hydrogen risk - Core

N u c l E n g T e c h n o l 4 7 ( 2 0 1 5 ) 3 3 e4 6

Available online at www.sciencedirect.com

ScienceDirect journal homepage: http://www.journals.elsevier.com/nuclearengineering-and-technology/

Invited Article

RESEARCH EFFORTS FOR THE RESOLUTION OF HYDROGEN RISK SEONG-WAN HONG*, JONGTAE KIM, HYUNG-SEOK KANG, YOUNG-SU NA, and JINHO SONG Division of Severe Accident & PHWR Safety Research, Korea Atomic Energy Research Institute, 989-111, Daedeok-daero, Yuseong-gu, Daejon, 305-353, South Korea

article info

abstract

Article history:

During the past 10 years, the Korea Atomic Energy Research Institute (KAERI) has per-

Received 12 December 2014

formed a study to control hydrogen gas in the containment of the nuclear power plants.

Received in revised form

Before the Fukushima accident, analytical activities for gas distribution analysis in ex-

22 December 2014

periments and plants were primarily conducted using a multidimensional code: the

Accepted 22 December 2014

GASFLOW. After the Fukushima accident, the COM3D code, which can simulate a multi-

Available online 21 January 2015

dimensional hydrogen explosion, was introduced in 2013 to complete the multidimensional hydrogen analysis system. The code validation efforts of the multidimensional

Keywords:

codes of the GASFLOW and the COM3D have continued to increase confidence in the use of

Hydrogen issue Multidimensional analysis Passive autocatalytic recombiner Severe accident

codes using several international experimental data. The OpenFOAM has been preliminarily evaluated for APR1400 containment, based on experience from coded validation and the analysis of hydrogen distribution and explosion using the multidimensional codes, the GASFLOW and the COM3D. Hydrogen safety in nuclear power has become a much more important issue after the Fukushima event in which hydrogen explosions occurred. The KAERI is preparing a large-scale test that can be used to validate the performance of domestic passive autocatalytic recombiners (PARs) and can provide data for the validation of the severe accident code being developed in Korea. Copyright © 2015, Published by Elsevier Korea LLC on behalf of Korean Nuclear Society.

1.

Introduction

After the Three Mile Island Reactor 2 (TMI-2) accident (near Middletown, PA, USA) in 1979, a substantial amount of research has been performed to control hydrogen gas in the containment. In European countries, an important research result to control hydrogen gas is the implementation of

passive autocatalytic recombiners for pressurized water reactors (PWRs) with a large dry containment. However, the boiling water reactor (BWR), which has a relatively very small containment volume in comparison to a PWR was adapted as an inert concept to control hydrogen gas by injecting nitrogen gas into a small containment. This concept to control hydrogen gas in a BWR was an action item after the TMI-2

* Corresponding author. E-mail address: [email protected] (S.-W. Hong). This is an Open Access article distributed under the terms of the Creative Commons Attribution Non-Commercial License (http:// creativecommons.org/licenses/by-nc/3.0) which permits unrestricted non-commercial use, distribution, and reproduction in any medium, provided the original work is properly cited. http://dx.doi.org/10.1016/j.net.2014.12.003 1738-5733/Copyright © 2015, Published by Elsevier Korea LLC on behalf of Korean Nuclear Society.

34


Fig. 1 e Hydrogen and steam source in loss of coolant accident.

accident [1]. However, hydrogen explosions unfortunately occurred in the reactor building of the BWRs of the Fukushima plants [2]. This explosion caused the large release of the fission product into the environment. After the large explosions in the Fukushima accident, a precise prediction of gas distribution under various containment thermal hydraulic conditions has become an important issue. The explosion behavior can be estimated from precise hydrogen distribution in the containment. However, it is not easy to precisely predict the hydrogen distribution in a containment where many structures and components exist. In addition, hydrogen experts are interested in hydrogen explosion behavior based on the hydrogen distribution because the reactor building was destroyed by hydrogen explosions in the Fukushima plants. Two international activities commenced after the Fukushima accident. The Organization for Economic Cooperation and Development/Nuclear Energy Agency (OECD/ NEA) Hydrogen Mitigation Experiments for Reactor Safety (HYMERS) project [3] was launched to precisely predict gas distribution in the containment. The main objective of this project is to enhance the reliability of advanced computational analysis devoted to containment thermal hydraulics. It is necessary to continue the validation of models of the code to improve confidence in the plant analysis results and the predictive ability of the models of the code; to have experiments with a well-instrumented spatial and temporal grid in line with modeling issues and with real control of the boundary conditions; to reach a significant level of accuracy on the gas mixture composition during the mixing process because the consequences of a combustion [e.g., 11e13 volume percent (vol%) of hydrogen in air] can be totally different; and to cover the broad spectrum of phenomena during a severe accident in a light water reactor (LWR) such as geometric effects and the effects of safety systems. Another international activity to remove the hydrogen threat is the performance test of PARs, which can simulate an accident without electricity under severe accidents conditions, thus, the OECD/NEA Thermal-hydraulic, Hydrogen, Aerosol and Iodine (THAI) Project was performed [4]. The frame of the first phase of the OECDeTHAI project was

