaks is 10-6 per year. These estimates are for all piping 12 inches or larger, regardless of its location in the RCS. Using the expert elicitation data from NUREG-1829 and the approach outlined in the previous paragraph, a refined estimate of the initiating frequency of breaks in piping 12 inches or larger on the CL side (i.e., the CL as well as piping connected to the CL) is 1.4 x 10-7 (i.e., the order of magnitude is 10-7 per year).
In its comparison with the spectrum of risk thresholds in Section 5.5.1, the NRC staff assumed that an initiating event as defined in this section leads directly to core damage. That is, given a HLB greater than 5 inches or a CLB greater than 12 inches, core damage is assumed. The staff concludes that the margins not included in the assessment of core blockage reduce the risk significantly below that based solely on break frequency and that if core blockage occurs, there is DID available to provide LTCC.
In summary, the estimated order of magnitude of the frequencies of the initiating events (LOCAs) that can generate enough debris to challenge LTCC via IVDEs, along with the sequence of events necessary to result in a high-resistance debris bed at the core inlet discussed in Section 5.2 of this report, supports the staffs conclusions in this TER.
5.5.3 Safety Margins Safety margins associated with the evaluation of IVDEs depend on plant-specific parameters.
The margins associated with the analyses are discussed throughout this document and summarized here.
- This evaluation is performed assuming limiting conditions for debris generation. Many breaks will not generate the amounts of debris assumed.
- The transport assumptions in the evaluation are conservative. All fine fibrous debris that penetrates the strainer is assumed to be transported to the core. All fine particulate debris is assumed to reach the core. Some of this debris would settle or deposit on surfaces.
- The filtering characteristics of the strainer make larger breaks less problematic than would be assumed from their debris generation amounts. For the HLB, a 6-inch break is assumed. Larger breaks may generate significantly more debris, but the amount of debris that passes through the strainer decreases quickly with strainer load.
- The assumed time for switchover to recirculation is conservative. It is based on the smallest RWST water volume and highest ECCS and CSS flows for each plant. The assumed switchover time is 20 minutes for all AFP HLB analyses and all TRACE analyses (21). The fastest time for many plants is about 30 minutes. Times for less limiting but more likely breaks are much longer than 30 minutes. Longer switchover times result in lower decay heat levels which allow AFPs more time to respond to prevent any secondary heatup. Longer switchover times also allow more time for debris to settle.
- Unqualified coatings are assumed to fail as 100 percent fine particulate that is transported to the strainer and core. These coatings would actually fail in a range of sizes and some would not transport to the core.
- The evaluation of particulate debris concentration ignored liquid that would be present in the UP and used a limiting plant value for particulate debris mass. The concentration of particulate will be lower than that calculated for the CLB.
- A 50 percent void fraction is assumed for the core region for the particulate debris concentration calculation. The actual void fraction will be lower and will decrease with time. The concentration of particulate debris for the CLB will be lower than assumed.
- The AFP analysis assumed a uniform bed. This leads to a much lower fibrous debris amount than would actually be required to form a high-resistance bed at the core inlet.
The fiber limits referred to in this analysis are based on a uniform bed. It is likely that a high resistance bed would not form even with significantly more fiber than represented by the limits.
- Chemical effects will probably not occur until after the plants have initiated HLSO. FA testing conducted for WCAP-16793 found that significantly larger amounts of fiber could be accommodated without causing excessive headloss if precipitates are not present.
The staff identified plants that have the potential to incur precipitation before 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />.
The PWROG chemical effects testing for WCAP-17788 identified that out of 44 groups, 37 would not incur precipitation before 24 hours2.777778e-4 days <br />0.00667 hours <br />3.968254e-5 weeks <br />9.132e-6 months <br />. Two plants were identified that could have precipitation before HLSO. Of these, one is a low fiber plant that has already closed GL 2004-02 including IVDEs. The second plant is a single unit for which the staff was unable to confirm that precipitation would occur after HLSO and without more information, the staff considers this unit to be outside the key parameters of this TER.
