- Email: [email protected]
Identification of important phenomena through the PIRT process for development of sodium fire analysis codes
Nuclear Engineering and Design 353 (2019) 110240
Contents lists available at ScienceDirect
Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes
Identification of important phenomena through the PIRT process for development of sodium fire analysis codes
T
⁎
Mitsuhiro Aoyagi , Akihiro Uchibori, Shin Kikuchi, Takashi Takata, Shuji Ohno, Hiroyuki Ohshima Japan Atomic Energy Agency (JAEA), Oarai, Japan
A R T I C LE I N FO
A B S T R A C T
Keywords: Sodium fire PIRT Verification & validation Assessment matrix
JAEA has developed sodium fire analysis codes, SPHINCS and AQUA-SF, to evaluate the consequences of sodium fire events, which are a key issue in sodium-cooled fast reactor (SFR) plants. This paper describes a PIRT (Phenomena Identification and Ranking Table) process for such events. The present PIRT can be utilized to validate the sodium fire analysis codes. Because a sodium fire event in an SFR plant involves complex phenomena, various figures of merit (FOMs) for importance ranking could exist in the PIRT process. Therefore, the FOMs are specified through a factor analysis. Associated phenomena in a sodium fire event are identified through both element- and sequence-based phenomena analyses. The importance of each associated phenomenon is evaluated by considering the sequence-based analysis of associated phenomena related to the FOMs. Then, a ranking table for the event is established. An assessment matrix of important phenomena and experiments is completed. Sufficiency of experimental data is confirmed for the validation of models corresponding to the identified important phenomena in the sodium fire analysis codes. Additional assessments are discussed specifically for the aerosol module and the CFD module in three-dimensional codes from a perspective of careful validation.
1. Introduction Sodium fires represent a key issue in sodium-cooled fast reactor (SFR) plants, which occur when sodium leaks from a coolant circuit and reacts with oxygen and moisture. A sodium fire may harm components in SFR plants as well as the surrounding environment due to reaction heat and reactive aerosols in combustion products. In the case of severe accidents involving leakage and combustion of the primary sodium coolant and core fuel damage, the behavior of aerosols is essential for source term transportation. To evaluate the consequences of sodium fire events, sodium fire analysis codes have been developed by JAEA (Japan Atomic Energy Agency), such as a zone model code SPHINCS (Yamaguchi et al., 2001) and a multi-dimensional field model code AQUA-SF (Takata et al., 2003). Verification and validation is absolutely necessary in the development of a numerical analysis code to ensure the code’s reliability. Because sodium fire events involve many complex phenomena, identification of key phenomena during such events becomes more appropriate for validating models in the sodium fire analysis codes. The PIRT (Phenomena Identification and Ranking Table) process is an effective
⁎
method to identify key phenomena involved in this type of event (Wilson et al., 1998). In this paper, a phenomena importance ranking table for sodium fire events and an assessment matrix of the important phenomena for the validation of the sodium fire analysis codes are presented. Identification of important phenomena is a significant milestone in a PIRT process. An assessment matrix ensures the sufficiency of existing experimental data or indicates the need for additional experiments to achieve more precise validation. Ultimately, systematical validation can be enabled. As displayed in Table 1, a phenomena importance ranking table consists of figures of merit (FOMs), phenomena associated with an event, and importance ratings of each phenomenon for the FOMs. Therefore, determining these three items is essential for the PIRT process. When the PIRT process is applied to an event involving complex phenomena, such as a sodium fire event in an SFR, various FOMs could exist. Particularly, two FOMs were specified in a PIRT process applied to a sodium fire event by Ohno et al. (2012). In the PIRT process by Olivier et al. (2008), both thermal and aerosol insults were of concern although the FOM was united into the ability to predict both insult. To enhance the precision of PIRT, the specification of appropriate FOMs
Corresponding author. E-mail address: [email protected] (M. Aoyagi).
https://doi.org/10.1016/j.nucengdes.2019.110240 Received 28 March 2019; Received in revised form 29 July 2019; Accepted 30 July 2019 Available online 17 August 2019 0029-5493/ © 2019 Elsevier B.V. All rights reserved.
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
temperature of the liner is specified as an FOM. FOMs relevant to concerns about the components and surrounding environment are also specified in the same manner through factor analysis. Factors and FOMs about component concerns are very similar to those for the building structure. Here, because zone model codes such as SPHINCS only analyze the average atmospheric temperature in an evaluating room, the local temperatures of components are not evaluated. Therefore, atmospheric temperature is substituted for component temperature in accordance with the general fire Probabilistic Risk Assessment (Nowlen et al., 2005). Factors of chemical damage include chemical reactions of sodium oxide and sodium hydrate with component materials. Thus, a relevant FOM is aerosol concentration since the sodium compounds transfer in the form of aerosol to the component surface. Considerable factors regarding the surrounding environment include the chemical effect of sodium compounds and the radiation effect of radioactive materials, such as radioactive sodium in a primary coolant and fission products from damaged core fuel. Because chemical effects on the surrounding environment are evaluated based on the amount of sodium-compound aerosol, aerosol concentration as a consequence of aerosol behavior must be quantified by a sodium fire analysis code. Aerosol behavior plays a major role in the transportation of radioactive materials when the materials are absorbed by the aerosol. Therefore, aerosol concentration is an FOM. As a result of the factor analysis, seven FOMs in total are specified in this PIRT process, as presented in Fig. 1. The FOMs for the concern about building structure are atmospheric pressure, hydrogen concentration, concrete temperature, and steel-liner temperature. The FOMs for the concern about components are atmospheric pressure, hydrogen concentration, component temperature, atmospheric temperature, steel-liner temperature, and aerosol concentration. The sole FOM for the concern about surrounding environment is aerosol concentration in this PIRT process.
Table 1 Examples of a ranking table. Phenomenon
Combustion Heat Transfer Aerosol Transfer
Figure of Merit Atmospheric Pressure
Component Temperature
H (high) L (low) M (medium)
M H L
should be considered carefully, especially in a complex sodium fire event. Additionally, importance evaluation of phenomena for each FOM should be conducted in an objective manner during the PIRT process. For this purpose, the process and results of the importance evaluation must be intelligibly expressed. This paper describes the methodology for assessing the following three items in the ranking table:
• Specification of FOMs, • Identification of associated phenomena, • Importance evaluation of phenomena. FOMs are specified through a factor analysis with attention to safety concerns relevant to sodium fire events. Associated phenomena are identified through element- and sequence-based phenomena analyses, which are commonly employed in PIRT processes. Importance of each associated phenomenon is evaluated by considering the sequence-based analysis of associated phenomena correlated with the FOMs. An assessment matrix is also presented in this paper, which summarizes the sufficiency of experimental data to validate models in the SPHINCS and AQUA-SF codes. Moreover, specific discussions are carried out for validations of aerosol behavior and thermal hydraulics. 2. Ranking table
2.2. Identification of the associated phenomena 2.1. Specification of figures of merit This section describes a methodology for identifying the associated phenomena listed in the first row of the ranking table. Associated phenomena are identified via element- and sequence-based phenomena analyses. The element-based phenomena analysis, for which a system such as an SFR plant is decomposed hierarchically (Zuber et al., 1998), has been widely used in PIRT processes (Wilson et al., 1997; Larson et al., 2007). In the sequence-based phenomena analysis, associated phenomena are identified by analyzing an event progression, as detailed by Kawada et al. (2014). The lists of the associated phenomena identified by the element- and sequence-based analyses are confirmed to be consistent each other.
