Evaluation of AP1000 Containment Hydrogen Control Strategies for Post- Fukushima Lessons Learned

Transcription

1 Evaluation of AP1000 Containment Hydrogen Control Strategies for Post- Fukushima Lessons Learned James H. Scobel and Hong Xu Westinghouse Electric Company, EEC 1000 Westinghouse Dr. Cranberry Township, PA Tel: , Abstract - The A P I000' plant primarily employs heated-coil igniters for the control o f combustible gases generated in postulated severe accident scenarios. As part o f a post- Eukushima lessons learned assessment o f the A P I000' plant, MAAP4.0.7 analyses were performed assuming the failure o f all AC power and the total failure o f igniter power. The analyses credit the passive autocatalytic recombiners (PARs) for hydrogen control when igniters are not available. The evaluation shows that for a scenario with hydrogen generation associated with 100% oxidation o f the active cladding, the PARs currently included in the plant design are sufficient to prevent flame acceleration and deflagration to detonation transition. Therefore, containment integrity is not challenged by the combustion of hydrogen inside containment. I. INTRODUCTION The AP1000 passive nuclear power plant severe accident management strategy primarily employs heatedcoil igniters for hydrogen and combustible gas control (Reference 1). The igniters are powered by offsite and onsite AC power, and, in the event of a loss of all ac power, safety class 2 batteries that are sized to provide at least 4-hours of dc power for the igniters. The passive containment also includes passive autocatalytic recombiners (PARs) located in the upper compartment that are sized to control the hydrogen releases from postaccident radiolysis of water and corrosion of metal components in the containment. Severe accident hydrogen control analyses performed as part of level 2 PRA (Reference 2) only credit the igniters and neglect the PARs for controlling the hydrogen releases associated with the oxidation of the zirconium fuel cladding during a core damage accident. During a Westinghouse internal post-fukushima review of the A PI000 plant, it was noted that, even though the assessments show that extreme external events will not likely result in a core damage accident in the passive plant, there was still a need to review lessons learned from a severe accident management perspective. Specifically, it was pointed out that while AP1000 plant s strengths relative to Fukushima events rely on its passive nature, severe accident hydrogen management relies on active equipment, e.g. the igniters. Therefore, questions were raised regarding A PI000 hydrogen control strategies for a postulated station blackout (SBO) scenario: 2015 Westinqhouse Electric Company LLC. All Rights Reserved. Is it possible that a severe accident scenario with slow, delayed, or sporadic hydrogen generation could lead to a situation where the batteries were depleted and the igniters were not available during the hydrogen release? Is there a potential need for further enhancement of passive hydrogen control in the AP10008 containment? The purpose of this paper is to evaluate the A PI000 hydrogen control strategy assuming the design configuration with 4-hour igniters and the installed PAR capacity and a passive configuration with the PARs alone for a SBO accident scenario that generates hydrogen associated with 100% cladding oxidation. The analysis will also evaluate alternate passive hydrogen control configurations (PARs only) for comparison to justify that the likelihood of containment failure due to hydrogen combustion is as low as reasonably practicable for the AP1000 plant design. II. AP1000 PLANT HYDROGEN MITIGATION STRATEGY The following diverse features provide defense-indepth severe accident hydrogen mitigation capability to protect the integrity of the A PI000 containment: Reactor Coolant System (RCS) Automatic depressurization system (ADS) valves that provide an engineered path of least resistance for the release of hydrogen into a compartment within the natural circulation flow in the containment (Figure 1) to promote hydrogen mixing. ( A Formatted: Justified ]