- In the AFP analysis, the limiting fiber amounts were determined from testing using the limiting flow rate, limiting particulate to fiber ratio, and limiting particulate size distribution.
Changing these parameters resulted in lower head loss values for the same amounts of fiber.
- The core inlet resistance value Kmax was determined using a 60 second ramp rate for increasing the resistance at the time of SSO for Westinghouse and B&W designs. An alternate approach was used for CE designs. This results in a conservative Kmax value, and therefore a conservative fiber limit, as described in Section 5.3.3.5.
- The design of RCS piping assures that it is robust. Implementation of code inspections for Class 1 piping provides some assurance that the piping does not have defects where LOCAs can initiate. Many plants have mitigated dissimilar metal welds to reduce the probability of primary water stress corrosion cracking. For unmitigated welds, the code requires augmented inspections for defects.
- The analysis does not credit leak before break. However, the most likely way for large piping to fail is for a crack to initiate and grow to a critical size, then rupture. All PWRs have systems designed to detect leakage at levels that would occur below the critical crack size. Technical specifications require that plants monitor for leakage and shut down when unidentified leakage indicates potential leakage in the RCS. The intention of the technical specification is to ensure the plant is shut down, leakage identified, and repairs implemented as required, before a LOCA occurs.
- The PCT limit of 800 °F used as an acceptance criterion in WCAP-17788 was based on long-term autoclave testing that showed no significant degradation in cladding properties would occur by spending 30 days at 800 °F. WCAP-16793-NP-A, Appendix A, discusses the basis for the temperature limit. A PCT up to this limit is not expected to challenge fuel integrity. The TH analyses from WCAP-17788 and TRACE show only short-term PCT increases.
- Decay heat in the AFP calculations was modeled similar to the requirements in 10 CFR Part 50, Appendix K (see Section 5.3.3.5).
- The limiting results for Kmax/tblock in the AFP analyses were obtained from minimum recirculation flows, which is a conservative methodology.
- For the limiting AFP analyses, tblock was determined using a conservative ramp rate of 60 seconds for Westinghouse and B&W designs. This results in a conservative tblock value (see Section 5.3.3.5).
- B&W plants exhibited adequate LTCC when the core inlet was blocked in the AFP analysis. The B&W plant response to IVDEs was evaluated assuming a complete blockage at the core inlet beginning at the time of SSO.
- The AFP analyses assumed the limiting FA was operating with a top-peaked power shape, similar to that assumed in small-break LOCA analyses. This maximizes any secondary heatup if the core uncovers during a core inlet blockage.
6.0 CONCLUSION
The NRC staff determined that for the operating US PWR fleet, and within the range of assumed plant parameters, debris penetrating the sump strainer is very unlikely to prevent adequate LTCC following LOCAs via IVDEs. This TER considered a wide range of analytical and deterministic information from industry and NRC sources, supplemented by engineering judgement and risk insights.
Based on integrated decisionmaking supported by information from a variety of sources including risk insights, traditional engineering evaluations and insights, operational experience and historical plant performance, engineering judgment, and current regulatory requirements, the NRC staff concludes that the safety significance of the challenge to LTCC following LOCAs via IVDEs is low. The staff believes this analysis is conservative compared to typical operating US PWRs, but it is not an evaluation of any specific plant. It does not address methods for plants to demonstrate compliance with regulations or define licensing basis parameters for the IVDE issue.
The concept of safety significance, and the evaluation methods, findings, and conclusions discussed in this TER are limited in scope to the evaluation of IVDEs with respect to GSI-191 and GL 2004-02. NRC staff does not consider this concept to be applicable to additional issues or topics without further justification.
7.0 REFERENCES
- 1. WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid," July 2013 (ADAMS Package Accession No. ML13239A111).
- 2. PWR Owners Group, Submittal of WCAP-17788: "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090)," July 17, 2015 (ADAMS Package Accession No. ML15210A667).
- 3. Title 10 of the Code of Federal Regulations Part 50 "Domestic Licensing of Production and Utilization Facilities".
- 4. Generic Letter 2004-02, "Potential Impact of Debris Blockage on Emergency Recirculation During Design Basis Accidents at Pressurized Water Reactors," September 13, 2004 (ADAMS Accession No. ML042360586).