The scenario evaluated in this PIRT process involves sodium leakage from a coolant circuit in a plant building room that consists of a concrete structure, steel-lined floor, ambient air, and reactor components, such as coolant circuit piping. The combustion of sodium and resulting transportation of heat and mass are examined. The primary purpose of the factor analysis is to specify appropriate FOMs from concerns in sodium fire events. Fig. 1 exhibits a chart of the factor analysis. To specify FOMs, we evaluate concerns about the building structure, components, and surrounding environment that could result from a sodium fire. Here, the concern about the surrounding environment denotes the influence of the event on living bodies in and around the plant. Factors and relevant phenomena are analyzed for each concern as shown in Fig. 1 and explained below. Factors relevant to building structure concerns include mechanical, thermal, and chemical damage. A further factor behind mechanical damage is the pressure rise that results from sodium and hydrogen combustion. Thus, atmospheric pressure becomes one of the FOMs. We discriminated the identified factors based on whether each factor should be evaluated by sodium fire analysis codes or by a separate method. For example, hydrogen combustion is not evaluated by sodium fire analysis codes, but sodium fire analysis codes should quantify hydrogen generation. Thus, hydrogen concentration is also identified as an FOM. Thermal damage is also triggered by both sodium and hydrogen combustion, hence the temperature of structural concrete is listed as an FOM. Considerable chemical damage to the concrete structure may occur from the reactivity between sodium and concrete, which is evaluated beyond sodium fire analysis codes. However, since sodium contacts the concrete only after penetrating the steel liner, the
2.2.1. Element-based phenomena analysis Fig. 2 demonstrates the hierarchical system decomposition. An SFR plant is decomposed into its subsystem, module, component, phase, geometry, field, and transfer process. Then, associated phenomena are identified from the transfer process. In our PIRT process, we focus on a building room in a primary or secondary cooling system. Considerable components in the building room are the ambient air, aerosols of sodium compounds, hydrogen generated by sodium reactions, a steel liner on the floor, the concrete of the building structure, and leaked sodium. Aerosols and hydrogen components, which exist as mixtures with ambient gas, are distinguished from ambient gas because they are only produced once combustion of leaked sodium has occurred. Each component is decomposed into a phase of gas, liquid, or solid. The concrete component has both liquid and solid phases, since the concrete contains water, which is released by heating. Aerosols of both liquid and solid phases exist as fine particles together. Leaked sodium is divided into the droplet or column state in the air and the pool state on the floor. All the elements have their heat energy field, while the gas and 2
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Concern
Related Effect Mechanical Damage
Building Structure / Components
Thermal Damage
Related Phenomenon to be evaluated
Pressure due to Hydrogen Combustion Heat due to Hydrogen Combustion
Hydrogen Combustion *1 *2
Surrounding Environment
Hydrogen Combustion
Pressure due to Sodium Combustion
Atmospheric Pressure
Hydrogen Generation due to Sodium-Moisture Reaction
Hydrogen Concentration
Heat due to Sodium Combustion *1: for Building Structure *2: for Components
Chemical Damage
Figure of Merit (FOM)
by a Separate Method by Sodium Fire Analysis Codes
SodiumConcrete Reaction
Corrosion due to Sodium Compounds
Chemical Effect
Harm from Sodium Compounds
Radiation Effect
Harm from Radioactive materials
Steel Liner Failure Sodium Compounds Transfer to Components
*1
Concrete Temperature
Component *2 Temperature Atmospheric *2 Temperature Steel Liner Temperature
Aerosol Concentration
Aerosol Behavior
Fig. 1. Chart of the factor analysis related to sodium fire events.
liquid phases have also mass and momentum fields. The mass field is considered for the steel liner because the thickness of the liner changes due to corrosion wastage. The mass field of water contained in the concrete is also considered. The transfer processes involved in sodium fire events are categorized into combustion of sodium, including reaction with moisture; heat and mass transfers; and chemical reactions in the atmosphere and with the structure. Then, the transfer processes and
associated phenomena are identified for each category. Here, in a strict sense, spray and pool “combustion” involves heat and mass transfer as well as chemical reactions around the combusting liquid sodium. Distinguished from the combustion phenomena, heat and mass transfer as well as the chemical reactions listed in Fig. 2 are phenomena in the atmosphere and structures. Fig. 3 displays individual correlations for each category extracted from Fig. 2. As a result of the element-based
Fig. 2. Hierarchical system decomposition chart of element-based phenomena analysis. 3
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Fig. 3. Correlation for each category.
into droplets, and then spray combustion occurs involving a sodiummoisture reaction. Since some of the sodium droplets fall to the floor prior to complete combustion, a sodium pool is formed, and pool combustion occurs on the floor. As the result of the spray and pool combustion, heat and reaction products are generated. Reaction heat is transported to the sodium droplets, pool, and surrounding atmosphere via heat convection and radiation. Direct heat transfer to the structure also occurs through heat radiation. Heat conduction, which negligibly contributes to the atmosphere, transfers the
phenomena analysis, fourteen associated phenomena are identified.
2.2.2. Sequence-based phenomena analysis The associated phenomena are also identified through a sequencebased analysis in addition to the element-based analysis. Fig. 4 presents a chart of the whole event sequence considered in the sequence-based analysis. The importance ranking indicated in the figure is explained in the next section. The event sequence is initiated by sodium leakage from a cooling system into a building room. The leaked sodium forms 4
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Fig. 4. Chart of event sequence in a sodium fire event as a result of sequence-based phenomena analysis with the importance of each phenomenon for atmospheric pressure and aerosol concentration. 5
6
L/L H/H L/L
Chemical Reaction
Mass Transfer
Heat Transfer
Pool Combustion
14)
* Concern about 1) Building Structure, 2) Components and 3) Surrounding Environment.
L/L L/L L/L
H/M H/M L/M L/M L/M L/L L/L L/M L/L M/H M/M H/H L/M L/L L/L L/L L/M L/M L/L L/M L/M L/M L/M H/H L/M L/M M/L M/L M/L L/M L/H L/H H/H L/M L/M L/L L/L L/M L/L H/L H/L H/L L/M L/M L/M L/L M/H M/M L/M L/L L/L L/L M/L M/L M/L L/M L/H L/H H/H L/M L/M L/L L/L L/M L/L M/L M/L M/L L/M L/H L/H H/H M/M M/M L/L L/L L/L L/L H/L H/L H/L L/M L/M L/M L/L H/M M/M M/L L/L L/L L/L 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) Spray Combustion
Droplet Generation Spray Combustion Reaction Heat Transfer (spray) Pool Enlargement Pool Combustion Reaction Heat Transfer (pool) Heat Conduction Heat Convection Heat Radiation Mass and Momentum Transfer Gas Species Transfer Aerosol Transfer Atmospheric Chemical Reaction Steel Liner Corrosion Wastage
Aerosol Concentration Hydrogen Concentration Steel Liner Temperature Atmospheric Temperature Component Temperature Concrete Temperature Atmospheric Pressure
1&2 Figure of Merit
Related Concern*
Table 2 Ranking table of important phenomena in a sodium fire event.
1
2
2
1
Importance of the associated phenomena are evaluated in view of their contribution to the maximum value of each FOM. For this purpose, the FOMs are arranged in the event sequence chart illustrated in Fig. 4. The importance of each phenomenon is evaluated by reference to the event sequence chart in addition to engineering judgment. This paper describes the detailed process of importance evaluation for FOMs of atmospheric pressure and aerosol concentration. Importance for both early and late phases is considered. The overall results of the importance ranking are exhibited in Table 2. The importance of each phenomenon for atmospheric pressure is indicated in Fig. 4 with blue characters. In the early phase, spray combustion is dominant over pool combustion. When pool combustion becomes dominant due to pool enlargement and oxygen deficiency around the spray fire region, the event transitions to late phase. In the early phase, droplet generation, spray combustion of sodium, and reaction heat transfer of spray combustion are judged as high importance. In the late phase, the phenomena categorized under pool combustion become dominant over those of spray combustion but are ranked as medium importance because pool combustion is less intense than spray combustion. Reaction heat generated by spray and pool combustion transfers to ambient air in the building room. Then, atmospheric pressure increases due to the thermal expansion of the air. Processes of heat transfer to the air are convection and radiation. Thus, heat convection and radiation are judged as high or medium importance. The heat convection in the early phase is dominant over the maximum atmospheric pressure and is hence judged as high importance. Mass and momentum transfer is ranked with medium importance because this phenomenon promotes convection heat transfer and, in addition, the inter-room convection decreases atmospheric pressure in the building room. The importance of each phenomenon for aerosol concentration is indicated in Fig. 4 with red characters. Because spray combustion in the early phase is dominant over aerosol generation for the same reason as atmospheric pressure, spray combustion of sodium and prior droplet generation in the early phase are judged as high importance. However, because aerosol concentration likely increases for a longer period than atmospheric pressure, associated phenomena in the late phase are considered more important for the maximum aerosol concentration. Therefore, both spray and pool combustion in the late phase are ranked with medium importance. The generated aerosol transfers to the ambient air and then reacts with moisture. The importance of gas species transfer in both phases and atmospheric chemical reaction in the late phase are judged as medium. Atmospheric chemical reaction in the early phase is judged as low importance. The moisture reaction becomes intense after rising aerosol and moisture concentrations. Aerosol transfer in both phases is ranked as high importance because the atmospheric aerosol concentration decreases significantly due to its transfer to solid
Phenomenon
2.3. Importance evaluation of the identified phenomena
Category
1&2
2&3
heat between the sodium pool and the structure as well as in the structure itself. Heat transfer to the steel liner affects its corrosion wastage. Mass and momentum transfers are driven by buoyancy due to the thermal expansion of atmosphere. Then, convection transfers of heat and mass are promoted. The sodium reaction with oxygen generates oxides, while the reaction with moisture generates hydrogen and sodium hydroxide, respectively. The sodium oxides and hydroxide become aerosols, which mix with the atmosphere. Therefore, the convection of mixed gas transports these reaction products. Amounts of the hydrogen and sodium oxides change due to their chemical reaction with oxygen and moisture in the atmosphere, respectively. The amount of aerosol is also affected by agglomeration, deposition to a solid surface, and collection in a pool surface. As a result of the sequence-based phenomena analysis, the same associated phenomena are identified as those in the element-based analysis listed in Fig. 2.