aimed at providing nonexistent or inaccessible experimental data for the validation of developed analysis methods and codes to predict hydrogen distribution, combustion behavior, and PAR behavior under representative reactor conditions. Significant progress has been achieved by demonstrating the transferability of helium to hydrogen distribution behavior and by providing comprehensive data for the validation of computational fluid dynamics (CFD) and lumped parameter simulation codes. With regard to code validation, phenomena such as the formation and dissolution of the stratification of different gases remain partly open. In spite of significant improvements in available computer codes, deficiencies still exist in modeling certain phenomena. Uncertainties continue to appear in the modeling of deflagrations and the modeling of the performance of PAR under special conditions such as an under spray or low oxygen supply. The general performance of the PAR is satisfactory, but some adverse effects occur with iodine present in the gases. An effort to validate the performance of PARs has continued. The original objectives of the OECD-THAI2 [5], which was a follow-up project to the OECD-THAI1, will address open questions concerning the behavior of: (1) graphite dust transport in a generic hightemperature gas-cooled reactor (HTGR) geometry; (2) release of gaseous iodine from a flashing jet and iodine deposition on aerosol particles; and (3) hydrogen combustion during spray operation and PAR operation under the condition of extremely low oxygen content. However, after the Fukushima accident, the test cases have been changed toward PAR performance tests with spray operation.

2.

Outstanding issues

The integral analysis codes, the so-called lumped codes such as the Modular Accident Analysis Program (MAAP) [6] and the Methods for Estimation of Leakages and Consequences of Releases (MELCOR) [7], were primarily used in the past and are now the main tools used to evaluate hydrogen safety in nuclear plants. These lumped codes are capable of easily analyzing an accident sequence for a long time with regard to the computational calculation time. The


multidimensional analysis codes in parallel have recently been used to compensate the limitation of lumped codes that cannot trace local hydrogen concentration in the containment. The necessity of multidimensional codes for hydrogen safety is introduced in this paper by reviewing the previous studies. KIT (Karlsruhe Institute of Technology) compared the gas distribution in the containment of the Western generic PWR reactor type using the MELCOR, lumped parameter code, and the GASFLOW [8], a multidimensional analysis code. The GASFLOW shows probable explosive mixtures in a specific area, whereas the MELCOR shows no explosive mixtures [9]. The MELCOR code, which is a lumped code, cannot take local hydrogen concentration, which affects the hydrogen combustion and the evaluation of PAR performance. It is accordingly believed that the multidimensional analysis has to be accompanied in parallel with the lumped parameter analysis for the evaluation of hydrogen safety. Hydrogen distribution analysis for the APR1400 using the GASFLOW code was performed for the loss of coolant accident (LOCA) sequence at the Korea Atomic Energy Research Institute (KAERI) [10]. Fig. 1 shows the hydrogen release rate into the containment for the LOCA sequence of the ARP1400, and Fig. 2 shows the amount of hydrogen removed by PARs with the passage of time. As Fig. 3 shows, if the remaining hydrogen is not removed by the PARs at a specific time and it accumulates in the local area and over the flammability limit, it may cause a hydrogen explosion from an ignition source by chance before the PARs can remove sufficient hydrogen. An accurate hydrogen distribution analysis in the containment using the multidimensional code is accordingly important to predict hydrogen explosion behavior and to effectively remove the hydrogen in the containment by the PAR through selecting the proper location for the PAR. JNES (Japan Nuclear Energy Safety Organization) performed a hydrogen explosion analysis of the Fukushima accident [11]. In Unit 1, it was assumed that the released flow rate of hydrogen gas was 100 kg/h for 4 hours. The released hydrogen gas accumulated evenly in the top of the floor (5F) with a high concentration of approximately 20%. In Unit 3,

they assumed that the released flow rate of the hydrogen gas was 200 kg/h for 5 hours and the released hydrogen gas accumulated evenly in the whole reactor building with a high concentration of approximately 16e17%. They insisted that hydrogen explosion simulation under the assumption of different hydrogen distributions in two plants was well performed by showing similar results to those of the accident. From the JNES analysis result, different hydrogen distributions may affect different hydrogen explosion behaviors. Hydrogen distribution in a reactor building in the Fukushima plants remains unknown and will not be predicted in the future because a hydrogen explosion occurred in the past. In summary, the use of a multidimensional code to predict the hydrogen distribution in a containment is important: (1) to support the calculation results from the lumped code; (2) to provide the best location of PARs which are installed in various regions of the containment; (3) to evaluate the regulatory requirement of hydrogen control; and (4) to provide information on hydrogen distribution to hydrogen explosion code. Detailed analysis on hydrogen distribution and explosion behavior using a multidimensional code has become more important because a hydrogen explosion occurred in the Fukushima plants. Performance tests to the PAR developed in other countries have meanwhile been performed in a large-scale test vessel because PAR operation induces a natural circulation flow. Even though the structure of the PAR is different, the general features of PAR performance were found using a 60-m3 vessel in a THAI facility. The general performance from the THAI experimental facility was known [12]. The performance of PAR increases with increasing initial pressure. The steam delays the start-up time of the PAR. Ignition of hydrogen by PARs is

800

200

Recombined amount

180

700

160

Source

140

Inventory

600 500

120

400

100 80

H 2 (kg)

Recombination rate (g/s)

35

300

60 200 40

103 kg

20 0

0

1000

2000

3000

4000

5000

Time (s)

6000

7000

8000

100 0 9000

Fig. 2 e Hydrogen removal rate by passive autocatalytic recombiners.