- 5. WCAP-17360-P, Rev. 0, "Small Scale Unbuffered and Buffered Boric Acid Nucleate Boiling Heat Transfer Tests with Sump Debris in a Vertical 3x3 Rod Bundle," May 31, 2012 (ADAMS Accession Nos. ML12167A435 and ML12167A437) (non-public).
- 6. PWROG-15091-NP, Rev. 0, "Subscale Brine Test Program Report, Analysis committee, PA-ASC-1188," Parts 1 and 2, November 30, 2015 (ADAMS Accession Nos. ML15342A069 and ML15342A070).
- 7. Enclosure, "Revised Content Guide for Generic Letter 2004-02 Supplemental Responses," November 2007 (ADAMS Accession No. ML073110278).
- 8. Nuclear Energy Institute (NEI) 04-07, Vol. 1, "Pressurized Water Reactor Sump Performance Evaluation Methodology," December 2004 (ADAMS Accession No. ML050550138).
- 9. WCAP-16406-P-A, Rev. 1, "Evaluation of Downstream Sump Debris Effects in Support of GSI 191," March 2008 (not publicly available; proprietary information).
- 10. SECY-12-0093 Closure Options For Generic Safety Issue-191, "Assessment Of Debris Accumulation On Pressurized-Water Reactor Sump Performance," July 12, 2012 (ADAMS Accession No. ML121310648).
- 11. Staff Requirements Memorandum SRM-SECY-12-0093, "Closure Options for Generic Safety Issue-191, Assessment of Debris Accumulation on Pressurized Water Reactor Sump Performance," December 14, 2012 (ADAMS Accession No. ML12349A378).
- 12. Final Safety Evaluation for Topical Report WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate Fibrous and Chemical Debris in the Recirculating Fluid," April 8, 2013 (ADAMS Package Accession No. ML13084A161).
- 13. "South Texas Project, Units 1 and 2, Issuance of Amendment Nos. 212 and 198 - Risk-Informed Approach to Resolving Generic Safety Issue 191," July 11, 2017 (ADAMS Package Accession No. ML17019A001).
- 14. WCAP-17788-P, Vol. 5, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090) - Autoclave Chemical Effects Testing for GSI-191 Long-Term Cooling," July 2015 (ADAMS Accession Nos. ML15210A678, ML15210A680, ML15210A681). (non-public).
- 15. WCAP-17788-NP, Vol. 3, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090) - Cold Leg Break (CLB) Evaluation Method for GSI-191 Long-Term Cooling," July 2015. (ADAMS Accession No. ML15210A671).
- 16. WCAP-17788-NP, Vol. 4, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SE-1090) - Thermal-Hydraulic Analysis of Large Hot Leg Break with Simulation of Core Inlet Blockage," July 2015. (ADAMS Accession Nos. ML15210A672 and ML15210A674).
- 17. "Staff Observations of Plant-Specific Testing for Generic Safety Issue 191 During February 26, 2006, Trip to the Alion Hydraulics Laboratory," April 17, 2006 (ADAMS Accession No. ML060750467).
- 18. "Staff Observations of Testing for Generic Safety Issue 191 During April 28, 2006, Trip to the Alion Hydraulics Laboratory," June 27, 2006 (ADAMS Accession No. ML061720514).
- 19. "Seabrook, Unit 1 - Updated Final Response to NRC Generic Letter 2004-02," January 31, 2018 (ADAMS Accession No. ML18031B248).
- 20. "Analysis Report - Post-LOCA PWR Core Inlet Blockage Assessment," February 26, 2007 (ADAMS Accession No. ML070860521).
- 21. "Assessment of Boric Acid Accumulation in a Reactor Core Blocked by Debris," NRC Office of Nuclear Regulatory Research, March 2019. (ADAMS Accession No. ML19086A270).
- 22. Transmittal of WCAP-17057-P/NP, Rev. 1, "GSI-191 Fuel Assembly Test Report for PWROG," PA-SEE-0312, September 2011 (ADAMS Package Accession No. ML112930439).