L/L
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
in sodium fire events, as listed in Table 3. The FD series experiment is a basic falling test of a single sodium droplet involving combustion in air (Miyahara et al., 1998; Doda et al., 2003; Doda et al., 2005). The spray combustion model can be assessed separately by the FD experimental data. Run-E1 is a typical spray fire experiment, where sodium was sprayed into a closed cylindrical vessel (Morii et al., 1987). Temperature distribution and pressure change in the vessel were measured in this experiment. These experimental data are available for the validation of spray combustion, heat conduction, and heat convection models. The Run-D1 experiment is a basic pool fire experiment, where sodium was supplied from the bottom of the pool, and the pool area was kept constant during the experiment (Miyahara et al., 1987). An important feature of this experiment is the measurement of heat flux from the combusting pool. Evaluation of heat radiation is enabled by this feature. In the Run-F7 experiment, sodium was supplied from a nozzle above the pool with column state flow. The pool area became enlarged and conceivable sodium leakage events occurred (Futagami et al., 2005; Nishimura et al., 2007). Heat and aerosol transfer resulting from pool fire was investigated in Run-D1, where a multi-cell facility was utilized (Ohno, 2015). An adjoining room was connected to the room used in the Run-D1 pool fire experiment through a horizontal slit. Inter-cell heat and mass transfer were investigated by this experiment. The Run-D4 experiment simulated the structural objects in the Monju plant, such as an air duct and a grating, to investigate the sodium fire incident in 1995 (Uchiyama et al., 2000; Nakagiri et al., 2000). The experimental data is used for comprehensive validation of the codes. It is notable that water release from concrete and the resulting atmospheric chemical reaction of moisture and sodium oxides were observed in this experiment. The six experiments comprise a database to cover all the associated phenomena. Since numerous experiments regarding sodium fires have been conducted, the selected experiments can be replaced by another one. In addition, some experiments can give wider knowledge than the present six experiments can. For example, the IGNA2002 experiment in France was a large-scale test where the sodium leakage rate was measured as 135 kg/s (Saux et al., 1996). On the other hand, the Run-E1 experiment produced a leakage rate of 0.5 kg/s. Although oxygen deficiency occurred in both experiments, local oxygen deficiency might be more noticeable in the IGNA2002 experiment. This experimental data could be utilized for the validation of local phenomena.
or sodium-pool surfaces. Mass and momentum transfer in the late phase is also identified as a phenomenon of high importance due to its promotion of mass transfer. Reaction heat transfer and heat convection in the late phase are judged as medium importance, as they relate to mass and momentum transfer. The importance evaluation for the other five FOMs is conducted in the same manner as previously explained. The overall result of the importance ranking is shown in Table 2. Key points of the importance evaluation are described below. The importance judgment for atmospheric temperature is very similar to that of atmospheric pressure, since most of the associated phenomena are in common between atmospheric pressure and temperature. Phenomena of higher importance for the component, concrete, and steel-liner temperature are heat conduction and those categorized under pool combustion. This is because the heat transfer process to these structural objects is mainly comprised of heat conduction from pool combustion heat. For hydrogen concentration, transfer of gas species including moisture and hydrogen is the most important phenomenon, since hydrogen is generated by the sodium-moisture reaction and then transferred to ambient air. Hydrogen concentration is also influenced by heat conduction to the concrete, which increases moisture concentration, and by aerosol transfer, which decreases it. The importance for hydrogen concentration in the late phase is ranked higher, since hydrogen concentration increases slowly, especially before starting moisture release from the concrete. Here, we should note that a typical SFR plant is considered in this PIRT process. Depending on plant features, such as its design and condition, as well as the scenario being evaluated, there are differences in the associated phenomena and their importance listed in the importance ranking table. However, the process and methodology of the PIRT described in this paper can be applied generally for other PIRT exercises, even if an SFR plant has special features. 3. Assessment matrix of important phenomena and experiments 3.1. Models in the sodium fire analysis codes Table 3 demonstrates an assessment matrix of the models in the analysis codes SPHINCS and AQUA-SF corresponding to the important phenomena and the experimental data to validate the models in the codes. All the important phenomena are modeled in both the SPHINCS and AQUA-SF codes, except for steel-liner corrosion wastage, which can be evaluated without the sodium analysis codes. While the spray and pool combustion models are in common for both the codes, the spatial dimension of atmosphere is the largest difference between the codes. In the SPHINCS code, where a zone model is employed for fast calculation, one computational cell is applied for one room in a reactor building in general. Temperature and concentrations of gas species are homogenized in the atmosphere of the room. Then, inter-room transfer of momentum, mass, and heat is calculated by the flow network model based on the conservation of momentum, mass, and heat energy while considering momentum transfer due to the pressure gradient and buoyancy. On the other hand, the AQUA-SF, which is based on a threedimensional CFD code, AQUA, calculates intra-room transfer of temperature and gas species as well as inter-room transfer. Because the importance ranking table of associated phenomena and sufficiency of experimental knowledge, which is described in the next section, are independent from numerical analysis codes, the assessment matrix for different analysis codes, such as SPHINCS and AQUA-SF, can be developed based on Table 3.
3.3. Detailed assessment for the aerosol behavior model Because aerosol concentration in the ambient air was measured in the spray and pool fire experiments, there are reasonable experimental data on integral behaviors of aerosol transfer. From a more detailed perspective, the aerosol transfer phenomenon (or aerosol behavior) can be subdivided to elemental phenomena, such as agglomeration and deposition, as shown in Fig. 4. Agglomeration does not directly affect aerosol concentration, but deposition rate is sensitive to change in the aerosol diameter due to agglomeration. Therefore, an individual assessment matrix for aerosol behavior is described in this section. Table 4 displays the assessment matrix for aerosol behavior. The key elemental phenomena of aerosol behavior are agglomeration, deposition, convection, and diffusion. Aerosol transfer occurs with the combination of these phenomena. In the SPHINCS and AQUA-SF codes, agglomeration and deposition are modeled by employing an aerosol behavior module based on the ABC-INTG code (Miyahara et al., 1984). Effects of gravity falling, turbulence, and random motion are considered in the aerosol behavior module to simulate detailed behavior. Aerosol transfer due to convection and diffusion is simulated by the flow network model in the SPHINCS code, where just convection is considered, and by the CFD module in the AQUA-SF code, respectively. These models allow aerosol behavior to be simulated during a sodium fire event.
3.2. Sodium fire analysis experiments Numerous experimental studies have been performed for sodium fire investigation. We select six representative experiments performed at JAEA to formulate our validation data involving the key phenomena 7
8
Flow Network Model CFD Agglomeration and Deposition Models Equilibrium Reaction Model
Governing Equation of Heat Conduction Flow Network Model CFD Radiation Model 6-Flux Gas Radiation Model Flow Network Model & Water CFD & Water Release Model Release Model from Concrete from Concrete
(NaFeO type Corrosion Model)*3
Highest rank of each phenomena in the ranking table. Including sodium-moisture reaction. Out of range in the present sodium fire evaluation. Out of range in the present matrix. Negligible small influence. Assessable but indirect measurement.
: : : : : :
H
AQUA-SF
Governing Equations for Droplet Falling and Unburnt Pool Flame Sheet Combustion Model
Nukiyama-Tanasawa Model Spray Combustion Model (NACOM)
*1 *2 *3 *4 *5 *6
14)
M
H H
H H M H
M H H
H H H
Atmospheric Chemical Reaction Steel Liner Corrosion Wastage
Droplet Generation Spray Combustion*2 Reaction Heat Transfer (spray) Pool Enlargement Pool Combustion*2 Reaction Heat Transfer (pool) Heat Conduction Heat Convection Heat Radiation Mass and Momentum Transfer Gas Species Transfer Aerosol Transfer
SPHINCS
Model
13)
11) 12)
7) 8) 9) 10)
4) 5) 6)
1) 2) 3)
Phenomenon & Rank*1
Table 3 Assessment matrix of important phenomena and experiments.