Fig. 3 e Hydrogen clouds at 10% and 5% at 6,200 seconds.

36


possible. The ignition source is suspected to be the high surface temperature of the catalyst at high hydrogen concentrations. A low oxygen concentration decreases the performance of PAR. The SNL (Sandia National Laboratories) performed PAR performance in a Surtsey vessel of 100 m3 [13]. Most general features are similar to the THAI test results. One important finding in the SNL test was that the PAR appeared to generate a convective flow loop in the Surtsey vessel from the PAR outlet to the dome, down the Surtsey wall until reaching a height near the PAR inlet, and then back to the PAR inlet. This was indicated by relatively flat and equal temperatures from the thermocouples located at elevations below the PAR. Because the PAR only consumes hydrogen in a portion of the total vessel volume, a depletion rate calculation that assumes a well-mixed condition in the code overpredicts consumption. In addition, counterpart experiments determined that the placement of a PAR near a wall yielded depletion rates smaller than those obtained with a PAR placed in the middle of the facility. The PARs in a Korean plant are located in the near wall of the structure, and the performance test of PARs developed in Korea are performed in a very relatively scale vessel [14]. It is not possible to see the overall performance because the small vessel height is insufficient to induce a natural circulation flow. Therefore, the performance test under various conditions of the Korean PAR need to be performed.

3. Research activities at the KAERI for hydrogen mitigation 3.1.

Analytical activities

It is commonly understood that having accurate numerical solutions of hydrogen behaviors and explosions occurring in a containment during a severe accident condition is very difficult. This is because many phenomena such as buoyant jet, steam condensation, heat transfer, hydrogen recombination, and combustion occur in a coupled manner, and because large discrepancies exist in length and time scales between accident progression in a real containment and in local phenomena. Numerical predictions of some local phenomena nevertheless rely on simple correlations instead of universal models. Sometimes reducing temporal or spatial resolution is required to obtain an approximate solution of a numerical analysis from a currently available computer program. However, the prediction of hydrogen distribution and explosion by experts with physical and numerical understanding using a code with well benchmarked numerical and physical models will increase the confidence of the results of plant analysis. The objective of a hydrogen analysis at KAERI is to analyze and evaluate the hydrogen safety of Korean nuclear plants during a severe accident. It is very important to have confidence in the computed results by the code, the validation of an analysis code, and evaluation of models implemented in the code. Some efforts have been made to validate the code in parallel with the plant analysis. From the analyses of a separate effect test such as the PANDA [15], it is possible to have physical insight to hydrogen

behaviors in compartments and choose optimal numerical schemes to obtain confident solutions from the GASFLOW code. Fig. 4 shows PANDA nodalization and Fig. 5 shows the simulation results for test 9bis in which the injected superheated vapor was condensed on the cold wall of the test vessel. The analysis results well agreed with the experimental results. The volume flow rates at the vent rapidly reduced over the inlet flow to some plateau values, which reflect the cooling of the injected steam by mixing with air. A further decay to a second plateau occurs in the test after the onset of condensation at approximately 3,000 seconds. This plateau is controlled by the constant pressure of 1.3 bar at the vent and is well predicted in the GASFLOW and Generation of Thermal Hydraulic Information in Containments (GOTHIC) codes. For the hydrogen distribution analysis, KAERI has performed code benchmark analyses and plant application for the past 10 years [16]. Some interest in multidimensional analysis of hydrogen explosion has recently increased to evaluate flame acceleration and obtain a mechanical load on a containment wall during a severe accident. Table 1 shows some specific features of the multidimensional analysis codes that are applied for a nuclear power plant [17e19]. To select the multidimensional hydrogen explosion code among the available codes, KAERI first considered the applicability of the code to a nuclear power plant and the efficiency of connection with the analysis code of hydrogen distribution, GASFLOW. On considering these factors, the COM3D code [20] was ultimately selected for the hydrogen explosion analysis. Fig. 6 shows the analysis result using the ANSYS CFX-13 [21] in comparison with the COM3D result for the ENACCEF test results [22] with a hydrogen concentration of 13% and an obstacle blockage ratio of 0.63. In the COM3D analysis, a three-dimensional grid model with a quarter symmetric condition was generated. Approximately 439,676 hexahedral cells with a cell length of 7 mm were generated in the grid model. The KYLCOMþ model with the Schmidt turbulent flame speed model was used to simulate hydrogen flame propagation in the COM3D calculation [20]. A turbulent flow

Fig. 4 e PANDA nodalization.

37

CRECOM [19]

BVM; COM3D; CRECOM; DDT; EBU; EDM; ENACCEF; eq., equation; H2, hydrogen; JANAF; KYLCOM; L, large scale; N-S; PAR, passive autocatalytic recombiner; REACFLOW; RUT; S, small scale; THAI, Thermal-hydraulic, Hydrogen, Aerosol and Iodine; TONUS CFD.