- 23. EPRI 1011753, "Design Basis Accident Testing of Pressurized Water Reactor Unqualified Original Equipment Manufacturer Coatings," Final Report, September 30, 2005 (ADAMS Accession No. ML071130075) (non-public).
- 24. "Keeler & Long PPG, "Design basis Accident Testing of Coating Samples from Unit 1 Containment, TXU Comanche Peak SES," April 13, 2006 (ADAMS Accession No. ML070230390).
- 25. Attachment 2: Revised Responses to NRC RAIs Related to WCAP-17788 Supporting the Closure of GSI-191 (PA-SEE-1090), September 28, 2018 (ADAMS Accession No. ML18285A019).
- 26. "Vogtle Electric Generating Plant Units 1 & 2, Supplemental Response to NRC Generic Letter 2004-02," April 21, 2017 (ADAMS Package Accession No. ML17116A096)
(non-publicly available).
- 27. Draft Safety Evaluation Report by the Office of Nuclear Reactor Regulation Topical Report WCAP-17788-P, "Comprehensive Analysis and Test Program for GSI-191 Closure," Vol. 5, "Autoclave Chemical Effects Testing for GSI-191 Long Term Cooling,".
January 10, 2019 (ADAMS Accession No. ML19016A310) (non-public).
- 28. Trip Report for May 2013 NRC Staff Visit to the Westinghouse Churchill Laboratory to Observe Autoclave Testing, June 10, 2013 (ADAMS Accession No. ML13159A007).
- 29. Westinghouse Trip Report of November 19, 2013, December 24, 2013 (ADAMS Accession No. ML13337A479) (non-public).
- 30. NUREG/CR-6913, Chemical Effects Head-Loss Research in Support of Generic Safety Issue 191," Argonne National Laboratory, December 2006 (ADAMS Accession No. ML070090553).
- 31. NUREG/CR-6914, Volume 2, Integrated Chemical Effects Test Project: Test #1 Data Report, September 30, 2006 (ADAMS Accession No. ML062560111) (non-public).
- 32. WCAP-16530-NP-A, "Evaluation of Post-Accident Chemical Effects in Containment Sump Fluids to Support GSI-191," March 2008 (ADAMS Accession No. ML081150379).
- 33. Staff Observations of Testing for Generic Safety Issue 191 During July 22 To July 24, 2009 Trip to the Alion Hydraulics Laboratory, September 24, 2009 (ADAMS Accession No. ML092670458).
- 34. "Aluminum Solubility in Boron Containing Solutions as a Function of pH and Temperature, Argonne National Laboratory, September 19, 2008 (ADAMS Accession No. ML091610696).
- 35. Attachment 1, "Responses to NRC RAIs Related to WCAP-17788, Volume 5 Supporting the Closure of GSI-191 (PA-SEE-1090) and Mark-ups to WCAP-17788 PROPPRIETARY Attachment," May 23, 2017 (ADAMS Accession No. ML17293A220) (non-public).
- 36. WCAP-17788-NP, Vol. 1, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SEE-1090)," July 2015 (ADAMS Accession No. ML15210A669).
- 37. Response to RAIs on WCAP-17788 Vol. 4, B&W Plants Licensing Report, December 31, 2017 (ADAMS Accession No. ML18029A213) (non-public).
- 38. Attachment 1: "Westinghouse Responses to NRC RAIs Related to WCAP-17788, Volumes 1 and 4 Supporting the Closure of GSI-191 (PA-SEE-1090) and Mark-ups to WCAP-17788 Proprietary," December 19, 2017 (ADAMS Accession Nos. ML18029A210 and ML18029A211). (non-public).
- 39. ANP-3583P, Rev. 2, "Response to RAls on WCAP-17788, Volume 4 - CE Plants:
Licensing Report," July 31, 2018 (ADAMS Accession No. ML18285A021) (non-public).
- 40. WCAP-17788-P, Vol. 6, Rev. 0, "Comprehensive Analysis and Test Program for GSI-191 Closure (PA-SE-1090) - Subscale Head Loss Test Program Report," July 2015.