–*4
–*4
✓
–*4
✓
✓ ✓*6 ✓
✓*6 ✓*6
–*4
✓ ✓
–*4 ✓ ✓
–*4 ✓
Constant Pool Area (Run-D1)
Pool Fire
n/a*5 n/a*5
Spray (RunE1)
Single Droplet (FD)
Spray Fire
Experiment
–*4
✓
✓ ✓ ✓*6
✓ ✓ ✓
–*4 n/a*5 n/a*5
Enlarging Pool Area (Run-F7)
–*4
✓ ✓
✓ ✓
–*4
✓
✓ ✓ ✓*6
✓ ✓ ✓
–*4 n/a*5 n/a*5
✓ ✓ ✓*6 ✓
✓ ✓
–*4
Multi-cell Pool (Run-D3)
Integrated Mock-up (Run-D4)
M. Aoyagi, et al.
Nuclear Engineering and Design 353 (2019) 110240
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Table 4 Assessment matrix for aerosol behavior. Phenomenon
Model SPHINCS
Experiment AQUA-SF
Spray Fire Single Droplet (FD)
Agglomeration Gravity falling Turbulence Random motion Deposition Gravity falling Random motion Convection & Diffusion
Pool Fire Spray (Run-E1)
Constant Pool Area (Run-D1)
Enlarging Pool Area (Run-F7)
Multi-cell Pool (Run-D3)
Integrated Mock-up (Run-D4)
✓*
✓*
✓
Gravitational Agglomeration Model Turbulent Agglomeration Model Brownian Agglomeration Model ✓*
✓*
Gravitational Deposition Model Brownian Deposition Model Thermophoretic Deposition Model Flow Network Model CFD (no-diffusion)
✓*
✓
* Integrated evaluation through measurement of atmospheric aerosol concentration.
4. Conclusions
As described before, aerosol concentration was measured in the reference experiments. These experimental data enable integral validation of the aerosol behavior module. To carefully validate the individual models in the aerosol behavior module, it is preferred to utilize measurement data corresponding to each elemental phenomenon, but available data are very limited. In the Run-F7 experiment, a diameter profile of aerosols was measured. The agglomeration model can be validated by assessing change in the diameter. In the Run-D3 experiment, inter-room aerosol transfer was examined, which can contribute to the validation of convection transfer. No experimental data for deposition is available as far as our survey of existing experiments of sodium fires. There exists experimental data on agglomeration, convection, and diffusion, but they are very limited. In addition, the amount of aerosol generation has some degree of uncertainty in sodium fire experiments. For these reasons, if more careful validation is required, additional experiments focusing on aerosol behavior are preferred.
For the validation of the sodium fire analysis codes SPHINCS and AQUA-SF, a ranking table and assessment matrix of important phenomena were completed through the PIRT process of a sodium fire event. Seven FOMs were specified by the factor analysis. Associated phenomena were identified through both element- and sequence-based analyses for adequate phenomena identification. As a result, fourteen phenomena were listed in the ranking table. Importance of the identified phenomena was evaluated by reference to the event sequence chart where the FOMs were corrected with the phenomena, in addition to engineering judgment. Through the assessment matrix, we confirmed the sufficiency of the models in the codes corresponding to the important phenomena. Six sodium fire experiments for validation were listed covering the key phenomena. The aerosol module and the CFD module in the three-dimensional code should be validated considering elemental phenomena for additional careful validation. References
3.4. Specific validation for the three-dimensional code Doda, N., et al., 2003. Falling Sodium Droplet Experiment (FD-2), JNC TN9400 2003-011, Ibaraki (in Japanese). Doda, N., et al., 2005. Falling Sodium Droplet Experiment (FD-3), JNC TN9400 2005-048, Ibaraki (in Japanese). Futagami, S., et al., 2005. Characteristics of Pool Burning in Small-Scale Sodium Leakage, JNC Technical Review No. 27, Ibaraki, pp. 31–40 (in Japanese). Kawada, K., et al., 2014. Development of PIRT (Phenomena Identification and Ranking Table) for SAS-SFR (SAS4A) Validation. Proc. of 22th Int. Conference on Nucl. Eng. (ICONE22, Prague, 2014). ASME, New York ICONE23-1586. Larson, T.K., et al., 2007. Iris small break loca phenomena identification and ranking table (PIRT). Nucl. Eng. Des. 237, 618–626. Miyahara, S., et al., 1984. Development and Validation of ABC-INTG Code, PNC TN943 84-08, Ibaraki. Miyahara, S., et al., 1987. A Large-Scale Test on Sodium Leak and Fire (I)-Sodium Pool Fire Test in Air, Run-D1-, PNC TN9410 87-081, Ibaraki (in Japanese). Miyahara, S., et al., 1998. Experimental Study of Sodium Droplet Burning in Free Fall -Evaluation of Preliminary Test Results-, PNC TN9410 98-065, Ibaraki (in Japanese). Morii, T., et al., 1987. Large-Scale Test on Sodium Leak and Fire (III) (Large-Scale Test of Sodium Spray Fire in Air, Run-E1), PNC TN9410 86-124Tr, Ibaraki (in Japanese). Nakagiri, T., 2000. Evaluation of water transport behavior in Sodium fire experiment – II, JNC TN9400 2000-030, Ibaraki (in Japanese). Nishimura, M., et al., 2007. Characteristics of sodium pool burning behavior in small leakage. Trans. Atomic Energy Soc. Japan 6 (2), 149–160 (in Japanese). Nowlen, S.P., et al., 2005. EPRI/NRC-RES Fire PRA Methodology for Nuclear Power Facilities: Volume 2: Detailed Methodology, EPRI TR-1011989, Palo Alto and NUREG/CR-6850, Rockville. Ohno, S., et al., 2012. Development of PIRT and assessment matrix for verification and validation of sodium fire analysis codes. J. Power Energy Syst. 6 (2), 241–254. Ohno, S., 2015. Evaluation of thermal consequence in sodium fire experiment in two-cell geometry, In: Proc. JSME annual meeting (Sapporo, 2015) JSME, Tokyo (CDROM) S0821105 (in Japanese). Olivier, T.J., et al., 2008. Metal Fire Implications for Advanced Reactors, Part 2: PIRT Results, Sandia Report, SAND2008-6855, Albuquerque. Saux, W., et al., 1996. Experimental study of sodium jet fires with high flow rates. In: IAEA-IWGFR Technical Committee Meeting on Evaluation of Radioactive Materials
The assessment matrix of important phenomena and experiments is common in the zone model code, SPHINCS, and the three-dimensional field model code, AQUA-SF, since the important phenomena are identified without relation to analytical codes. However, detailed assessment is needed for three-dimensional codes, which can analyze detailed thermal-hydraulic behavior via the CFD module. The CFD module should be assessed without a sodium fire for more careful validation. A basic thermal-hydraulic phenomenon during a sodium fire event is buoyancy convection due to sodium combustion heat followed by convective heat and mass transfers. As for fundamental assessment, the thermal cavity problem (Vahl et al., 1983) is an appropriate benchmark to validate the thermal-hydraulics model, which is comprised of equations regarding the coupling of mass, momentum, and energy conservation laws. Another important phenomenon is turbulence. Airflow during sodium fire events is expected to be turbulent, especially in the case of intense sodium combustion. A turbulent plume experiment dominated by buoyant convection with thermal effects (Shabbir et al., 1994) is suitable for validation of the turbulence model in the sodium fire analysis code. Moreover, some sodium fire experiments listed in Table 3 provide useful data related to multi-dimensional effects. For instance, thermo-couple trees are used in the Run-E1 experiment to measure spatial temperature distribution. It is possible to assess multi-dimensional effects by means of these transient data.
9
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
AP600 small break loss-of-coolant accident, main steam line break, and steam generation tube rupture scenarios, NUREG/CR-6541, INEL94-0061, Washington DC. Wilson, G.E., et al., 1998. The role of the PIRT process in experiments, code development and code applications associated with reactor safety analysis. Nucl. Eng. Des. 186, 23–37. Yamaguchi, A., et al., 2001. Numerical methodology to evaluate fast reactor sodium combustion. Nucl. Technol. 136, 315–330. Zuber, N., et al., 1998. An integrated structure and scaling methodology for severe accident technical issue resolution: development of methodology. Nucl. Eng. Des. 186, 1–21.