2 Figure 1. Automatic Depressurization System (ADS) and Passive Emergency Core Cooling System (PXS) 1 of 2 Trains Shown An active distributed hydrogen igniter system that is designed to single-failure criteria. PARs located in the containment upper compartment. A steel containment vessel that is structurally designed to accommodate the peak pressure produced by the global deflagration of hydrogen associated with 100% cladding reaction without exceeding the ASME Service Level C pressure limit (Reference 2). A passive containment cooling system (PCS) that rapidly mixes the large, open containment volume by natural circulation, minimizing the concentration of hydrogen released to the containment (Reference 1). Passive cooling also maintains an elevated long-term steam partial pressure within the containment that dilutes the hydrogen/air mixture, reducing the potential for flame acceleration and deflagration to detonation transition (Reference 2). Potential hydrogen vent paths in containment are located away from the containment shell to protect the structure from overtemperature due to sustained hydrogen diffusion flames (Reference 2). The ADS system depressurizes the RCS to support passive injection of water for emergency core cooling. In a severe accident, ADS stage 4 valves provide an engineered hydrogen release path-of-least-resistance from the RCS to the containment. The hydrogen flows from the RCS into the loop compartments that participate in the containment natural circulation circuit, which promotes the hydrogen mixing in the containment. The containment hydrogen igniter system is composed of two independent trains of 33 heated-coil igniters per train distributed strategically throughout the containment. The igniters are manually actuated through either the Plant Control System (PLS) or through the Diverse Actuation System (DAS) prior to the onset of cladding oxidation as one of the first steps in the emergency operating procedure for inadequate core cooling. The successful operation of one train of igniters is sufficient to maintain the containment hydrogen concentration near the lower flammability limit to preclude global deflagration and mitigate the potential for flame acceleration (FA) and deflagration-to-detonation transition (DDT). The igniters are powered continuously by offsite ac power and Safety Class 2 diesel generators. In the event of station blackout, each igniter train can be powered by the Safety Class 2 dc power batteries for at least 4 hours (Reference 1). The containment has the equivalent of 4 NIS Type 44 PARs that are located in the containment upper compartment to reduce excess hydrogen in the containment atmosphere over time without the need for power. The volume of the A PI000 passive containment is comparable to conventional pressurized water reactor (PWR) large-dry containments, which historically in the USA, are not required to have active hydrogen control for severe accident mitigation. The large containment volume passively mitigates hydrogen combustion by minimizing

3 the partial pressure of the hydrogen mixed within the containment atmosphere. PCS water cooling of the containment shell effectively mixes the containment atmosphere by natural circulation in conjunction with the release of steam and hydrogen at the bottom of the loop compartments. The open design and flow paths in the ceilings of all subcompartments mitigate locally high hydrogen concentrations. Flow paths where the release of hydrogen-rich plumes may be postulated to produce standing diffusion flames are placed at locations away from the containment shell and penetrations so as to not create a potential for containment overtemperature failure. More than 75% active cladding oxidation is required for the global dry hydrogen concentration to reach 13%, the minimum limit where transition to detonation could be reasonably postulated assuming dry air. However, in a post-accident containment environment with hydrogen, there will always be steam in the AP1000 containment atmosphere. Passive containment cooling produces a sustained elevated steam partial pressure of approximately 0.5 to 1.0 bar. The elevated steam concentration and limited hydrogen concentration reduce the likelihood of flame acceleration (FA) and deflagration-to-detonation transition (DDT). If a global deflagration is postulated, the containment structure is design to withstand the maximum pressure that could be produced by the combustion of the hydrogen associated with 100% active cladding oxidation. III. HYDROGEN SOURCE TERM FOR ANALYSIS The vast majority of in-vessel hydrogen generation that occurs during a severe accident is a result of fuel rod cladding oxidation. The reactor core is uncovered due to a loss of coolant accident (LOCA) and decay heat cannot be effectively removed from the uncovered portion of the core and it overheats. Steam from water boiling at the bottom portion of the core passes over the overheated zirconium cladding surface area in the upper core and oxidation of the cladding produces hydrogen in an exothermic reaction: Zr + 2H20 >2H2 + Z r0 2 + Qrx The extent of the core that is oxidized is controlled by, and may be limited by, the availability of water and the availability of overheated, unreacted zirconium surface area. The water must boil away to uncover and overheat the cladding, and if the water is not replenished it will be depleted and the reaction will be terminated. As the core overheats, oxidizes and loses geometry, the zirconium surface area available for the reaction diminishes and attenuates the reaction. The containment hydrogen control system must be capable of mitigating hydrogen releases associated with 100% cladding oxidation. Thus, a scenario used to establish a hydrogen source term must oxidize 100% of the zirconium cladding. For scenarios to have such extensive cladding oxidation, the major hydrogen generation and release rate to containment occur during the reflooding of an overheated, but relatively intact core. Otherwise the oxidation reaction is water limited and the fraction of cladding reaction is much lower. Rapidly reflooding the core relatively soon after it has overheated maximizes the water availability and cladding surface area participating in the oxidation and exacerbates the hydrogen release. If the reflood rate is slow, the exothermic oxidation of the cladding will cause the core to lose geometry and the reaction will be surface area limited. Reflooding the core quickly will quench the core and will terminate the oxidation reaction. Therefore, it is reasonable to assume that for a 100% cladding oxidation scenario, a large fraction of the hydrogen generation will occur over a short period of time during a core reflood. Subsequent hydrogen release rates are expected to be much slower due to the loss of core geometry and quenching of the core that limits the surface area of unoxidized, overheated zirconium cladding available to support the metal-water reaction. The hydrogen source term used for this analysis is generated by the MAAP4 model of the A PI000 plant using an early core reflooding scenario that passively refloods the core via the break as the break compartment floods with water. MAAP4 parameters are set to maximize cladding oxidation. The integrated mass of hydrogen generated in-vessel is presented in Figure 2. Approximately 60% of the cladding is oxidized during the core reflood and the remainder of the cladding is oxidized over the remainder of a 24 hour period. The total mass of hydrogen generated by the oxidation of 100% of the active cladding in the AP1000 plant is 788 kg. IV. UNCERTAINTY CONSIDERATIONS The scenario considerations for this design extension condition analysis are primarily established by the need to create a core damage event, to oxidize 100% of the active zirconium cladding and to release the hydrogen to the containment. The initiating event is a double-ended direct vessel injection (DVI) line (Figure 1) break within the passive emergency core cooling system (PXS) valve room. The PXS injection squib valves in the intact DVI line fail to open thus creating the core damage scenario. The PXS valve room floods with water draining from the in-containment refueling water storage tank (IRWST) through the PXS side of the break. ADS operation is credited to depressurize the RCS to allow passive reactor vessel and core reflooding through the broken DVI line. The core makeup tank (CMT) drains through the broken DVI line and automatically actuates the ADS on low CMT level. Open ADS valves establish potential hydrogen flow paths from the pressurizer to the IRWST compartment through ADS stages 1, 2 and 3 valves. ADS stage 4 valves release most of the hydrogen from the RCS hot legs to the loop compartments. Some hydrogen is released through the DVI line break. Figure 2 depicts the hydrogen releases distributions to the containment.

