Research activities on nuclear reactor physics and thermal-hydraulics in Japan after Fukushima-Daiichi accident

ABSTRACT

Research and development in nuclear reactor physics and thermal-hydraulics continue to be vital parts of nuclear science and technology in Japan. The Fukushima accident not only brought tremendous change in public attitudes towards nuclear engineering and technology, but also had huge influence towards the research and development culture of scientific communities in Japan. After the Fukushima accident, thorough accident reviews were completed by independent committees, namely, Tokyo Electric Power Company (TEPCO), the Japanese government, the Diet of Japan, the Rebuild Japan Initiative Foundation, and the Nuclear and Industrial Safety Agency. Reactor physics and thermal-hydraulics divisions of Atomic Energy Society of Japan (AESJ) also issued the roadmaps after the accident. As a result, lessons learned from the accident were made clear, and a number of new research activities were initiated. The present paper reviews ongoing nuclear engineering research activities in Japanese institutes, universities, and corporations, focusing on the areas in reactor physics and thermal-hydraulics since the Fukushima accident to the present date.

KEYWORDS:

• Fukushima-Daiichi
• nuclear fuels
• nuclear reactor
• reactor safety
• safety analysis
• severe accident
• thermal hydraulics

1. Introduction

Severe accidents (SAs) in Fukushima-Daiichi nuclear power plant accident (Fukushima accident) brought enormous change in public attitudes towards nuclear technology not only in Japan, but also around the globe. Immediately after the accident, all existing nuclear power plants in Japan were taken offline for the maintenance and safety inspections, which had powered nearly one-quarter of electricity. This resulted in huge economic and environmental impacts due to the increase in coal/natural gas import, and restart of the old fossil-fuel power plants to make up the electricity supplied by the nuclear power. Meanwhile, investment to promote renewable energy has grown rapidly, especially in the area of photovoltaic (PV) technology, aided by the feed-in tariff (FIT) program under Japanese government's promotion [1]. The program obliged utility companies to buy power generated by PV at above market price, but such effort only improved the solar power's contribution on the power mix in Japan from 0.4% in 2012 to 3.4% in 2015. Meanwhile, the subsidizing program has resulted in addition of the renewable power surcharge to consumer electric bill, and created shortage of the grid capacity in various sites. As a result, a number of FIT-approved PV projects were terminated in Japan and green energy investments have plateaued since 2016. Moreover, following the shortfall in power output by nuclear, Japan became one of the world's largest importers of liquefied natural gas (LNG), and this increase in import has widened Japan's trade deficit. After facing these realities, public acceptance for nuclear technology has begun to come back slowly. In order to restart nuclear power plants, utility companies must satisfy a series of requirements and standards set by Japan's Nuclear Regulation Authority (NRA) which include completion of reactor upgrades and requiring the local government's consent. As of September 2017, out of 42 operable reactors in Japan, five nuclear power plants (Ikata 3, Sendai 1 and 2, and Takahama 3 and 4), which have met the post-Fukushima regulatory standards set by NRA, have resumed for operation, seven reactors (Genkai 3 and 4, Mihama 3, Ohi 3 and 4, and Takahama 1 and 2) are currently under pre-operation inspections for possible restart, and restart-related documents were filed for twelve reactors (Tomari 1, 2 and 3, Shika 2, Tsuruga 2, Shimane 2, Hamaoka 3 and 4, Kashiwazaki-kariwa 6 and 7, Onagawa 2, and Higashidori 1). In addition, as was recently announced by the Japanese government on the energy policy, power generation by nuclear energy will continue to be a vital part of primary baseload power source in Japan. For the future energy mix proposed by METI, about 20%–22% of electricity is projected to be generated by nuclear power by 2030 [2].

While these drastic changes took place since March 2011, nuclear communities in Japan continued their effort on the researches and developments on nuclear science and engineering to make the technology safer and more reliable by adopting the lessons learned from Fukushima-Daiichi accident. Thorough accident investigations performed by independent committees including Tokyo Electric Power Company (TEPCO), the Japanese government, the Diet of Japan, the Rebuild Japan Initiative Foundation and the Nuclear and Industrial Safety Agency summarize lessons learned from the accident [3]. In the area of thermal-hydraulics, Atomic Energy Society of Japan (AESJ) has established the roadmap for light water reactor (LWR) safety improvement and development to achieve higher safety standards for nuclear power plants [4]. A similar roadmap was established by the reactor physics division of AESJ in 2012 [5].

The present paper reviews ongoing nuclear engineering researches since the Fukushima accident to the present date at Japanese research institutes, universities, and industries, focusing on the core nuclear engineering disciplines including reactor physics and thermal-hydraulics, as of year 2017.