Release and Sodium Fires in Fast Reactors (IWGFR, Oarai, 1996). IAEA, Vienna, pp. 343–348. Shabbir, A., et al., 1994. Experiments on a round turbulent buoyant plume. J. Fluid Mech. 272, 1–32. Takata, T., et al., 2003. Numerical investigation of multi-dimensional characteristics in sodium combustion. Nucl. Eng. Des. 220, 37–50. Uchiyama, N., et al., 2000. Investigation for the Sodium Leak in Monju-Sodium Fire TestII-, JNC TN9400 2000-090, Ibaraki (in Japanese). Davis, Vahl G.D., et al., 1983. Natural convection in a square cavity: a comparison exercise. Int. J. Numerical Methods Fluids 3, 227–248. Wilson, G.E., et al., 1997. Phenomena identification and ranking tables for Westinghouse
10
Contents lists available at ScienceDirect
Nuclear Engineering and Design journal homepage: www.elsevier.com/locate/nucengdes
Identification of important phenomena through the PIRT process for development of sodium fire analysis codes
T
⁎
Mitsuhiro Aoyagi , Akihiro Uchibori, Shin Kikuchi, Takashi Takata, Shuji Ohno, Hiroyuki Ohshima Japan Atomic Energy Agency (JAEA), Oarai, Japan
A R T I C LE I N FO
A B S T R A C T
Keywords: Sodium fire PIRT Verification & validation Assessment matrix
JAEA has developed sodium fire analysis codes, SPHINCS and AQUA-SF, to evaluate the consequences of sodium fire events, which are a key issue in sodium-cooled fast reactor (SFR) plants. This paper describes a PIRT (Phenomena Identification and Ranking Table) process for such events. The present PIRT can be utilized to validate the sodium fire analysis codes. Because a sodium fire event in an SFR plant involves complex phenomena, various figures of merit (FOMs) for importance ranking could exist in the PIRT process. Therefore, the FOMs are specified through a factor analysis. Associated phenomena in a sodium fire event are identified through both element- and sequence-based phenomena analyses. The importance of each associated phenomenon is evaluated by considering the sequence-based analysis of associated phenomena related to the FOMs. Then, a ranking table for the event is established. An assessment matrix of important phenomena and experiments is completed. Sufficiency of experimental data is confirmed for the validation of models corresponding to the identified important phenomena in the sodium fire analysis codes. Additional assessments are discussed specifically for the aerosol module and the CFD module in three-dimensional codes from a perspective of careful validation.
1. Introduction Sodium fires represent a key issue in sodium-cooled fast reactor (SFR) plants, which occur when sodium leaks from a coolant circuit and reacts with oxygen and moisture. A sodium fire may harm components in SFR plants as well as the surrounding environment due to reaction heat and reactive aerosols in combustion products. In the case of severe accidents involving leakage and combustion of the primary sodium coolant and core fuel damage, the behavior of aerosols is essential for source term transportation. To evaluate the consequences of sodium fire events, sodium fire analysis codes have been developed by JAEA (Japan Atomic Energy Agency), such as a zone model code SPHINCS (Yamaguchi et al., 2001) and a multi-dimensional field model code AQUA-SF (Takata et al., 2003). Verification and validation is absolutely necessary in the development of a numerical analysis code to ensure the code’s reliability. Because sodium fire events involve many complex phenomena, identification of key phenomena during such events becomes more appropriate for validating models in the sodium fire analysis codes. The PIRT (Phenomena Identification and Ranking Table) process is an effective
⁎
method to identify key phenomena involved in this type of event (Wilson et al., 1998). In this paper, a phenomena importance ranking table for sodium fire events and an assessment matrix of the important phenomena for the validation of the sodium fire analysis codes are presented. Identification of important phenomena is a significant milestone in a PIRT process. An assessment matrix ensures the sufficiency of existing experimental data or indicates the need for additional experiments to achieve more precise validation. Ultimately, systematical validation can be enabled. As displayed in Table 1, a phenomena importance ranking table consists of figures of merit (FOMs), phenomena associated with an event, and importance ratings of each phenomenon for the FOMs. Therefore, determining these three items is essential for the PIRT process. When the PIRT process is applied to an event involving complex phenomena, such as a sodium fire event in an SFR, various FOMs could exist. Particularly, two FOMs were specified in a PIRT process applied to a sodium fire event by Ohno et al. (2012). In the PIRT process by Olivier et al. (2008), both thermal and aerosol insults were of concern although the FOM was united into the ability to predict both insult. To enhance the precision of PIRT, the specification of appropriate FOMs
Corresponding author. E-mail address: [email protected] (M. Aoyagi).
https://doi.org/10.1016/j.nucengdes.2019.110240 Received 28 March 2019; Received in revised form 29 July 2019; Accepted 30 July 2019 Available online 17 August 2019 0029-5493/ © 2019 Elsevier B.V. All rights reserved.
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
temperature of the liner is specified as an FOM. FOMs relevant to concerns about the components and surrounding environment are also specified in the same manner through factor analysis. Factors and FOMs about component concerns are very similar to those for the building structure. Here, because zone model codes such as SPHINCS only analyze the average atmospheric temperature in an evaluating room, the local temperatures of components are not evaluated. Therefore, atmospheric temperature is substituted for component temperature in accordance with the general fire Probabilistic Risk Assessment (Nowlen et al., 2005). Factors of chemical damage include chemical reactions of sodium oxide and sodium hydrate with component materials. Thus, a relevant FOM is aerosol concentration since the sodium compounds transfer in the form of aerosol to the component surface. Considerable factors regarding the surrounding environment include the chemical effect of sodium compounds and the radiation effect of radioactive materials, such as radioactive sodium in a primary coolant and fission products from damaged core fuel. Because chemical effects on the surrounding environment are evaluated based on the amount of sodium-compound aerosol, aerosol concentration as a consequence of aerosol behavior must be quantified by a sodium fire analysis code. Aerosol behavior plays a major role in the transportation of radioactive materials when the materials are absorbed by the aerosol. Therefore, aerosol concentration is an FOM. As a result of the factor analysis, seven FOMs in total are specified in this PIRT process, as presented in Fig. 1. The FOMs for the concern about building structure are atmospheric pressure, hydrogen concentration, concrete temperature, and steel-liner temperature. The FOMs for the concern about components are atmospheric pressure, hydrogen concentration, component temperature, atmospheric temperature, steel-liner temperature, and aerosol concentration. The sole FOM for the concern about surrounding environment is aerosol concentration in this PIRT process.
Table 1 Examples of a ranking table. Phenomenon
Combustion Heat Transfer Aerosol Transfer
Figure of Merit Atmospheric Pressure
Component Temperature
H (high) L (low) M (medium)
M H L
should be considered carefully, especially in a complex sodium fire event. Additionally, importance evaluation of phenomena for each FOM should be conducted in an objective manner during the PIRT process. For this purpose, the process and results of the importance evaluation must be intelligibly expressed. This paper describes the methodology for assessing the following three items in the ranking table:
• Specification of FOMs, • Identification of associated phenomena, • Importance evaluation of phenomena. FOMs are specified through a factor analysis with attention to safety concerns relevant to sodium fire events. Associated phenomena are identified through element- and sequence-based phenomena analyses, which are commonly employed in PIRT processes. Importance of each associated phenomenon is evaluated by considering the sequence-based analysis of associated phenomena correlated with the FOMs. An assessment matrix is also presented in this paper, which summarizes the sufficiency of experimental data to validate models in the SPHINCS and AQUA-SF codes. Moreover, specific discussions are carried out for validations of aerosol behavior and thermal hydraulics. 2. Ranking table
2.2. Identification of the associated phenomena 2.1. Specification of figures of merit This section describes a methodology for identifying the associated phenomena listed in the first row of the ranking table. Associated phenomena are identified via element- and sequence-based phenomena analyses. The element-based phenomena analysis, for which a system such as an SFR plant is decomposed hierarchically (Zuber et al., 1998), has been widely used in PIRT processes (Wilson et al., 1997; Larson et al., 2007). In the sequence-based phenomena analysis, associated phenomena are identified by analyzing an event progression, as detailed by Kawada et al. (2014). The lists of the associated phenomena identified by the element- and sequence-based analyses are confirmed to be consistent each other.