4 Hydrogen Release Location from the RCS to the Containment Intg Maes of H2 Gen In-Vessel Intg Mass of H2 Released thru AD5-1/2/ Intg Mass of H2 Released via Loop 1 ADS Intg Mass of H2 Released via Loop 2 ADS Intg Maes of H2 Released via Break Time (hr) Figure 2. Integrated Hydrogen Generation and Distribution through RCS Flow Paths to the Containment The emergency operating procedure (EOP) for inadequate core cooling instructs the operator to actuate the igniters prior to performing actions that will reflood the core. The procedure is entered based on a core-exit thermocouple temperature of 650 C, before in-vessel hydrogen generation begins. Thus the rapid hydrogen releases during core reflooding events are mitigated by the igniters operating on the 4-hour batteries. If there is hydrogen release after the batteries are depleted, these releases would be expected to be mitigated by the containment volume, passive mixing and recombination by the PARs In-vessel retention of core debris is credited in this analysis. The operator manually floods the reactor cavity to prevent reactor vessel failure by externally cooling the reactor vessel surface. This action is performed within the same EOP as the manual actuation of the hydrogen igniters which is entered when the core-exit thermocouple temperature exceeds 650 C. In addition, in this scenario, the reactor vessel and damaged core are reflooded, so core debris in the vessel is water-cooled both inside the vessel and through the vessel wall. Because reactor vessel failure does not occur, the hydrogen generation is due only to invessel cladding oxidation and molten core concrete interaction does not occur in this analysis. The passive containment cooling system (PCS) water is actuated automatically on high containment pressure following the blowdown from the LOCA initiating event. PCS water cooling capacity is maximized assuming 100% wetting of the containment shell to minimize the containment steam concentration, which is conservative for hydrogen combustion because excess steam in the atmosphere attenuates the conditions for flame acceleration and DDT. Therefore, this particular scenario and hydrogen source term address significant uncertainties in severe accident hydrogen analysis. Conservatism is achieved by maximizing the hydrogen release rates by reflooding the core, oxidizing 100% of the cladding to maximize the hydrogen concentrations and minimizing the containment steam concentrations by passively cooling the containment. V. PASSIVE AUTOCATALYTIC RECOMBINER MODELING ASSUMPTIONS PAR units are modeled with the NIS Type 44 PAR model in the MAAP4 code. The MAAP4 PAR model with the parameters as used in this analysis has been benchmarked against test data for the NIS PARs. For this analysis, the PARs are assumed to recombine the hydrogen and oxygen at 80% of their recombination capacity as predicted by MAAP4 to account for uncertainties in the PAR performance in a severe accident environment. It has been seen in testing that the PARs may potentially generate high temperatures and ignite the hydrogen if it reaches globally flammable concentrations (Reference 3). However, the ignition of the hydrogen by the PARs is not credited in this analysis. Assuming the containment gas mixtures are not in the fast flame regime, ignition by the PARs would be beneficial. Therefore, if the hydrogen mixture concentration is controlled such that the containment does not reach conditions that support flame acceleration, it is conservative in this analysis to assume that the PARs do not ignite the hydrogen.