2. Nuclear reactor physics

In the present chapter, studies and works relevant to the nuclear reactor physics carried out by Japanese reactor physics community after the Fukushima accident are described. Those presented in Journal of Nuclear Science and Technology published by the AESJ and in Nuclear Science and Engineering published by the American Nuclear Society (ANS) are mainly addressed, and some important works presented in other journals and in proceedings of international conferences are also cited. Studies and works cited in the present chapter are classified into the following six technical areas: (1) works relevant to the Fukushima accident, (2) nuclear data, (3) deterministic methods for reactor analyses, (4) uncertainty quantification for neutronics parameters, (5) particle transport and criticality simulation with Monte Carlo method, and (6) developments relevant to future nuclear systems. Some important technical subjects in the nuclear reactor physics such as criticality safety, reactor noise analysis, and reactor monitoring/diagnosis methods are not explicitly addressed in the present review because of limitation of authors’ knowledge.

2.1. Research activities relevant to the Fukushima-Daiichi nuclear power plant accident

Just after the Fukushima accident, concerned issues related to the reactor physics were possibility of recriticality and quantification of radioactive materials inventories in reactor cores and spent fuel pools.

Possibility of recriticality in the Fukushima-Daiichi power plant cannot be eliminated, but it has been widely recognized from various measurement data obtained at the site that recriticality with significant increases in fission reactions has never occurred after the accident. The recriticality issue in the plant has been summarized by Nakajima [6].

Quantification of total radioactive materials inventories in the plant is important from many aspects such as estimation of release rates of radioactive materials to the environment, potential risk of the plant, and identification of damaged nuclear fuels in the cores. Nishihara et al. have evaluated fuel compositions in the plant from available data [7]. This work is quite helpful since detailed information such as nuclide-wise number densities, weight and radioactivities with cooling time after the accident have been presented with detailed information on the evaluation conditions.

Measurement data of radioactivities measured in and around the site, which have been released by TEPCO, are helpful to identify damaged fuels in the cores and to estimate release rates of radioactive materials to the environment. It has been well known that radioactivity ratio of cesium-134 to cesium-137 can be an indicator of fuel burnup, so estimation of burnup of damaged fuels from measurement data of this radioactivity ratio has been attempted by several researchers [8–11].

After the emergency period, concerns from a view point of the reactor physics have been shifted to recriticality risk during defueling and materials identification of fuel debris for its proper management. In order to accumulate fundamental criticality data relating critical control technology for fuel debris treatment, a modification program of the static experiment critical facility, STACY, is now undergoing in Japan Atomic Energy Agency (JAEA) [12]. Izawa et al. have calculated neutron multiplication factors of uranium dioxide–concrete mixture, which might be generated through the molten core–concrete interaction (MCCI) during the accident, with various conditions, and have concluded that concrete provides efficient neutron moderation which increases the recriticality risk [13]. Figure 1 shows infinite neutron multiplication factors of spherical systems in which UO₂ sphere is surrounded by concrete moderator with perfect reflective boundary conditions. This result shows that infinite neutron multiplication factors can exceed unity even if spent fuel with fission products nuclides is assumed. Uniformness of materials cannot be guaranteed and material spatial distributions are expected in fuel debris, so Ueki has conducted criticality analysis with the Monte Carlo particle transport method under material distribution uncertainty [14]. Sub-criticality monitoring is also essential to prevent recriticality, so both experimental and theoretical works have been conducted [15,16]. For efficient management of recriticality risk, Nauchi et al. have proposed concept of neutron capture credit based on neutron-induced gamma ray spectroscopy, and this concept has been validated through experiments conducted at Kyoto University Critical Assembly (KUCA) [17]. Takezawa et al. have evaluated effect of debris dust generated by drilling in defueling on neutron multiplication by fissions, and have proposed a passive measure which can prevent rapid increase of neutron multiplication factors [18]. Works done by Tuya and Obara on supercritical transient analysis in coupled fuel debris systems using the integral kinetic model should be also cited [19].

Figure 1. Infinite neutron multiplication factors of UO₂–concrete systems [13].

From a view point of special nuclear material accountancy, Sagara et al. have pointed out the importance of inventory quantification of low-volatile fission product and special nuclear materials, and have conducted feasibility study of passive gamma spectrometry of fuel debris which had been applied to the Three Mile Island Nuclear Power Plant Unit 2 [20].

2.2. Nuclear data

Nuclear data are fundamental physical quantities which are mandatory in neutronics calculations in every nuclear system. Since it is impossible to obtain true value to each of nuclear data, evaluation processes based on nuclear physics model calculations and/or (differential) measurement data are essential. Evaluation results are summarized in evaluated nuclear data files or libraries, and these files are supplied to neutronics calculations. As widely known, there are three major nuclear data files in the world: ENDF in US, JEFF in Europe and JENDL in Japan. The JENDL library has been developed by Japanese experts on nuclear physics and nuclear engineering, and the latest version, JENDL-4.0, has been released in 2010 [21]. JENDL-4.0 mainly focuses on minor actinoid nuclides and fission products which are important in neutronics calculations of future innovative nuclear systems. Since the release of JENDL-4.0, many advancements have been attained in the field of nuclear data. In this subsection, some of them which directly contribute to the advancement in nuclear engineering are described. Needless to say, other activities which are rather close to nuclear physics than to nuclear engineering are also important, whereas these are not addressed in the present paper.