The scenario evaluated in this PIRT process involves sodium leakage from a coolant circuit in a plant building room that consists of a concrete structure, steel-lined floor, ambient air, and reactor components, such as coolant circuit piping. The combustion of sodium and resulting transportation of heat and mass are examined. The primary purpose of the factor analysis is to specify appropriate FOMs from concerns in sodium fire events. Fig. 1 exhibits a chart of the factor analysis. To specify FOMs, we evaluate concerns about the building structure, components, and surrounding environment that could result from a sodium fire. Here, the concern about the surrounding environment denotes the influence of the event on living bodies in and around the plant. Factors and relevant phenomena are analyzed for each concern as shown in Fig. 1 and explained below. Factors relevant to building structure concerns include mechanical, thermal, and chemical damage. A further factor behind mechanical damage is the pressure rise that results from sodium and hydrogen combustion. Thus, atmospheric pressure becomes one of the FOMs. We discriminated the identified factors based on whether each factor should be evaluated by sodium fire analysis codes or by a separate method. For example, hydrogen combustion is not evaluated by sodium fire analysis codes, but sodium fire analysis codes should quantify hydrogen generation. Thus, hydrogen concentration is also identified as an FOM. Thermal damage is also triggered by both sodium and hydrogen combustion, hence the temperature of structural concrete is listed as an FOM. Considerable chemical damage to the concrete structure may occur from the reactivity between sodium and concrete, which is evaluated beyond sodium fire analysis codes. However, since sodium contacts the concrete only after penetrating the steel liner, the
2.2.1. Element-based phenomena analysis Fig. 2 demonstrates the hierarchical system decomposition. An SFR plant is decomposed into its subsystem, module, component, phase, geometry, field, and transfer process. Then, associated phenomena are identified from the transfer process. In our PIRT process, we focus on a building room in a primary or secondary cooling system. Considerable components in the building room are the ambient air, aerosols of sodium compounds, hydrogen generated by sodium reactions, a steel liner on the floor, the concrete of the building structure, and leaked sodium. Aerosols and hydrogen components, which exist as mixtures with ambient gas, are distinguished from ambient gas because they are only produced once combustion of leaked sodium has occurred. Each component is decomposed into a phase of gas, liquid, or solid. The concrete component has both liquid and solid phases, since the concrete contains water, which is released by heating. Aerosols of both liquid and solid phases exist as fine particles together. Leaked sodium is divided into the droplet or column state in the air and the pool state on the floor. All the elements have their heat energy field, while the gas and 2
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Concern
Related Effect Mechanical Damage
Building Structure / Components
Thermal Damage
Related Phenomenon to be evaluated
Pressure due to Hydrogen Combustion Heat due to Hydrogen Combustion
Hydrogen Combustion *1 *2
Surrounding Environment
Hydrogen Combustion
Pressure due to Sodium Combustion
Atmospheric Pressure
Hydrogen Generation due to Sodium-Moisture Reaction
Hydrogen Concentration
Heat due to Sodium Combustion *1: for Building Structure *2: for Components
Chemical Damage
Figure of Merit (FOM)
by a Separate Method by Sodium Fire Analysis Codes
SodiumConcrete Reaction
Corrosion due to Sodium Compounds
Chemical Effect
Harm from Sodium Compounds
Radiation Effect
Harm from Radioactive materials
Steel Liner Failure Sodium Compounds Transfer to Components
*1
Concrete Temperature
Component *2 Temperature Atmospheric *2 Temperature Steel Liner Temperature
Aerosol Concentration
Aerosol Behavior
Fig. 1. Chart of the factor analysis related to sodium fire events.
liquid phases have also mass and momentum fields. The mass field is considered for the steel liner because the thickness of the liner changes due to corrosion wastage. The mass field of water contained in the concrete is also considered. The transfer processes involved in sodium fire events are categorized into combustion of sodium, including reaction with moisture; heat and mass transfers; and chemical reactions in the atmosphere and with the structure. Then, the transfer processes and
associated phenomena are identified for each category. Here, in a strict sense, spray and pool “combustion” involves heat and mass transfer as well as chemical reactions around the combusting liquid sodium. Distinguished from the combustion phenomena, heat and mass transfer as well as the chemical reactions listed in Fig. 2 are phenomena in the atmosphere and structures. Fig. 3 displays individual correlations for each category extracted from Fig. 2. As a result of the element-based
Fig. 2. Hierarchical system decomposition chart of element-based phenomena analysis. 3
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Fig. 3. Correlation for each category.
into droplets, and then spray combustion occurs involving a sodiummoisture reaction. Since some of the sodium droplets fall to the floor prior to complete combustion, a sodium pool is formed, and pool combustion occurs on the floor. As the result of the spray and pool combustion, heat and reaction products are generated. Reaction heat is transported to the sodium droplets, pool, and surrounding atmosphere via heat convection and radiation. Direct heat transfer to the structure also occurs through heat radiation. Heat conduction, which negligibly contributes to the atmosphere, transfers the
phenomena analysis, fourteen associated phenomena are identified.
2.2.2. Sequence-based phenomena analysis The associated phenomena are also identified through a sequencebased analysis in addition to the element-based analysis. Fig. 4 presents a chart of the whole event sequence considered in the sequence-based analysis. The importance ranking indicated in the figure is explained in the next section. The event sequence is initiated by sodium leakage from a cooling system into a building room. The leaked sodium forms 4
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Fig. 4. Chart of event sequence in a sodium fire event as a result of sequence-based phenomena analysis with the importance of each phenomenon for atmospheric pressure and aerosol concentration. 5
6
L/L H/H L/L
Chemical Reaction
Mass Transfer
Heat Transfer
Pool Combustion
14)
* Concern about 1) Building Structure, 2) Components and 3) Surrounding Environment.
L/L L/L L/L
H/M H/M L/M L/M L/M L/L L/L L/M L/L M/H M/M H/H L/M L/L L/L L/L L/M L/M L/L L/M L/M L/M L/M H/H L/M L/M M/L M/L M/L L/M L/H L/H H/H L/M L/M L/L L/L L/M L/L H/L H/L H/L L/M L/M L/M L/L M/H M/M L/M L/L L/L L/L M/L M/L M/L L/M L/H L/H H/H L/M L/M L/L L/L L/M L/L M/L M/L M/L L/M L/H L/H H/H M/M M/M L/L L/L L/L L/L H/L H/L H/L L/M L/M L/M L/L H/M M/M M/L L/L L/L L/L 1) 2) 3) 4) 5) 6) 7) 8) 9) 10) 11) 12) 13) Spray Combustion
Droplet Generation Spray Combustion Reaction Heat Transfer (spray) Pool Enlargement Pool Combustion Reaction Heat Transfer (pool) Heat Conduction Heat Convection Heat Radiation Mass and Momentum Transfer Gas Species Transfer Aerosol Transfer Atmospheric Chemical Reaction Steel Liner Corrosion Wastage
Aerosol Concentration Hydrogen Concentration Steel Liner Temperature Atmospheric Temperature Component Temperature Concrete Temperature Atmospheric Pressure
1&2 Figure of Merit
Related Concern*
Table 2 Ranking table of important phenomena in a sodium fire event.
1
2
2
1
Importance of the associated phenomena are evaluated in view of their contribution to the maximum value of each FOM. For this purpose, the FOMs are arranged in the event sequence chart illustrated in Fig. 4. The importance of each phenomenon is evaluated by reference to the event sequence chart in addition to engineering judgment. This paper describes the detailed process of importance evaluation for FOMs of atmospheric pressure and aerosol concentration. Importance for both early and late phases is considered. The overall results of the importance ranking are exhibited in Table 2. The importance of each phenomenon for atmospheric pressure is indicated in Fig. 4 with blue characters. In the early phase, spray combustion is dominant over pool combustion. When pool combustion becomes dominant due to pool enlargement and oxygen deficiency around the spray fire region, the event transitions to late phase. In the early phase, droplet generation, spray combustion of sodium, and reaction heat transfer of spray combustion are judged as high importance. In the late phase, the phenomena categorized under pool combustion become dominant over those of spray combustion but are ranked as medium importance because pool combustion is less intense than spray combustion. Reaction heat generated by spray and pool combustion transfers to ambient air in the building room. Then, atmospheric pressure increases due to the thermal expansion of the air. Processes of heat transfer to the air are convection and radiation. Thus, heat convection and radiation are judged as high or medium importance. The heat convection in the early phase is dominant over the maximum atmospheric pressure and is hence judged as high importance. Mass and momentum transfer is ranked with medium importance because this phenomenon promotes convection heat transfer and, in addition, the inter-room convection decreases atmospheric pressure in the building room. The importance of each phenomenon for aerosol concentration is indicated in Fig. 4 with red characters. Because spray combustion in the early phase is dominant over aerosol generation for the same reason as atmospheric pressure, spray combustion of sodium and prior droplet generation in the early phase are judged as high importance. However, because aerosol concentration likely increases for a longer period than atmospheric pressure, associated phenomena in the late phase are considered more important for the maximum aerosol concentration. Therefore, both spray and pool combustion in the late phase are ranked with medium importance. The generated aerosol transfers to the ambient air and then reacts with moisture. The importance of gas species transfer in both phases and atmospheric chemical reaction in the late phase are judged as medium. Atmospheric chemical reaction in the early phase is judged as low importance. The moisture reaction becomes intense after rising aerosol and moisture concentrations. Aerosol transfer in both phases is ranked as high importance because the atmospheric aerosol concentration decreases significantly due to its transfer to solid
Phenomenon
2.3. Importance evaluation of the identified phenomena
Category
1&2
2&3
heat between the sodium pool and the structure as well as in the structure itself. Heat transfer to the steel liner affects its corrosion wastage. Mass and momentum transfers are driven by buoyancy due to the thermal expansion of atmosphere. Then, convection transfers of heat and mass are promoted. The sodium reaction with oxygen generates oxides, while the reaction with moisture generates hydrogen and sodium hydroxide, respectively. The sodium oxides and hydroxide become aerosols, which mix with the atmosphere. Therefore, the convection of mixed gas transports these reaction products. Amounts of the hydrogen and sodium oxides change due to their chemical reaction with oxygen and moisture in the atmosphere, respectively. The amount of aerosol is also affected by agglomeration, deposition to a solid surface, and collection in a pool surface. As a result of the sequence-based phenomena analysis, the same associated phenomena are identified as those in the element-based analysis listed in Fig. 2.