5 VI. ACCEPTANCE CRITERION The AP1000 containment is able to structurally withstand the peak pressure from a slow deflagration of the hydrogen generated from 100% cladding oxidation without exceeding the ASME service level C pressure limit. Potential challenges from diffusion flames to the containment pressure boundary are mitigated by design. Thus the primary challenge to the containment integrity from inadequately controlled hydrogen combustion is due to the potential for flame acceleration and DDT. The conditions that support flame acceleration are calculated based on the CSNI/OECD State-of-the-Art Report on (SOAR) Flame Acceleration and DDT Methodology (Reference 4) that relates the occurrence of flame acceleration to the expansion ratio (o), which is the ratio of the unburned gas mixture density to the burned gas mixture density. The expansion ratio is a function of the mixture composition and gas temperature and provides a measure of the reactivity of the mixture. A critical value of the expansion ratio (ocrit) defines the boundary between slow flame and fast flame combustion regimes. The critical expansion ratio is correlated from a large amount of experimental data as presented in Reference 4 and summarized in Table I. The critical expansion ratio is a function of temperature for lean hydrogen concentrations and is constant for rich concentrations. TABLE I Critical Expansion Ratios (Reference 4) Temperature ^crit (K) Xh2 < 2 X02 Xh2 > 2 X The flame acceleration index a /a crit< 1.0 provides a measure that the hydrogen combustion event would be outside the fast flame regime (Reference 4) and the burn would be a slow deflagration. a /a ciit > 1.0 is a prerequisite for DDT. Therefore, if the flame acceleration index is less than 1.0, flame acceleration and DDT are not possible and containment integrity during a burn is assured. The flame acceleration index is used as a limiting acceptance criterion in this analysis. The potential for flame acceleration is a necessary condition for DDT that based on the thermodynamic properties of the mixture. There is no geometric factor in this criterion. VII. HYDROGEN CONTROL SYSTEM CONFIGURATIONS Five hydrogen control system configurations were modeled and compared to perform the assessment of the AP1000 hydrogen control strategies: Configuration 0: No hydrogen control measures are modeled. No hydrogen combustion is modeled. The purpose of Configuration 0 is to assess the degree of hydrogen challenge if no hydrogen mitigation measures were employed. Configuration 1: The current hydrogen control strategy with 1 successful train of igniters powered for 4-hrs by batteries and 4 PARs in the upper compartment. The purpose of Configuration 1 is to evaluate the current active/passive strategy as the base line. Configuration 2: Current hydrogen control strategy with failure of the igniter power, crediting only the 4 PARs in the upper compartment. The purpose of Configuration 2 is to evaluate the effectiveness of the current design in the event of complete igniter failure. Configuration 3: Passive hydrogen control strategy with 4 PARs in the upper compartment and 2 PARs in each of the loop compartments. The purpose of Configuration 3 is to evaluate the benefit of adding local PARs to the lower compartments that receive the major hydrogen releases through ADS stage 4. Configuration 4: Passive hydrogen control strategy with 8 PARs in the upper compartment. The purpose of Configuration 4 is to evaluate the impact of increasing the passive hydrogen control capacity by twofold. VIII. RESULTS OF THE MAAP4 ANALYSES The results of the MAAP4 analyses are presented in Figures 3 through 12. VIII.A. Configuration 0: No Hydrogen Control Measures Credited For Configuration 0, 100% cladding oxidation with no hydrogen control measures modeled, containment pressure and temperatures are presented in Figures 3 and 4. Local hydrogen/oxygen/steam mixture compositions are presented for the loop compartment control volume (Figure 5), which receives the major hydrogen and steam releases through the ADS stage 4 valves, the containment upper compartment control volume (Figure 6), which represents the well-mixed volumes of the containment, and the IRWST control volume (Figure 7), which is a confined volume with limited hydrogen releases (see Figure 2) through the ADS stages 1, 2, and 3 spargers. The PXS valve room that receives the break flow is water-filled and vents any hydrogen releases (see Figure 2) to the lower compartment, which is well-mixed by the containment natural circulation. Thus the PXS valve room is not considered as a potential location for flame acceleration. The value of a/(jcrit for each compartment is presented in Figure 8 for Configuration 0. Configuration 0 defines the overall A PI000 containment hydrogen conditions that need to be controlled for a hydrogen release profile corresponding to 100%