In a research area of nuclear data measurements, a lot of works have been done using the neutron beam of Accurate Neutron-Nucleus Reaction measurement Instrument (ANNRI) at the Japan Proton Accelerator Research Complex/Materials and Life science experimental Facility (J-PARC/MLF) [22]. In ANNRI, neutron beam emitted from moderator goes through T0 chopper, neutron filter, disk choppers, collimators and two detector systems, array of large Ge detectors and NaI detectors. High beam intensity of ANNRI makes it possible to measure neutron–nuclide reaction cross sections in high-background environment, so cross-section measurements of radioactive nuclides with high precisions are possible. With the ANNRI beamline, capture cross sections can be accurately measured, and those which are important for innovative future nuclear systems, such as curium-244 and -246 [23], neptunium-237 [24], americium-241 [25] and some fission product nuclides [26] have been successfully measured. It is notable that measurement of Cm-244 capture cross section has never been conducted in the world so far. With this beamline, simultaneous measurement of fission and capture cross sections of americium-241 has been also carried out [27]. Nuclear data measurements for some actinoid nuclides and fission products have been also conducted at Tokyo Institute of Technology (TIT) and at Kyoto University Research Reactor Institute (KURRI).

In a research area of nuclear data evaluations, re-evaluations on many intermediate-heavy nuclides have been conducted, and revised data files have been accumulated by JAEA. Revised isotopes are platinum, mercury, xenon, tantalum, iodine, tellurium, rhodium, antimony, ruthenium, rhenium, praseodymium, gallium, erbium, technetium and cesium. All of these new evaluations have been reported in scientific papers in Journal of Nuclear Science and Technology.

In addition to neutron–nuclide reaction cross sections, radioactive decay and fission yields data of fission product nuclides are also important in inventory and decay heat calculations for spent nuclear fuels. JENDL FP Decay Data File 2011 (JENDL/FPD-2011) and JENDL FP Fission Yields Data File 2011 (JENDL/FPY-2011) have been developed and released in 2012 [28]. Furthermore, in 2015, decay data of non-FP nuclides have been added, and a more comprehensive file, JENDL Decay Data File 2015 (JENDL/DDF-2015), has been released [29]. With this new file, consistent nuclide transmutation calculations with all nuclides including actinoids, fission products and light elements have been realized.

In order to adopt evaluated nuclear data files to neutronics calculations, these files need to be converted into specific forms which are suited to neutronics codes, such as Monte Carlo particle transport codes. This conversion is generally referred to as processing of evaluated nuclear data files. So far, the NJOY code developed by Los Alamos National Laboratory has been mainly utilized for nuclear data file processing in the worldwide, but in Japan, own computer code has been developed to obtain more flexibility. This computer code is named FRENDY, and the current version of FRENDY can provide the specifically formatted nuclear data for general Monte Carlo particle transport codes [30].

In the development of nuclear data files, performance tests (integral validation) against measurement data of neutronics parameters obtained at facilities like critical assemblies are mandatory. Various kinds of these measurement data (or integral data) should be collected and supplied to performance tests. There have been many contributions from Japanese community of the reactor physics experiments to the world. We cite here the experimental data on highly enriched uranium systems with different neutron flux energy spectra measured at the Fast Critical Assembly (FCA) of JAEA [31], the data for the large-size fast reactor development acquired during an experimental program conducted by the United States and Japan [32], and the data of burnup reactivity swing obtained at Japanese fast reactor JOYO [33]. Works done by Kitamura et al. to revisit measurement techniques such as the pulsed neutron source method, the neutron correlation method and the critical-water-level method for reactor physics experiments should be also cited here [34–36].

In Japan, a new project of general-purpose evaluated nuclear data file development has been just launched. All the recent and current activities on the field of nuclear data will be summarized as a release of the new version of JENDL, JENDL-5, in future.

2.3. Deterministic methods for reactor analyses

Numerical simulations of neutron transport, prompt and delayed neutrons emissions by fission reactions, and nuclides transmutation through neutron–nuclide reactions have been one of the most important technical subjects in the nuclear reactor physics, and currently this importance still remains without any changes. Neutron transport calculations for nuclear reactor cores with the Monte Carlo method, which enable us to do precise treatment on phase space variables such as energy, space and particle direction, have been realized recently with significant advancements in computers, but the deterministic methods which discretize these phase space variables have played important roles in various aspects in nuclear reactor analyses, and there have been many advancements and developments in the deterministic methods for reactor analyses in Japanese reactor physics community.