L/L
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
in sodium fire events, as listed in Table 3. The FD series experiment is a basic falling test of a single sodium droplet involving combustion in air (Miyahara et al., 1998; Doda et al., 2003; Doda et al., 2005). The spray combustion model can be assessed separately by the FD experimental data. Run-E1 is a typical spray fire experiment, where sodium was sprayed into a closed cylindrical vessel (Morii et al., 1987). Temperature distribution and pressure change in the vessel were measured in this experiment. These experimental data are available for the validation of spray combustion, heat conduction, and heat convection models. The Run-D1 experiment is a basic pool fire experiment, where sodium was supplied from the bottom of the pool, and the pool area was kept constant during the experiment (Miyahara et al., 1987). An important feature of this experiment is the measurement of heat flux from the combusting pool. Evaluation of heat radiation is enabled by this feature. In the Run-F7 experiment, sodium was supplied from a nozzle above the pool with column state flow. The pool area became enlarged and conceivable sodium leakage events occurred (Futagami et al., 2005; Nishimura et al., 2007). Heat and aerosol transfer resulting from pool fire was investigated in Run-D1, where a multi-cell facility was utilized (Ohno, 2015). An adjoining room was connected to the room used in the Run-D1 pool fire experiment through a horizontal slit. Inter-cell heat and mass transfer were investigated by this experiment. The Run-D4 experiment simulated the structural objects in the Monju plant, such as an air duct and a grating, to investigate the sodium fire incident in 1995 (Uchiyama et al., 2000; Nakagiri et al., 2000). The experimental data is used for comprehensive validation of the codes. It is notable that water release from concrete and the resulting atmospheric chemical reaction of moisture and sodium oxides were observed in this experiment. The six experiments comprise a database to cover all the associated phenomena. Since numerous experiments regarding sodium fires have been conducted, the selected experiments can be replaced by another one. In addition, some experiments can give wider knowledge than the present six experiments can. For example, the IGNA2002 experiment in France was a large-scale test where the sodium leakage rate was measured as 135 kg/s (Saux et al., 1996). On the other hand, the Run-E1 experiment produced a leakage rate of 0.5 kg/s. Although oxygen deficiency occurred in both experiments, local oxygen deficiency might be more noticeable in the IGNA2002 experiment. This experimental data could be utilized for the validation of local phenomena.
or sodium-pool surfaces. Mass and momentum transfer in the late phase is also identified as a phenomenon of high importance due to its promotion of mass transfer. Reaction heat transfer and heat convection in the late phase are judged as medium importance, as they relate to mass and momentum transfer. The importance evaluation for the other five FOMs is conducted in the same manner as previously explained. The overall result of the importance ranking is shown in Table 2. Key points of the importance evaluation are described below. The importance judgment for atmospheric temperature is very similar to that of atmospheric pressure, since most of the associated phenomena are in common between atmospheric pressure and temperature. Phenomena of higher importance for the component, concrete, and steel-liner temperature are heat conduction and those categorized under pool combustion. This is because the heat transfer process to these structural objects is mainly comprised of heat conduction from pool combustion heat. For hydrogen concentration, transfer of gas species including moisture and hydrogen is the most important phenomenon, since hydrogen is generated by the sodium-moisture reaction and then transferred to ambient air. Hydrogen concentration is also influenced by heat conduction to the concrete, which increases moisture concentration, and by aerosol transfer, which decreases it. The importance for hydrogen concentration in the late phase is ranked higher, since hydrogen concentration increases slowly, especially before starting moisture release from the concrete. Here, we should note that a typical SFR plant is considered in this PIRT process. Depending on plant features, such as its design and condition, as well as the scenario being evaluated, there are differences in the associated phenomena and their importance listed in the importance ranking table. However, the process and methodology of the PIRT described in this paper can be applied generally for other PIRT exercises, even if an SFR plant has special features. 3. Assessment matrix of important phenomena and experiments 3.1. Models in the sodium fire analysis codes Table 3 demonstrates an assessment matrix of the models in the analysis codes SPHINCS and AQUA-SF corresponding to the important phenomena and the experimental data to validate the models in the codes. All the important phenomena are modeled in both the SPHINCS and AQUA-SF codes, except for steel-liner corrosion wastage, which can be evaluated without the sodium analysis codes. While the spray and pool combustion models are in common for both the codes, the spatial dimension of atmosphere is the largest difference between the codes. In the SPHINCS code, where a zone model is employed for fast calculation, one computational cell is applied for one room in a reactor building in general. Temperature and concentrations of gas species are homogenized in the atmosphere of the room. Then, inter-room transfer of momentum, mass, and heat is calculated by the flow network model based on the conservation of momentum, mass, and heat energy while considering momentum transfer due to the pressure gradient and buoyancy. On the other hand, the AQUA-SF, which is based on a threedimensional CFD code, AQUA, calculates intra-room transfer of temperature and gas species as well as inter-room transfer. Because the importance ranking table of associated phenomena and sufficiency of experimental knowledge, which is described in the next section, are independent from numerical analysis codes, the assessment matrix for different analysis codes, such as SPHINCS and AQUA-SF, can be developed based on Table 3.
3.3. Detailed assessment for the aerosol behavior model Because aerosol concentration in the ambient air was measured in the spray and pool fire experiments, there are reasonable experimental data on integral behaviors of aerosol transfer. From a more detailed perspective, the aerosol transfer phenomenon (or aerosol behavior) can be subdivided to elemental phenomena, such as agglomeration and deposition, as shown in Fig. 4. Agglomeration does not directly affect aerosol concentration, but deposition rate is sensitive to change in the aerosol diameter due to agglomeration. Therefore, an individual assessment matrix for aerosol behavior is described in this section. Table 4 displays the assessment matrix for aerosol behavior. The key elemental phenomena of aerosol behavior are agglomeration, deposition, convection, and diffusion. Aerosol transfer occurs with the combination of these phenomena. In the SPHINCS and AQUA-SF codes, agglomeration and deposition are modeled by employing an aerosol behavior module based on the ABC-INTG code (Miyahara et al., 1984). Effects of gravity falling, turbulence, and random motion are considered in the aerosol behavior module to simulate detailed behavior. Aerosol transfer due to convection and diffusion is simulated by the flow network model in the SPHINCS code, where just convection is considered, and by the CFD module in the AQUA-SF code, respectively. These models allow aerosol behavior to be simulated during a sodium fire event.
3.2. Sodium fire analysis experiments Numerous experimental studies have been performed for sodium fire investigation. We select six representative experiments performed at JAEA to formulate our validation data involving the key phenomena 7
8
Flow Network Model CFD Agglomeration and Deposition Models Equilibrium Reaction Model
Governing Equation of Heat Conduction Flow Network Model CFD Radiation Model 6-Flux Gas Radiation Model Flow Network Model & Water CFD & Water Release Model Release Model from Concrete from Concrete
(NaFeO type Corrosion Model)*3
Highest rank of each phenomena in the ranking table. Including sodium-moisture reaction. Out of range in the present sodium fire evaluation. Out of range in the present matrix. Negligible small influence. Assessable but indirect measurement.
: : : : : :
H
AQUA-SF
Governing Equations for Droplet Falling and Unburnt Pool Flame Sheet Combustion Model
Nukiyama-Tanasawa Model Spray Combustion Model (NACOM)
*1 *2 *3 *4 *5 *6
14)
M
H H
H H M H
M H H
H H H
Atmospheric Chemical Reaction Steel Liner Corrosion Wastage
Droplet Generation Spray Combustion*2 Reaction Heat Transfer (spray) Pool Enlargement Pool Combustion*2 Reaction Heat Transfer (pool) Heat Conduction Heat Convection Heat Radiation Mass and Momentum Transfer Gas Species Transfer Aerosol Transfer
SPHINCS
Model
13)
11) 12)
7) 8) 9) 10)
4) 5) 6)
1) 2) 3)
Phenomenon & Rank*1
Table 3 Assessment matrix of important phenomena and experiments.