Fixed/adaptive grid (Tetra mesh) N-S eq. (Compressible flow) Total enthalpy CP, JANAF eq. r, Ideal gas law Standard k-e Unclear EDM (1) Forced chemical reaction (2) Possible (3) Variable for H2% (4) No Driver (S, 13%) ENACCEF (S,10e14%) RUT (L, 10%)

REACFLOW [18] COM3D [20]

Fixed Cartesian grid (Hexa mesh) N-S eq. (Compressible flow) Total enthalpy CP, JANAF eq. r, Ideal gas law Standard k-e Yes EDM/KYLCOM (1) Forced chemical reaction (2) No (3) Unclear (4) No Driver (S,13%) ENACCEF (S,10e14%) THAI (L, 10%) RUT (L, 10%) Fixed Cartesian grid (Hexa mesh) N-S eq. (low Ma) Euler eq. (high Ma) Temp. (low Ma) CP, JANAF eq. r, Ideal gas law Standard k-e Yes Arrhenius/EBU (1) Combusted area (2) Unclear (3) Variable for H2% (4) Yes Driver (S,13%) ENACCEF (S,10e14%) RUT (L, 10%)

TONUS CFD [17] Combustion Code

Table 1 e Features of multidimensional analysis codes.

was modeled using the standard k-e turbulent model. A wall condition with a constant temperature of 298 K was applied on the outer surface of the grid model. The time step size for the COM3D calculations was automatically controlled to assure a CouranteFriedrichseLewy (CFL) number below 0.9. The proposed COM3D model accurately predicted the measured flame front time-of-arrivals (TOAs) with an error range of approximately 10% over the entire range (Fig. 6A). The flame position in the CFD result was defined at the time the gas temperature increased to approximately 850 K at locations PM1 to PM16. The COM3D results overpredicted the peak pressure of the test results with an error range of approximately 20%, whereas the CFX accurately predicts it with an error range of approximately 10% (Fig. 6B). In the CFX calculation, the turbulent flame closure model constant changed from 2.0 to 5.0 after the flame arrived at the top of the dome region to prevent the retardation of the flame propagation rising from a local pressure buildup at an edge of the dome region in the grid model [23]. However, the computational time to complete the hydrogen flame propagation in the grid model was approximately 10 times faster than that of the CFX results, even though the number of mesh between the grid models is different. Therefore, the COM3D is a very efficient code for predicting the pressure behavior resulting from hydrogen flame acceleration. The code validation efforts will be continued by joining the OECD/THAI-2 and OECD-HYMERS programs, and will finally be applied to Korean nuclear power plants. To determine the applicability of the COM3D in a real plan, a preliminary COM3D analysis was performed to analyze the hydrogen flame propagation in the APR1400 containment using the calculated hydrogen distribution by the GASFLOW code for a station blackout accident [24]. The grid model representing the APR1400 containment (Fig. 7) was also transferred from the GASFLOW to the COM3D by reducing the cell length to approximately 0.25 m. Therefore, 5,931,678 hexahedral cells were generated for the hydrogen combustion in the grid model. The cell length was determined to resolve accurately the pressure

Grid model Hydrodynamic solver Energy equation Properties (r, CP) Turbulent model Multicomponent Combustion model (1) Ignition (2) DDT simulation (3) Model Const. (4) PAR Validation test facility (scale, H2%)

Fig. 5 e Comparison of the results by GASFLOW and by experiment.

Fixed Cartesian grid (Hexa mesh) Unclear Total enthalpy CP, Constant r, Ideal gas law Unclear Unclear BVM/Burning rate model (1) Unclear (2) Unclear (3) Unclear (4) No Explosion Channel (S,?) RUT (L, 10%)


38


wave propagation generated from the combusted region [25] and to model the important structures in the containment. The KYLCOMþ model with the Kawanabe turbulent flame speed model [20] was used to simulate the hydrogen flame propagation in the COM3D calculation. The turbulent flow was modeled by the standard k-e turbulent model. A time step size for the COM3D calculations was automatically controlled to assure a CFL number below 0.9. The wall condition with a constant temperature of 298 K was applied to the outer surface of the grid models. The ignition point was assumed around the hydrogen release location in the steam generator (SG) compartment, as Fig. 7B shows. The ignition process was modeled by using a hot spot region (with a radius of 0.5 m) where the hydrogen-air chemical reaction takes place. As Figs. 8 and 9 demonstrate, the COM3D results showed that the hydrogen flame was propagated to approximately 18 m along the vertical direction and 10 m along the radial direction in 0.155 seconds after the start of the ignition.

This flame acceleration induced the pressure to increase to approximately 3.99 105 Pa in the neighboring structure of the SG compartment. A detailed analysis on the calculated temperature and pressure results will be conducted after the hydrogen flame arrives at the top of the containment. Aside from using the code introduced from overseas, the KAERI has been developing a combustion module using an open-source CFD code, called OpenFOAM [26], and validating it by applying it to experiments. Figs. 10 and 11 show the simulation results of the ENACCEF experiments. The flame speed and length were reasonably predicted. In the figures, the quantitatively measured flame speeds were compared for uniform hydrogen distribution. The prediction seems to be reasonable, except for overprediction in the dome region. The simulations by other research group also show similar trends. It appears that experimental uncertainties may exist in the larger volume regions (i.e., the dome). In this situation, the maximum flame speed reaches approximately

Fig. 6 e Predicted results by COM3D and by CFX-13. The images show (A) the flame front time-of-arrival (TOA) from PM1 to PM16 and (B) the pressure behaviors at PCB2.