The two-step reactor analysis procedure, which consists of assembly-level calculations with detailed heterogeneous structure and whole-core-level calculations with assembly-homogenized structure, has resulted in great success, but recently many researchers and engineers have attempted to realize direct heterogeneous three-dimensional whole-core transport calculations. Since direct application of the method of characteristics (MOC), which has been generally used in 2D single- or multi-assembly calculations, to a three-dimensional reactor core is unrealistic even with current high-performance computers, most researchers and engineers have proposed, tested and improved the planar MOC approach, which treats whole-core as a set of two-dimensional planes which are coupled with each other by axial neutron leakage. The CHAPLET-3D code developed by Kosaka et al. is one of them [37]. The planar MOC approach is efficient from a view point of calculation cost and accuracy, but it has a problem in convergence stability in some cases. In order to overcome this problem, an alternative method, the axially simplified MOC in 3-D (ASMOC3D) method has been proposed by Giho et al. [38]. Based on the idea of ASMOC3D, a new three-dimensional heterogeneous transport solver GENESIS has been developed by Yamamoto et al. [39]. The Legendre polynomial Expansion of Angular Flux (LEAF) method has been devised and implemented to GENESIS. In the LEAF method, a geometry is covered by sets of parallel planes that are considered as an extension of ray traces drawn in 2D x–y geometry to z-direction as shown in Figure 2. Since explicit ray traces are not considered for z-direction, memory storage and computational cost are smaller than those of direct 3D MOC calculations. The performance of GENESIS has been verified through various benchmark calculations, and the results have indicated high fidelity of GENESIS.

Figure 2. A side view of ANNRIFig. 3 Concept of the Legendre polynomial expansion of angular flux method implemented to GENESIS [39].

Since MOC has become the most popular deterministic method for reactor analyses, several researches to improve efficiency of MOC have been carried out. Tabuchi et al. have proposed a method to improve iteration convergence in multi-group (e.g. 172-group) heterogeneous calculation with the transport cross section [40]. A study by Takeda et al. on preconditioners for the generalized minimal residual method applied to MOC calculations can be also cited [41]. For MOC in fuel depletion calculations, Tabuchi et al. have proposed a practical method to correct neutron flux spatial distribution obtained by coarse-spatial mesh MOC calculations [42].

As mentioned above, direct three-dimensional whole-core heterogeneous transport calculations have been realized recently, but the conventional two-step procedure is still important in design calculations. One of important key issues in the two-step procedure is homogenization without significant degradation of accuracy. The generalized equivalence theory with the neutron flux discontinuity factors is a widely used approach to mitigate errors associated with the homogenization. In general, the diffusion theory has been employed in whole-core calculations with homogenized assemblies (nodal calculations) and the application of discontinuity factors to this kind of problems has been well established. Advanced methods such as the simplified Pn (SPn) theory with homogenized pincells (pin-by-pin calculations) have become practical in recent years, and more accurate and rigorous method, such as the transport theory with the integro-differential equation form, is expected to be used in future. However, application of the neutron flux discontinuity factors to these advanced methods and theories has not been addressed by any researchers. Yamamoto et al. have initially addressed this issue and have proposed the homogenization method with the even-parity discontinuity factors (EPDFs) which can be applied to the integro-differential transport equation [43]. While EPDF requires even-parity and odd-parity angular fluxes calculations, a different method, an angular flux discontinuity factor method, which does not use even- and odd-parity fluxes, has been also proposed by Sakamoto et al. [44]. More recently, discontinuity factors for the SP3 theory has been derived from explicit angular flux representation for the SPn method by Yamamoto et al. [45].

On reactor core calculations of boiling water reactors, a series of researches to improve calculation accuracy of pin-by-pin calculations have been carried out by Fujita et al. This includes a correction technique to capture spectral interference effect caused by adjacent loadings of different types of fuel assemblies on energy-collapsed cross sections [46,47], and a new macroscopic cross section model for a core simulator [48]. A work done by Mitsuyasu et al. on neutronics and thermal-hydraulics coupling in the frame of the two-step procedure is also cited here [49].

There are other technical subjects on deterministic methods for reactor analyses, and several researches on these subjects have been carried out by Japanese reactor physicists. As examples, we refer here development of the multigrid amplitude function method [50] and extension of the integral kinetics model [51] for space-dependent reactor kinetics calculations. On resonance self-shielding treatment which is to generate energy-averaged neutron–nuclide cross sections supplied to lattice and core neutron transport calculations, works done by Koike et al. should be cited [52,53]. Development of acceleration methods for iterative calculations has been also important subject and a new coarse-mesh finite difference (CMFD) acceleration method for the SP3 advanced nodal method has been proposed by Yamamoto et al. [54] On fuel depletion (nuclide transmutation) calculations, while the Chebyshev rational approximation method to calculate matrix exponential has become general, an alternative method, the mini-max polynomial approximation method, has been proposed by Kawamoto et al. [55]. A next-generation framework for reactor physics calculations named MARBLE developed by JAEA is also notable [56].