(ADAMS Accession Nos. ML15210A681 and ML15210A682) (non-public).
- 41. WCAP-12945-P-A, "Code Qualification for Best Estimate LOCA Analysis," March 31, 1998 (ADAMS Package Accession No. ML093070051) (non-public).
- 42. Transmittal of WCAP-16009-P-A & WCAP-16009-NP-A, "Realistic Large Break LOCA Evaluation Methodology Using Automated Statistical Treatment of Uncertainty Method (ASTRUM)," January 31, 2005 (ADAMS Package Accession No. ML050910156) (non-public).
- 43. Safety Evaluation for Framatome ANP Topical Report EMF-2103(P) Rev. 1, Realistic Large Break LOCA Methodology for PWRs," Appendix H, "Containment Pressure" (TAC No. MC4259), January 19, 2015 (ADAMS Accession No. ML050190227) (non-public).
- 44. AREVA, Inc. Regarding Final Safety Evaluation for Topical Report EMF-2103(P),
Rev. 3, Realistic Large Break LOCA Methodology for Pressurized Water Reactors,(TAC No. ME2904), June 2016 (ADAMS Package Accession No. ML16286A579).
- 45. Acceptance of Final Safety Evaluation for AREVA NP, Inc Topical Report BAW-10164(P), Rev. 6, RELAP5/MOD2-B&W - An Advanced Computer Program for Light Water Reactor LOCA and Non LOCA Transient Analysis, June 25, 2007. (ADAMS Accession No. ML091410456) (non-public).
- 46. Framatome Technologies, Inc., BAW-10192P-A, Rev. 0, BWNT LOCA: BWNT Loss-of-Coolant Accident Evaluation Model for Once-Through Steam Generator Plants, BAW-10192P-A, Revision 0, June 1998, (ADAMS Accession No. ML093080467) (non-public; proprietary).
- 47. American Nuclear Society, ANS-5 Standards Subcommittee, Proposed ANS Standard, "Decay Energy Release Rates Following Shutdown of Uranium Fueled Thermal Reactors," October 1971.
- 48. American National Standards Institute/American Nuclear Society, ANSI/ANS-5.1-1979, "Decay Heat Power in Light Water Reactors," August 29, 1979.
- 49. "Assessment of Boric Acid Accumulation in a Reactor Core Blocked by Debris,"
March 2019 (ADAMS Accession No. ML19086A270) (non-public).
- 50. LIC-504, R. 4 "Integrated Risk-Informed Decision-Making Process for Emergent Issues," May 30, 2014 (ADAMS Accession No. ML14035A143).
- 51. NRC, "Regulatory Analysis Guidelines of the U.S. Nuclear Regulatory Commission,"
NUREG/BR-0058, Revision 5, draft for comment, April 30, 2017 (ADAMS Package No.
- 52. Regulatory Guide 1.174, Rev. 3, "An Approach for Using Probabilistic Risk Assessment in Risk-Informed Decisions on Plant-Specific Changes to the Licensing Basis," January 2018 (ADAMS Accession No. ML17317A256).
- 53. IMC 0609, "Significance Determination Process," October 23, 2018 (ADAMS Accession No. ML18187A187).
- 54. SECY 97-077, "Draft Regulatory Guides, Standard Review Plans and NUREG Document in Support of Risk Informed Regulation for Power Reactors," April 8, 1997 (ADAMS Accession No. ML992920137).
- 55. NUREG-1829, "Estimating Loss-of-Coolant Accident (LOCA) Frequencies Through the Elicitation Process," Main Report, April 2008 (ADAMS Accession Nos. ML082250436 and ML081060300).
- 56. PWR Piping Raw Data for NUREG-1829, June 15, 2011 (ADAMS Accession No. ML080560011).
- 57. WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid," July 2013 (ADAMS Package Accession No. ML13239A111). .
- 58. WCAP-16793-NP-A, Rev. 2, "Evaluation of Long-Term Cooling Considering Particulate, Fibrous and Chemical Debris in the Recirculating Fluid," July 2013 (ADAMS Package Accession No. ML13239A111). .