Fig. 7 e Isosurface of hydrogen concentration (10%) at 47,400 seconds after the start of the station blackout accident, described by Kim et al. [24], using (A) GASFLOW and (B) COM3D.

Fig. 8 e Isotemperature of 1,000 K as time progresses after the start of the ignition. (A) 0.106s, (B) 0.123s, (C) 0.155s.

Fig. 9 e Calculated pressure distribution at 0.155 seconds after the start of the ignition.

40


Fig. 10 e Flame propagation in the uniform-distribution case of ENACCEF.

600 m/s, which indicates that turbulence generated by obstacles installed in the pipe rapidly accelerates flame propagation. In an effort to evaluate mechanistically hydrogen safety during a severe accident, multidimensional codes for the distribution and explosion analyses codes were applied to the APR1400 containment. During a high pressure accident such as a station blackout (SBO), the release location of hydrogen

Fig 11 e Comparison of the flame speed in the uniformdistribution case of ENACCEF.

generated in a reactor pressure vessel is specific for a nuclear power plant. For the APR1400, it is released in the incontainment refueling water storage tank (IRWST), which is a closed large annular compartment, except for the vent holes on the ceiling. During a GASFLOW simulation for an SBO accident, Kim et al. [24] found that a detonable hydrogen mixture cloud could be developed in the IRWST. In one study [27], the effect of the IRWST vents on hydrogen flame acceleration was studied mechanistically using the OpenFOAM code. In a low hydrogen concentration of up to approximately 20 vol%, the vents of the IRWST tend to decelerate considerably the propagation of the hydrogen flame, compared to the propagation without the vents at the same hydrogen concentration. On the contrary, the hydrogen flame can be accelerated, even with the vents, when the concentration of the hydrogen is above 25 vol%. Shin-Ulchin 1 and 2, which are construction plants of an APR 1400, have adopted a new hydrogen mitigation strategy in which hydrogen is released into a SG compartment by actively turning a three-way valve when a severe accident management (SAM) for the plant is initiated. A parametric study [28] was conducted to mechanically evaluate the flame acceleration characteristics, depending on the hydrogen release location. In that study, a coupled analysis method of the GASFLOW code for hydrogen distribution and the OpenFOAM code for hydrogen combustion was applied to assess the possibility of hydrogen flame acceleration during a SBO accident in the plants. A hydrogen distribution and other thermal hydraulic data from a GASFLOW simulation are transferred to an OpenFOAM mesh for a combustion analysis using the procedure shown in Fig. 12. Some utility programs were needed to transfer data because of different coordinate systems and mesh resolutions between the GASFLOW and OpenFOAM codes. Two instances of a lower hydrogen release near a SG pedestal and an upper hydrogen release near a pressurizer bottom were simulated and compared. Fig. 13 shows the mapped hydrogen distributions for a combustion and flame acceleration analysis at the moment of a maximum release rate of hydrogen. The study showed that the flame acceleration is much stronger when

Fig. 12 e The procedure to transfer data from GASFLOW to OpenFOAM.


41

Fig. 13 e Hydrogen (H2) distribution obtained by GASFLOW simulations at the moment of maximum release rate of hydrogen in (A) the lower release condition and (B) the upper release condition.

Fig. 14 e Hydrogen flame fronts using a temperature isosurface of 1,000 K and pressure distributions on the containment wall in (A) the lower release condition and (B) the upper release condition.

42


Table 2 e Design parameters of the APR1400 containment. Parameters

Value

Height Diameter Free air volume Zirconium mass Number of PAR

75.7 m 45.7 m 88,575 m3 24,307 kg 30

PAR, passive autocatalytic recombiner.

Table 3 e Calculated Ra number, based on the safety analysis results [25]. Variable Distance to dome Pressure Temperature H2 molar fraction Density Ra

IRWST

SG compartment

29.5 m 1.45 bar 1,200 K 0.14 ~0.246 kg/m3 ~1.83 1011

41.6 m 1.45 bar 600 K 0.08 ~0.879 kg/m3 ~1.32 1011

H2, hydrogen; IRWST, in-containment refueling water storage tank; Ra, Rayleigh; SG, steam generator.

the release point is located near the pedestal of the SG, compared to the pressurizer bottom release, as Fig. 14 shows.

3.2.

Experimental activities

3.2.1.

Scaling analysis

To determine the vessel size of a scaled-down test facility simulating the APR1400 containment, scaling analysis was conducted on the basis of a severe accident scaling methodology [29]. With regard to the hydrogen distribution in the containment during a severe accident, the system scaling approach (top-down) may be preserving the ratio of the height to the diameter of the containment, the ratio of the produced hydrogen mass from the reactor core to an air mass in the containment, the location of hydrogen release, and the locations of the PARs.

Fig. 15 e The Rayleigh number, based on the test facility height.