2.4. Uncertainty quantification for neutronics parameters

Designs of nuclear systems and demonstrations of their safety performance generally rely on numerical simulations. Neutronics parameters including various kinds of safety parameters are predicted by neutronics calculation codes with nuclear data. Due to uncertainty in nuclear data and numerical treatments in neutronics calculations, neutronics parameters predicted are uncertain. Historically, nuclear data used for neutronics calculations of fast neutron systems such as fast reactors have been considered one of the important technical issues, so quantification of nuclear data-induced uncertainty of neutronics parameters and its reduction by using measurement data obtained at critical facilities have been important technical subjects in fast reactors development [57,58]. In Japan, this kind of development has been benefitted to actual works in fast reactor core designs [59,60]. On the other hand, in thermal neutron systems such as LWRs, numerical errors in neutronics calculations had been dominant. Recently, numerical errors in neutronics calculations of LWR cores have been significantly reduced by virtue of rapid advancements in computer technology, and hence quantification of nuclear data-induced uncertainty of neutronics parameters has been paid significant attention also in LWRs calculations.

While a lot of numerical procedures for uncertainty quantification of neutronics parameters exist, these can be classified into the adjoint-based procedure and the stochastic-based (random sampling-based) procedure [61,62]. In the former procedure, sensitivity coefficients of target neutronics parameters with respect to nuclear data are calculated with the well-established perturbation theory (PT) or the generalized perturbation theory (GPT), and then uncertainty propagation from nuclear data to target parameters is considered with these sensitivity coefficients calculated. This procedure is quite efficient since it does not require any significant computational cost once sensitivity coefficients are calculated. The adjoint-based procedure, however, has to introduce an important assumption that target neutronics parameters are dependent linearly on input nuclear data. Also, it is quite difficult to treat target parameters which are determined by complicated physical process such as fuel depletion over a whole reactor core or time-dependent transient with thermal feedback. By this background and advancement of computer technology, the random sampling-based procedure has been utilized more often in recent years. The random sampling-based procedure does not require any assumptions such as linearity of target parameters, and therefore it can be applied to any statistical parameters with arbitrary statistical distributions although it has a drawback that a much larger amount of computer resources is required than the adjoint-based procedure.

On nuclear data-induced-uncertainty quantification of neutronics parameters of LWRs, a work by Yamamoto et al. focusing on safety parameters such as assembly power and control rod worth of commercial pressurized water reactors (PWRs) should be cited [63]. In this work, uncertainty propagation from nuclear data to safety parameters of reactor core is realized by step-wise sensitivity profiles: sensitivity of assembly property to nuclear data and sensitivity of core safety parameter to assembly property. These sensitivities are calculated by numerical differentiation or GPT. While this work did not take into account nuclear fuel depletion, uncertainty of neutronics parameters during fuel depletion has been quantified by using the depletion perturbation theory (DPT) by Chiba et al. [64]. This work, however, is limited to fuel pincell model of LWRs. Uncertainty quantification of neutronics parameters of LWRs core with fuel depletion and thermal feedback effect has been realized by Yamamoto et al. [65] In this work, uncertainties of various neutronics parameters such as critical eigenvalue, maximum assembly burnup during operation cycle, shutdown margin, peaking factor and power spatial distribution have been quantified in commercial BWR and PWR cores with the random sampling-based procedure.

The random sampling-based procedure, which has been used in uncertainty quantification, has been extended. The research group of Nagoya University led by Yamamoto and Endo has applied the random sampling-based procedure to the nuclear data adjustment method [66] and the bias-factor method [67]. These works have enhanced the usefulness of the random sampling-based procedure. A method of estimating sensitivity profiles through random sampling calculations has been also proposed by Chiba et al. [68]. Efficient procedure to quantify statistical uncertainty of estimated statistical quantities inherent in the random sampling-based procedure has been proposed by Endo et al. [69].

Uncertainty quantification of decay heat and delayed neutron emissions, which are quite important in safety analyses, has also been carried out. Katakura has quantified decay heat uncertainty using several recent evaluated nuclear data files [70], and Chiba et al. have quantified delayed neutron emission rates uncertainty using the PT for time-dependent problems [71]. Uncertainty reduction of decay heat by the nuclear data adjustment using post-irradiation examination data has been attempted by Kawamoto et al. [72]. Since decay heat and delayed neutron emission rates are dependent on common nuclear data such as radioactive decay data and fission yields data, simultaneous nuclear data adjustment for these quantities have been performed by Chiba et al. [73].