–*4
–*4
✓
–*4
✓
✓ ✓*6 ✓
✓*6 ✓*6
–*4
✓ ✓
–*4 ✓ ✓
–*4 ✓
Constant Pool Area (Run-D1)
Pool Fire
n/a*5 n/a*5
Spray (RunE1)
Single Droplet (FD)
Spray Fire
Experiment
–*4
✓
✓ ✓ ✓*6
✓ ✓ ✓
–*4 n/a*5 n/a*5
Enlarging Pool Area (Run-F7)
–*4
✓ ✓
✓ ✓
–*4
✓
✓ ✓ ✓*6
✓ ✓ ✓
–*4 n/a*5 n/a*5
✓ ✓ ✓*6 ✓
✓ ✓
–*4
Multi-cell Pool (Run-D3)
Integrated Mock-up (Run-D4)
M. Aoyagi, et al.
Nuclear Engineering and Design 353 (2019) 110240
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
Table 4 Assessment matrix for aerosol behavior. Phenomenon
Model SPHINCS
Experiment AQUA-SF
Spray Fire Single Droplet (FD)
Agglomeration Gravity falling Turbulence Random motion Deposition Gravity falling Random motion Convection & Diffusion
Pool Fire Spray (Run-E1)
Constant Pool Area (Run-D1)
Enlarging Pool Area (Run-F7)
Multi-cell Pool (Run-D3)
Integrated Mock-up (Run-D4)
✓*
✓*
✓
Gravitational Agglomeration Model Turbulent Agglomeration Model Brownian Agglomeration Model ✓*
✓*
Gravitational Deposition Model Brownian Deposition Model Thermophoretic Deposition Model Flow Network Model CFD (no-diffusion)
✓*
✓
* Integrated evaluation through measurement of atmospheric aerosol concentration.
4. Conclusions
As described before, aerosol concentration was measured in the reference experiments. These experimental data enable integral validation of the aerosol behavior module. To carefully validate the individual models in the aerosol behavior module, it is preferred to utilize measurement data corresponding to each elemental phenomenon, but available data are very limited. In the Run-F7 experiment, a diameter profile of aerosols was measured. The agglomeration model can be validated by assessing change in the diameter. In the Run-D3 experiment, inter-room aerosol transfer was examined, which can contribute to the validation of convection transfer. No experimental data for deposition is available as far as our survey of existing experiments of sodium fires. There exists experimental data on agglomeration, convection, and diffusion, but they are very limited. In addition, the amount of aerosol generation has some degree of uncertainty in sodium fire experiments. For these reasons, if more careful validation is required, additional experiments focusing on aerosol behavior are preferred.
For the validation of the sodium fire analysis codes SPHINCS and AQUA-SF, a ranking table and assessment matrix of important phenomena were completed through the PIRT process of a sodium fire event. Seven FOMs were specified by the factor analysis. Associated phenomena were identified through both element- and sequence-based analyses for adequate phenomena identification. As a result, fourteen phenomena were listed in the ranking table. Importance of the identified phenomena was evaluated by reference to the event sequence chart where the FOMs were corrected with the phenomena, in addition to engineering judgment. Through the assessment matrix, we confirmed the sufficiency of the models in the codes corresponding to the important phenomena. Six sodium fire experiments for validation were listed covering the key phenomena. The aerosol module and the CFD module in the three-dimensional code should be validated considering elemental phenomena for additional careful validation. References
3.4. Specific validation for the three-dimensional code Doda, N., et al., 2003. Falling Sodium Droplet Experiment (FD-2), JNC TN9400 2003-011, Ibaraki (in Japanese). Doda, N., et al., 2005. Falling Sodium Droplet Experiment (FD-3), JNC TN9400 2005-048, Ibaraki (in Japanese). Futagami, S., et al., 2005. Characteristics of Pool Burning in Small-Scale Sodium Leakage, JNC Technical Review No. 27, Ibaraki, pp. 31–40 (in Japanese). Kawada, K., et al., 2014. Development of PIRT (Phenomena Identification and Ranking Table) for SAS-SFR (SAS4A) Validation. Proc. of 22th Int. Conference on Nucl. Eng. (ICONE22, Prague, 2014). ASME, New York ICONE23-1586. Larson, T.K., et al., 2007. Iris small break loca phenomena identification and ranking table (PIRT). Nucl. Eng. Des. 237, 618–626. Miyahara, S., et al., 1984. Development and Validation of ABC-INTG Code, PNC TN943 84-08, Ibaraki. Miyahara, S., et al., 1987. A Large-Scale Test on Sodium Leak and Fire (I)-Sodium Pool Fire Test in Air, Run-D1-, PNC TN9410 87-081, Ibaraki (in Japanese). Miyahara, S., et al., 1998. Experimental Study of Sodium Droplet Burning in Free Fall -Evaluation of Preliminary Test Results-, PNC TN9410 98-065, Ibaraki (in Japanese). Morii, T., et al., 1987. Large-Scale Test on Sodium Leak and Fire (III) (Large-Scale Test of Sodium Spray Fire in Air, Run-E1), PNC TN9410 86-124Tr, Ibaraki (in Japanese). Nakagiri, T., 2000. Evaluation of water transport behavior in Sodium fire experiment – II, JNC TN9400 2000-030, Ibaraki (in Japanese). Nishimura, M., et al., 2007. Characteristics of sodium pool burning behavior in small leakage. Trans. Atomic Energy Soc. Japan 6 (2), 149–160 (in Japanese). Nowlen, S.P., et al., 2005. EPRI/NRC-RES Fire PRA Methodology for Nuclear Power Facilities: Volume 2: Detailed Methodology, EPRI TR-1011989, Palo Alto and NUREG/CR-6850, Rockville. Ohno, S., et al., 2012. Development of PIRT and assessment matrix for verification and validation of sodium fire analysis codes. J. Power Energy Syst. 6 (2), 241–254. Ohno, S., 2015. Evaluation of thermal consequence in sodium fire experiment in two-cell geometry, In: Proc. JSME annual meeting (Sapporo, 2015) JSME, Tokyo (CDROM) S0821105 (in Japanese). Olivier, T.J., et al., 2008. Metal Fire Implications for Advanced Reactors, Part 2: PIRT Results, Sandia Report, SAND2008-6855, Albuquerque. Saux, W., et al., 1996. Experimental study of sodium jet fires with high flow rates. In: IAEA-IWGFR Technical Committee Meeting on Evaluation of Radioactive Materials
The assessment matrix of important phenomena and experiments is common in the zone model code, SPHINCS, and the three-dimensional field model code, AQUA-SF, since the important phenomena are identified without relation to analytical codes. However, detailed assessment is needed for three-dimensional codes, which can analyze detailed thermal-hydraulic behavior via the CFD module. The CFD module should be assessed without a sodium fire for more careful validation. A basic thermal-hydraulic phenomenon during a sodium fire event is buoyancy convection due to sodium combustion heat followed by convective heat and mass transfers. As for fundamental assessment, the thermal cavity problem (Vahl et al., 1983) is an appropriate benchmark to validate the thermal-hydraulics model, which is comprised of equations regarding the coupling of mass, momentum, and energy conservation laws. Another important phenomenon is turbulence. Airflow during sodium fire events is expected to be turbulent, especially in the case of intense sodium combustion. A turbulent plume experiment dominated by buoyant convection with thermal effects (Shabbir et al., 1994) is suitable for validation of the turbulence model in the sodium fire analysis code. Moreover, some sodium fire experiments listed in Table 3 provide useful data related to multi-dimensional effects. For instance, thermo-couple trees are used in the Run-E1 experiment to measure spatial temperature distribution. It is possible to assess multi-dimensional effects by means of these transient data.
9
Nuclear Engineering and Design 353 (2019) 110240
M. Aoyagi, et al.
AP600 small break loss-of-coolant accident, main steam line break, and steam generation tube rupture scenarios, NUREG/CR-6541, INEL94-0061, Washington DC. Wilson, G.E., et al., 1998. The role of the PIRT process in experiments, code development and code applications associated with reactor safety analysis. Nucl. Eng. Des. 186, 23–37. Yamaguchi, A., et al., 2001. Numerical methodology to evaluate fast reactor sodium combustion. Nucl. Technol. 136, 315–330. Zuber, N., et al., 1998. An integrated structure and scaling methodology for severe accident technical issue resolution: development of methodology. Nucl. Eng. Des. 186, 1–21.
Release and Sodium Fires in Fast Reactors (IWGFR, Oarai, 1996). IAEA, Vienna, pp. 343–348. Shabbir, A., et al., 1994. Experiments on a round turbulent buoyant plume. J. Fluid Mech. 272, 1–32. Takata, T., et al., 2003. Numerical investigation of multi-dimensional characteristics in sodium combustion. Nucl. Eng. Des. 220, 37–50. Uchiyama, N., et al., 2000. Investigation for the Sodium Leak in Monju-Sodium Fire TestII-, JNC TN9400 2000-090, Ibaraki (in Japanese). Davis, Vahl G.D., et al., 1983. Natural convection in a square cavity: a comparison exercise. Int. J. Numerical Methods Fluids 3, 227–248. Wilson, G.E., et al., 1997. Phenomena identification and ranking tables for Westinghouse
10