Major design parameters of the APR1400 containment are summarized in Table 2. The process approach (bottom-up) for preserving the most important physical phenomenon is to simulate a plume flow of hydrogen gas from the IRWST vent hole or the SG compartment to the dome area because the hydrogen generated in the core is discharged into the IRWST or the SG compartment during a severe accident. The selected nondimensional number for the process approach is the Rayleigh number, as defined using Equation 3.2-1 [24,30]. The calculated Rayleigh number using the safety analysis results for the APR1400 SBO sequence at approximately 60,000 seconds is 1.32 1011 1.83 1011 [24] (Table 3). The plume of the hydrogen gas may be a turbulent flow because the Rayleigh number is higher than 1.0 1010. To ensure the turbulent plume flow of the hydrogen gas in the experiment, the height of the test facility should be increased to approximately 9e10 m (Fig. 15), when the hydrogen gas at temperature 310 K discharges into the air environment at 290 K under atmospheric conditions. The diameter of the test facility can be determined by considering the height to diameter ratio of the APR1400 containment, as follows:

Ra ¼

00

bgq H ¼ ank 4

bg k DT H4 H ank

¼

bgDTH3 an

(3.2-1)

in which DT ¼ the temperature difference between the hydrogen gas and ambient gas; a ¼ thermal diffusivity; b ¼ thermal expansion ratio; g ¼ gravitational acceleration; H ¼ distance over the hydrogen gas can travel; and n ¼ kinematic viscosity.

3.2.2.

Facility description

The KAERI is preparing a large test facility to examine the safety issues, as stated in the outstanding issues related to hydrogen distribution and combustion and the performance

Fig. 16 e Schematic of the test vessel.

43


Fig. 17 e Corrosion of stainless steel. bromine, and chlorine.

a

Heat transfer fluid to control the temperature.

of safety devices such as the PAR under various severe accident situations. The content of detailed design specification was described in the paper [31]. The pressure vessel of the KAERI test facility is 9.5 m in height and 3.4 m in diameter, and the total free volume is 80 m3. The aspect ratio of height to diameter is 2.8, which is similar with that of the THAI test facility and is adequate for scaling the containment building down to examine the gas behavior under a severe accident. A larger free volume, compared to that of the THAI test facility, can offer a wide range of scaling. The design pressure of the vessel is 1.5 MPa at 450 K, and the six legs of the vessel can support a maximum of 130 tons (i.e., the vessel could be entirely filled with water as a scrubbing solution). The pressure vessel includes five separate double walls from the bottom head to the top head to control independently the wall temperature at various elevations, and to reduce the heat loss from a vessel to the environment. Fifty-one flanges of different sizes are installed on the wall at various elevations and orientations for the supply systems of steam, gases, and aerosols, and for measurement systems of gas distribution and aerosol characteristic. Fig. 16 shows a schematic of the KAERI test facility. Stainless steel 316L was chosen as the vessel material because of its corrosion resistance for a long operating time. Fig. 17 shows the possibility of corrosion of the test vessel. Stainless steel iron-based alloy with low carbon is a good choice to resist high corrosion under experimental conditions of steam and fission products. In March 2014, a pressure vessel with the specifications mentioned previously was constructed and installed at the KAERI site. The construction of additional facilities and systems for operating the KAERI test facility is an ongoing process. The commissioning tests for controlling the thermal

b

Examples of halogens are iodine,

hydraulic conditions and gas concentration at the KAERI test facility will be conducted soon.

3.3.

Test plan

The KAERI derived the research items to study the phenomena in the containment and to determine the performance of the accident mitigation systems (Fig. 18). Several phenomena have to be handled and have a connection with each other. The first use of the KAERI facility will focus on the study of the characteristics of a hydrogen mitigation system because in Korea various types of PARs have been installed in the containments of Korean nuclear plants as one of postFukushima actions [32]. Other research items will be focused on later because the test results of the PARs performance are first on demand from the utility and regulatory bodies. Table 4 shows the PAR installation status and plan in Korea. For operating plants, the utility company Korea Hydro and Nuclear Power (KHNP) completed the installation of PARs by the end of 2013. Most PARs used in Korean nuclear power plants are PARs developed by domestic companies. The APR1400 containment has its own compartmentalization features. The compartments in the APR1400 containment (Fig. 19) are a relatively open structure between compartments to induce gas released from the primary system to mix well and thereby result in protection from the local accumulation of hydrogen. The number of PARs in the APR1400 containment is also relatively small. Fig. 20 shows the location of the PARs in the containment. The KAERI plans to conduct performance tests of Korean PARs in a large-scale vessel with a volume over 80 m3 under various severe accidents. In addition, interaction tests between PARs located in the containment need to be conducted because the APR1400 has

44


its own compartmentalization features in the containment. The test results under severe accident conditions will be useful in compensating the PAR correlations made by the venders and will be used for optimization of PAR installation in a containment.

4.

Discussion

The activities on hydrogen safety after the Fukushima accident are briefly reviewed. The outstanding issues from the recent research results concerning hydrogen safety are

Fig. 18 e Phenomena occurring in a nuclear power plant containment in relation to hydrogen safety. CFVS (containment filtered venting system); HMS (Hydrogen Mitigation System); IRWST (in-containment refueling water storage tank); PAR (passive autocatalytic recombiner).