Theoretical studies on uncertainty quantification and reduction method should also be addressed. Yokoyama et al. have proposed an extended nuclear data adjustment method (ENA) which minimizes uncertainties of target neutronics parameters, not uncertainties of adjusted nuclear data [74]. Furthermore, they have proposed another method, a regressive nuclear data adjustment method (RNA), and also they have derived unified formulation of several variants of nuclear data adjustment methods including ENA and RNA [75]. This theoretical work provides important suggestion that the normal distribution assumption, which has been considered mandatory in the nuclear data adjustment method, is unnecessary. A work done by Umano et al. on representativity factors, which quantify similarities of several mock-up experiments to target system property, has been also carried out [76].

2.5. Particle transport and criticality simulation with Monte Carlo

The Monte Carlo methods have recently played important roles in nuclear reactor analyses due to their high precision on numerical modeling on neutrons behavior and complicated system geometry. There are several Monte Carlo production codes for reactor analyses in the world: MCNP in US and TRIPOLI in France, for example. In Japan, the MVP code has been developed by JAEA, and the new version, MVP-III, has been just released [77]. The recent developments of MVP including new models/capabilities such as perturbation calculation for eigenvalues, delayed neutrons simulations, group constant generation, new resonance elastic scattering model and reactor kinetics parameters calculations [78] have been summarized in a paper in Annals of Nuclear Energy [79]. Several works with MVP, such as fuel temperature coefficient calculations with the new resonance elastic scattering model [80] and benchmark calculations for whole core of LWR [81], should be also cited.

A new concept, a Monte Carlo method with complex-valued weights, has been proposed by Yamamoto of KURRI, and this has been applied to calculate/generate leakage-corrected energy-averaged cross sections and anisotropic diffusion coefficients [82], to perform frequency domain analyses of neutron noise [83,84], to calculate kinetics parameters [85], and to solve the frequency domain neutron transport equation [86] by Monte Carlo. Important works on statistical error estimations and convergence diagnosis done by Ueki should be addressed [87–91]; further descriptions of them are not provided here due to limitation of authors’ knowledge on this research field. A recent work done by Nauchi, who has carried out a breakthrough work on kinetics parameter calculations with Monte Carlo [92], is also cited here [93]. An efficient Monte Carlo simulation using graphic processing unit done by Okubo et al. is also noted [94].

2.6. Researches and developments on next-generation nuclear systems

In Japan, researches and developments to realize the nuclear fuel cycle have been carried out by national research institutes, electric power companies, vendors and universities over 60 years. Through the development, an experimental fast reactor JOYO, a prototype fast reactor MONJU and several pilot plants for spent fuel reprocessing and fuel manufacturing have been built and operated. After the Fukushima-Daiichi accident, however, shutdown and decommissioning of MONJU have been decided by Japanese government in 2016 while JOYO has been in preparation for licensing to new regulation. Although MONJU is going to be decommissioned, the nuclear fuel cycle is still one important candidate as future energy systems in Japan, and relevant works are now being conducted continuously.

Fast reactors have an important role to recycle minor actinide nuclides as well as power generation, and core design for efficient minor actinide recycling has high priority. Ohki et al. have proposed a fast reactor core concept for heterogeneous minor actinide loading [95]. If fast reactors become main energy sources in Japan in future, there should be a transition period from conventional LWRs to fast reactors, and during this period, fuel composition of fast reactors should be varied with the transition phase. Maruyama et al. have found correlations between fuel compositions and neutronics parameters of reactors [96]. The effect of innerduct of fuel subassembly, which is introduced to cope with hypothetical core disruptive accidents, on neutronics parameters has been also investigated by Ohgama et al. [97].

Measurement data obtained through various tests conducted at MONJU have been accumulated, and analyses of some of these measurement data have been conducted and published. Those for important neutronics parameters, criticality, control rod worth and isothermal temperature coefficients, of a restarting core in 2010 have been summarized in a set of papers presented in Nuclear Technology [98–100]. Recent work on feedback reactivity of this core done by Kitano et al. should be also cited [101].

Metallic-fuel fast reactors have been pursued by the Central Research Institute of Electric Power Industry (CRIEPI). Minor actinide transmutation has been confirmed through experiments and analyses on uranium–plutonium–zirconium alloys containing minor actinides and rare earth irradiated at the Phenix fast reactor in France [102].