6 cladding reaction. Containment conditions that briefly approach the potential for flame acceleration may occur in the loop compartments during rapid hydrogen generation that occurs due to the vessel reflooding. As the degree of cladding oxidation approaches 100% over the long term, if the hydrogen is not controlled, the containment atmosphere may reach a condition that will globally support flame acceleration. The small IRWST volume may see elevated hydrogen concentrations, but does not reach conditions supporting flame acceleration due to relatively low temperatures, lower oxygen concentrations and elevated steam concentration. VIII.B. Configuration 1: 4-hr Igniters and Upper Compartment PARs Configuration 1 evaluates the hydrogen control capability of the AP1000 existing active/passive hydrogen control strategy. The igniters are assumed to be on dc power and are credited for 4-hrs following activation. After the batteries are depleted, the upper compartment PARs provide control for long term hydrogen releases. As shown in Figure 9, in all containment compartments, the gas mixtures do not support flame acceleration at any time during the transient. During the reflooding, the elevated hydrogen releases to the loop compartment are mitigated effectively by the igniters. VIII.C. Configuration 2: Upper Compartment PARs Only Configuration 2 evaluates the hydrogen control capability of the AP1000 crediting only the PARs. The results are shown in Figure 10. The PARs do not mitigate local conditions in the loop compartment during rapid hydrogen generation that occurs during core re flooding. However, the containment long-term conditions do not reach the flame acceleration limits. VIII.D. Configuration 3: Addition o f 2 PARs in Each Loop Compartment Configuration 3 evaluates the benefit of adding 2 PARs to each of the loop compartment to mitigate local hydrogen releases from the ADS-4 valves during reflood (Figure 11). However, given the PAR heat up time and limited recombination rates, the results show that the added PARs are not able to mitigate the rapid hydrogen releases associate with early core reflooding. If reflooding were delayed the PARs may be more effective; however later core reflooding would not have such rapid hydrogen release rates due to loss of core geometry and the local conditions in the loop compartment would be mitigated by natural circulation and would not approach the flame acceleration limit. VIII. E. Configuration 4: 2 x Existing PAR Capacity Configuration 4 evaluates the benefit of doubling the PAR capacity in the upper compartment (Figure 12). As expected, the PARs have no effect on the local conditions and the long-term conditions have increased margin to the limits. However, given that the existing configuration adequately controls the hydrogen releases below the flame acceleration limits, increased capacity is not required. IV. CONCLUSIONS The in-vessel hydrogen releases from a severe accident scenario that oxidizes 100% of the fuel rod cladding in the active core region are used for the analysis of the A PI000 containment hydrogen control strategies. Approximately 60% of the cladding is oxidized during reactor vessel reflooding and the remainder of the cladding oxidizes over the course of a 24-hr transient. The hydrogen source term and the distribution to the containment through the various RCS openings are depicted in Figure 2. The analysis credits the in-vessel retention of core debris for mitigating ex-vessel hydrogen releases. According to the results of the analysis, the existing A PI000 hydrogen control configuration of active heatedcoil igniters and 4 NIS Type 44 PARs in the upper compartment provides the optimum hydrogen control configuration for the AP1000 containment. The igniters may be powered continuously by offsite ac power or onsite emergency diesel generators and for more than 4 hours by the Safety Class 2 batteries. The igniter are arranged in two power trains and designed to single failure criteria. One train of igniters is sufficient to provide effective hydrogen control. Even if ac power is lost, the 4-hr igniters on dc power will control rapid hydrogen releases effectively, because such rapid hydrogen release is only expected to occur early after core uncovery before core geometry is lost. The PARs do not require power and are effective on the slower long-term hydrogen releases. Together with the engineered hydrogen release location provided by ADS stage 4 and containment natural circulation mixing induced by passive containment cooling, the AP1000 active/passive hydrogen control strategies prevent hydrogen concentrations from reaching the limits for flame acceleration with margin for the most likely severe accident scenarios. Hydrogen combustion is not predicted to challenge containment integrity. In the event of a station blackout with the additional failure of dc power and igniters, the existing PARs alone are able to prevent the containment from globally reaching flame acceleration limits. Locally, conditions in the loop compartments may approach flame acceleration limits and reduce the margin during rapid hydrogen releases. However the addition of PARs to the containment does not provide any additional benefit to mitigate these rapid hydrogen releases and local concentrations without igniters. Therefore, the existing AP1000 hydrogen control features and strategies consisting of active igniters and PARs provide reasonable assurance of containment