45


Table 4 e The Korean Hydro and Nuclear Power passive autocatalytic recombiner installation status and plan. Plant

Manufacturer

Schedule

Plant

Manufacturer

Schedule

Kori 1 Kori 2 Kori 3 Kori 4 YGN 1 YGN 2 YGN 3 YGN 4 YGN 5 YGN 6 UCN 1 UCN 2 UCN 3 UCN 4

AECL Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb Ceracomb

09.12 '13.05 12.10 '13.01 '13.08 '13.01 '12.10 '13.10 '13.09 '12.11 '13.09 '13.04 12.09 '12.12

UCN 5 UCN 6 Woulsung 1 Woulsung 2 Woulsung 3 Woulsung 4 ShinKori 1 ShinKori 2 ShinWonsung 1 ShinWonsung 2 ShinKori 3 ShinKori 4 Shin UCN 1 Shin UCN 2

Ceracomb Ceracomb KNT Ceracomb Ceracomb Ceracomb AREVA/KNT AREVA/KNT AREVA/KNT AREVA/KNT KNT KNT KNT KNT

'13.05 12.10 11.05 '13.05 '13.06 '13.02 12.01 11.10 11.10 12.03 12.08 12.10 '15.09 '16.07

AECL (Canadian nuclear vendor), AREVA (French nuclear vendor), KNT (Koran PAR vendor), Ceracomb (Korean PAR vendor), UCN (Ulchin), YGN (Younggwang).

discussed. Multidimensional analysis for hydrogen distribution and explosion in parallel is required with lumped codes to resolve hydrogen issues. It is specifically expected that multidimensional analysis for hydrogen distribution is necessary for determining the location and number of PARs to be installed in the containment. The KAERI research activities for coping with the hydrogen threat have been introduced. Code validation efforts for hydrogen distribution and explosion using multidimensional codes such as GASFLOW and COM3D have been performed, and the results of these codes have reasonably agreed with experimental results. Based on the experience of code validation, the analysis on hydrogen distribution and explosion using the multidimensional codes, GASFLOW, COM3D, and

OpenFOAM, have been preliminarily evaluated for APR1400. This application effort will continue so that a reasonable level of the calculation results can be reached for plant application. In addition, the conservative hydrogen distribution and explosion at 100% MWR in a reactor vessel will be performed using this multidimensional analysis system. KAERI is preparing a large-scale test to resolve hydrogen issues. The first use of the KAERI facility will focus on studying the characteristics of a hydrogen mitigation system because various types of PARs have been installed in the containments of Korean nuclear plants as a post-Fukushima action. The Korean utility and regulatory bodies are first demanding the test results of PAR performance. The experiment will be accompanied by validation with multidimensional analysis.

Fig. 19 e Schematic of the ARP1400 containment dimension in meter.

Fig. 20 e The passive autocatalytic recombiners (PARs) location in the ARP1400 containment (the pink squares indicate each PAR location).

46


Conflicts of interest The authors have no conflicts of interest to declare.

Acknowledgments This work was supported by a National Research Foundation of Korea grant funded by the Korean government (Ministry of Science, Information and Communications Technology, and Future Planning; grant number 2012M2A8A4025889).

references

[1] US Nuclear Regulatory Commission, TMI-2 Lessons Learned Task Force Status Report and Short-Term Recommendations [Internet]. Office of Nuclear Reactor Regulation, US Nuclear Regulatory Commission, Washington, DC, 1979 [cited 2015 Jan 5]. Available at: http://pbadupws.nrc.gov/docs/ML0900/ ML090060030.pdf. [2] National Academies Press, Lessons Learned from the Fukushima Nuclear Accident for Improving Safety of U.S. Nuclear Plants, National Academies Press, Washington, DC, 2014. [3] Organisation for Economic Co-operation and Development/ Nuclear Energy Agency (OECD/NEA), Agreement on the OECD/NEA HYMERS Project, Paris, France, 2012. [4] Organisation for Economic Co-operation and Development/ Nuclear Energy Agency (OECD/NEA), Agreement on the OECD/NEA THAI-Project, Paris, France, 2007. [5] Organisation for Economic Co-operation and Development/ Nuclear Energy Agency (OECD/NEA), Agreement on the OECD/NEA THAI-2 Project, Paris, France, 2011. [6] MAAP4, Modular Accident Analysis Program, for LWR Plants, Code Manual, vols. 1e4, Fauske and Associates, Inc., Burr Ridge, IL, Palo Alto, CA, USA, 1994. [7] R.O. Gauntt, R.K. Cole, C.M. Erickson, R.G. Gido, R.D. Gasser, S.B. Rodriguez, M.F. Young, S. Ashbaugh, M. Leonard, A. Hill, MELCOR Computer Code Manuals, Version 1.8.6, Users Guide and Reference Manual, NUREG/CR-6119, SAND 2005-5713, USNRC, Sandia National Laboratories, Albuquerque, NM, 2005. [8] J.R. Travis, J.W. Spore, P. Royl, K.L. Lam, T.L. Wilson, C. Muler, G.A. Necker, B.D. Nichols, R. Redinger, E.D. Haughes, H. Wikening, W. Baumann, G.F. Niederauer, GAFLOW: a Computational Fluid Dynamics Code for Gases, Aerosols, and Combustion, LA-13357-M, FZK-5994, Los Alamos National Laboratory, Los Alamos, NM, 1988. [9] A. Miassoedov, Overview of Severe Accident Research Activities at KIT, CSARP Meeting, Bethesda, MD, Sep. 11e13, 2012. [10] J. Kim, S.-W. Hong, S.-B. Kim, Analysis of Hydrogen. Behaviors during the Severe Accident by a Loss-of-coolant in the APR1400 Containment, Int'l Workshop on Passive Safety Systems in Advanced PWRs (IPASS), Shanghai Jiao Tong Univ., Shanghai, China, April 28e30, 2008. [11] H. HoshiI, Severe Accident Activity at JNES, CSARP Meeting, Bethesda, MD, Sep. 11-13, 2012. [12] T. Kanzleiter, S. Gupta, K. Fischer, G. Ahrens, G. Langer, A. Ku¨hnel, G. Poss, G. Langrock, F. Funke, Hydrogen, Fission Product, Issues Relevant for Containment Safety Assessment under Severe Accident Conditions, Becker Technologies GmbH, Eschborn, Germany, 2010. [13] T.K. Blanchat, A. Maliakos, Performance Testing of Pass Autocatalytic Recombiners, NUREG/CR-6580, SAND97e2632, USNRC, Sandia National Laboratory, Albuquerque, NM, 1998.