Accelerator-driven subcritical system (ADS) is also a promising future nuclear system, which can transmute long-lived minor actinoids to reduce total radiotoxicity of spent nuclear fuels. Various experimental works have been carried out at KUCA for the ADS development, and invaluable and important data have been accumulated. Recent publications on mockup experiments on the thorium-loaded ADS [103,104] and sample reactivity experiments for lead nuclear data validation [105] are cited here. Specific technical subjects on ADS, such as estimation of effective delayed neutron fraction, which is one of important kinetics parameters, in sub-critical state [106,107], and sub-criticality monitoring by power spectral analysis [108,109] and by pulsed spallation neutron source [110], should also be noted. Development of a code system dedicated to ADS neutronics design calculations [111] and design study for reactivity control of ADS with use of gadolinium hydride [112] have also been carried out.

Other various innovative future nuclear systems have been proposed by Japanese researchers and engineers. It is impossible to cover the whole here, so some of them are addressed: light water-cooled fast reactors for plutonium breeding [113], molten-salt-type reactor for powering space probes [114], LWRs with rock-like oxide fuel in the phase-out scenario [115], plutonium burner system based on high-temperature gas-cooled reactor with high proliferation resistance [116], and axially heterogeneous boiling water reactor for burning transuranium elements developed by Hitachi, Ltd. [117] It should be noted that a research group of TIT led by Obara has proposed several unique concepts, such as long-life small prismatic high-temperature gas-cooled reactor with passive decay heat removal [118], and has advanced the CANDLE-type reactor concept also [119].

3. Research activities in nuclear thermal-hydraulics

In this chapter, research activities in nuclear thermal-hydraulics after the Fukushima -Daiichi accident are reviewed. Review of thermal-hydraulic researches in LWR's SAs published in Journal of Nuclear Science and Technology are well summarized by Prof. Kataoka [120], and summary of the SA researches in Japan before/after the Fukushima accident, based on the conference proceedings, is given in the article by Prof. Sugimoto [121]. AESJ's roadmap for thermal-hydraulics researches to improve LWR safety is reported by Nakamura et al. [4]. In the current review chapter, major nuclear thermal-hydraulic researches carried out in Japan after 2011 are classified into following groups: (1) severe accident researches, (2) Fukushima-Daiichi (1F) accident analysis, (3) development of new safety systems, (4) LWR safety research, and (5) development of next-generation reactors. For reviewing the notable thermal-hydraulics researches, works published in major nuclear engineering related journals and conference proceedings are mainly selected.

3.1. Severe accident (SA) researches

As was reviewed by the researchers [121,122], SA researches in Japan became highly active after the Fukushima accident, aiming to improve the severe accident analysis codes for better understanding of in-vessel and ex-vessel phenomena. A number of numerical works have been carried out by several institutions to simulate multi-dimensional molten metal dynamics. For the experimental works, in-vessel core degradation phenomena and ex-vessel molten debris spreading experiments are reported mainly from the universities. Some of the notable works are reviewed in this section.

Development of the in-vessel detailed core degradation model for BWR was initiated by Regulatory Standard and Research Department, Secretariat of Nuclear Regulation Authority following the Fukushima-Daiichi accident. The Multifunction Model developed by the Regulatory Authority focuses on in-vessel core degradation from cladding temperature escalation to molten metal mixture flow at RPV lower plenum. Assessment of the Multifunction Model against in-vessel core degradation phases was conducted by Okawa et al. [123] with respect to the experimental works performed at Sandia National Laboratories (DF-4 and XR-2 experiments), Karlsruhe Institute of Technology (QUENCH-06 and CORA-18 experiments), and Joint Research Centre facilities (FARO L-19 and KROTOS K-37 experiments). It was reported that the Multifunction Model is capable of simulating the main features of in-vessel core degradation phases including phenomena such as cladding temperature escalation and hydrogen generation, axial- and lateral directional motion of molten corium, and molten corium interaction with water pool in RPV lower head.

Development of the SA analysis code, SAMPSON (Severe Accident Analysis Code with Mechanistic, Parallelized Simulations Oriented toward Nuclear Field) has been actively carried out at industries under IMPACT project [124]. Since 2011, SAMPSON code has been utilized to understand the accident progression of Fukushima-Daiichi [125–127]. Hidaka et al. [128] reported on the improvement of debris spreading analysis (DSA) module in SAMPSON code, which simulates the molten core spreading on a reactor containment floor and MCCI phenomena, by considering thermal degradation of concrete for the erosion rate. Hidaka et al. [129] advanced DSA module by considering three-dimensional advection, diffusion, and volume reduction of eroded concrete into molten core for the 3D MCCI analysis (Figure 3). Wei et al. [130] at the University of Tokyo performed the experiment called LIVE-L4 test using KNO₃–NaNO₃ melt to validate the debris coolability analysis (DCA) module in SAMPSON code. The DCA module calculates the debris heat removal rate at the lower plenum to provide the safety margin of the reactor vessel. Alternative usage of heat transfer correlations was suggested by Wei et al. [130] for more accurate prediction of the average melt pool temperature in the lower plenum. SAMPSON also enables one to estimate the cesium accumulation level in the reactor, and based on the results, contaminated water accumulated in Fukushima-Daiichi can be estimated [131]. While the improvement of SAMPSON modules is still underway, it can potentially become one of the major SA analysis tools along with MAAP and MELCOR, by adopting the lessons learned from the Fukushima accident.