7 survivability and maintain challenges to containment structural integrity from postulated hydrogen combustion to be as low as reasonably practicable during a severe accident. NOMENCLATURE ADS - Automatic Depressurization System AS ME - American Society of Mechanical Engineers CMT - Core Makeup Tank CSNI - Committee on Safety of Nuclear Installations DAS - Diverse Actuation System DDT - Deflagration to Detonation Transition DVI - Direct Vessel Injection EOP - Emergency Operating Procedure FA - Flame Acceleration IRWST - In-containment Refueling Water Storage Tank LOC A - Loss of Coolant Accident NIS - German Company the makes PARs OECD- Organization for Economic Co-Operation and Development PAR - Passive Autocatalytic Recombiner PCS - Passive Containment Cooling System PLS - Plant Control System PRA - Probabilistic Risk Assessment PWR - Pressurized Water Reactor PXS - Passive Emergency Core Cooling System RCS - Reactor Coolant System SBO - Station Blackout SOAR - State-of-the-Art Report REFERENCES 1. AP1000 Design Control Document, Rev. 19, Westinghouse Electric Company, June AP1000 Probabilistic Risk Assessment, Rev. 8, Westinghouse Electric Company, July NEA/CSNI/R(2010)3, OECD/NEA THAI Project: Hydrogen and Fission Product Issues Relevant for Containment Safety Assessment under Severe Accident Conditions, Final Report, June (2010). 4. Breitung, W., Chan, C., Dorofeev, S. et al, OECD Nuclear Energy Agency, State-of-the Art Report on Flame Acceleration and Deflagration-to-Detonation Transient in Nuclear Safety, NEA/CSNI/R(2000)7 August (2000). AP1000 is a trademark or registered trademark in the United States o f Westinghouse Electric Company EEC, its subsidiaries and/or its affiliates. This mark may also be used and/or registered in other countries throughout the world. All rights reserved. Unauthorized use is strictly prohibited. Other names may be trademarks o f their respective owners.

8 AP1000 Hydrogen Control Config 0; 0 Igniters, 0 PARs Containment Pressure AP100G Hydrogen Control Config 0; 0 Igniters, 0 PARs Compartment Gas Temperatures Loop Compartment Upper Compartment Lower Compartment I RW5 T Figure 4

9 AP1000 Hydrogen Control Config 0; 0 Igniters, 0 PARs Loop Compartment Mixture Composition HZ Mole Fraction (in Dry Air) ' Steam Mole Fraction Oxygen Mole Fraction AP100G Hydrogen Control Config 0; 0 Igniters, 0 PARs Upper Compartment Mixture Composition HZ Mole Fraction (in Dry Air) Steam Mole Fraction Oxygen Mole Fraction Figure 6

10 AP1000 Hydrogen Control Config 0; 0 Igniters, 0 PARs IRWST Compartment Mixture Composition HZ Mole Fraction (in Dry Air) ' Steam Mole Fraction Oxygen Mole Fraction AP100G Hydrogen Control Config 0; 0 Igniters, 0 PARs Flame Acceleration Index In Containment Compartments Loop Compartment Upper Compartment Lower Compartment IRWST Figure 8

11 AP1D00 Hydrogen Control Config b 4 -h r Igniters and PARs Flame Acceleration Index in Containment Compartments Loop Compartment Upper Compar tme n t Lower Compartment IRWST API000 Hydrogen Control Config 2: Upper Compt PARs Only Flame Acceleration Index in Containment Compartments Loop Compartment Upper Compartment Lower Compartment IRWST Figure 10

12 AP1000 Hydrogen Control Config 3; Upper Compt PARs + Loop Compt PARs Flame Acceleration Index in Containment Compartments Loop Compartment ~ Upper C omp a r tme n t Lower Compartment AP1000 Hydrogen Control Config 4= 2x Upper Compt PARs Flame Acceleration Index in Containment Compartments Loop Compartment " Upper Compartment Lower Compartment Figure 12