[14] J.-W. Park, B.-R. Koh, K. Suh, Demonstrative testing of honeycomb passive autocatalytic recombiner for nuclear power plant, Nucl. Eng. Des. 241 (2011) 4280e4288. [15] O. Auban, R. Zboray, D. Paladino, Investigation of large-scale gas mixing and stratification phenomena related to LWR containment studiers in the PANDA facility, Nucl. Eng. Des. (2006) 409e419. [16] H.D. Kim, D.H. Kim, S.B. Kim, J.H. Song, S.W. Hong, Investigation on the resolution of severe accident issues for Korean nuclear power plants, Nucl. Eng. Technol. 41 (2009) 617e648. [17] S. Kudriakov, F. Dabbene, E. Studer, A. Beccantini, re, A. Bentaib, A. Bleyer, J. Malet, J.P. Magnaud, H. Paille E. Porcheron, C. Caroli, The TONUS CFD code for hydrogen risk analysis: physical models, numerical scheme and validation matrix,, Nucl. Eng. Des. 238 (2008) 551e565. [18] D. Baraldi, M. Heitsch, H. Wilkening, CFD simulation of hydrogen combustion in a simplified EPR containment with CFX and REACFLOW, Nucl. Eng. Des. 237 (15e17) (2007) 1668e1678. [19] A.A. Efimenko, S.B. Dorofeev, CREBCOM code system for description of gaseous combustion, J. Loss Prev. Process Ind. 14 (2001) 575e581. [20] A. Kotchourko, A. Lelyakin, J. Yanez, Z. Xu, K. Ren, COM3D: Turbulent Combustion Code, User Guide, Version 4.3, Karlsruhe Institute of Technology, Karlsruhe, Germany, 2012. [21] ANSYS Inc, CFX-13 User Manual, Simpsonville, MD, 2013. [22] A. Bentaib, N. Chaumeix, SARNET H2 Combustion Benchmark; Specification of ENACCEF Tests, Technical Nucle aire Report, L'Institut de Radioprotection et de Suˆrete (IRSN), Fontenay-aux-Roses, France, 2011. [23] H.S. Kang, S.B. Kim, S.W. Hong, CFD Analysis for Diluents Effect on H2 Flame Acceleration in the ENACCEF Facility, Nuclear Thermal Hydraulics and Safety (NTHAS9), Nov. 16e19, Buyeo, Republic of Korea, 2014. [24] J. Kim, S.W. Hong, S.B. Kim, H.-D. Kim, Hydrogen mitigation strategy of the APR1400 nuclear power plant for a hypothetical station blackout accident, Nucl. Technol. 150 (2005) 263e282. [25] M.A. Movahed-Shariat-Panahi, Recommendation for maximum allowable mesh size for plant combustion analyses with CFD codes, Nucl. Eng. Des. 253 (2012) 360e366. [26] H. Weller, C. Greenshields, M. Janssens, OpenFOAM: the Open Source CFD Toolbox User Guide [Internet], 2014. Available from: http://www.openfoam.org/download (accessed 05.01.15). [27] J. Kim, S.B. Kim, H.J. Kim, Numerical method for evaluation of hydrogen flame acceleration in a compartment of a nuclear power plant, J. Comput. Fluid Eng. 15 (2010) 64e72. [28] J. Kim, S.W. Hong, Analysis of hydrogen flame acceleration in APR1400 containment by coupling hydrogen distribution and combustion analysis codes, Prog. Nucl. Energy 78 (2015) 101e109. [29] N. Zuber, G.E. Wilson, M. Ishii, W. Wulff, B.E. Boyack, A.E. Dukler, P. Griffith, J.M. Healzer, R.E. Henry, J.R. Lehner, S. Levy, F.J. Moody, M. Pilch, B.R. Sehgal, B.W. Spencer, T.G. Theofanous, J. Valente, An integrated structure and scaling methodology for severe accident technical issue resolution: development of methodology, Nucl. Eng. Des. 186 (1e2) (1998) 1e21. [30] A. Bejan, Convection Heat Transfer, second ed., John Wiley & Sons, Inc., Hoboken, NJ, 1995. [31] Y.S. Na, S.-W. Hong, S.-H. Honga, B.-T. Min, Introduction to Large-sized Test Facility for validating Containment Integrity under Severe Accidents, Transactions of the Korean Nuclear Society Spring Meeting Jeju, Korea, May 29-30, 2014. [32] Nuclear Safety and Security Commission (NSSC), NSSC Meeting, Nov. 2011.