Figure 3. Schematic of the debris spreading analysis (DSA) module in the SAMPSON code [129].

JAEA has developed the THALES2 [132] for level 2 probabilistic risk assessment, which analyzes the progression and the source term of SAs in LWRs. Ishikawa et al. [133] applied the THALES2 code for a BWR-5 with a Mark-II containment to conduct level 3 probabilistic risk assessment through source term evaluation. KICHE code [134] was also developed at JAEA to simulate iodine chemistry based on kinetics of chemical reactions and mass transfer. The accident progression and the transportation of radioactive materials are analyzed with the THALES2 coupled with the KICHE to take account of the iodine reaction kinetics. Iodine release rate from Fukushima-Daiichi unit 3 was analyzed with THALES2/KICHE codes by Ishikawa et al. [133] and the effects of aqueous phase of iodine chemistry on the source term was evaluated by the sensitivity analysis.

Molten metal spreading simulation using commercial CFD (computational fluid dynamics) code and in-house developed code are reported by the research institutions and universities. Utilization of moving particle semi-implicit (MPS) method, originally proposed by Koshizuka and Oka [135], to calculate molten metal spreading has been actively conducted at Waseda University. The MPS method is capable of simulating molten metal-free surface without introducing computational grid, and it is suitable for analyzing solidification behavior of molten metal involving large deformation without relying heavily on empirical correlations. In the MPS method, phase change model with temperature-dependent surface tension, viscosity correlation proposed by Ramacciotti et al. [136], and implicit calculation of pressure Poisson equation are typically utilized. Kawahara and Oka [137] used the MPS method to calculate the leading-edge position of the molten metal spreading behavior, and its 3D calculation results showed a good agreement with existing data without utilizing empirical parameters. Ability to simulate the crest formation was confirmed in the MPS simulation. Chen et al. [138] simulated the melt freezing behavior in an instrument guide tube penetration of BWR's RPV bottom head. Calculation results were compared with various experimental programs utilizing molten metal and showed good agreement. For the analysis of the MCCI phenomena, which is a crucial element in new regulatory requirement, usage of the MPS method was done by Li and Yamaji [139] and Chai et al. [140]. Analysis of the stratification and solidification/melting phenomena of molten tin was conducted by Li et al. [141] and the MPS method was capable of capturing stratified layer of the molten metal. These studies that are cited in this section showed the MPS method's potential as one of the promising SA analysis tools.

JAEA has developed the in-house CFD code called JUPITER (JAEA Utility Program with Immersed boundary Technic and Equations of multi-phase flow analysis to simulate Relocation behavior of molten materials) to simulate the three-dimensional molten core behavior including solidification and relocation [142]. At a moment, the code does not include the effect of crust formation, but it was possible to simulate the spreading front position of the corium spreading experiment done at VULCANO facility. Simulation of the molten metal spreading using commercial CFD code was conducted by Kobayashi et al. [143] at Hokkaido University using STAR-CCM+. As was the case for the MPS method, by utilizing Ramacciotti et al.'s viscosity model [136], it was possible to track molten metal's front position measured at VULCANO facility. However, the VULCANO experiment was designed for core-catcher design purpose, where corium sample spreads smoothly on the floor surface. Numerical investigation involving more realistic SA scenario, such as corium penetration through RPV bottom surface, inclusion of the collisional effect against floor surface, and interaction with stagnant pool water, will be the future challenge for 3D CFD.

Basic experiments simulating breakup and fragmentation behaviors of molten metal jet in the lower plenum of BWR using simulated rod bundle test section were conducted by Saito et al. [144,145] at Tsukuba University. Empirical correlation to predict fragmentation diameter as a function of Weber number, Reynolds number, density ratio, and viscosity ratio was proposed in their work. For the experimental works targeting on ex-vessel phenomena, molten metal's spreading and deposition behaviors were investigated by Matsumoto et al. [146] at Hokkaido University. In their experiment, high-frequency induction heating system was utilized to create molten metal and it was dropped onto a flat heated surface by adjusting outlet nozzle diameter, fall height, and initial melt temperature. Scaling analysis was performed to identify the key parameters associated with spreading and solidification. Such database will be useful to benchmark spreading models utilized in SA codes.

3.2. 1F accident analysis

A number of studies were carried out to understand the Fukushima-Daiichi accident progression. TEPCO analyzed the accident progression behaviors and source terms with the MAAP code for units 1, 2, and 3 of Fukushima-Daiichi, which were based on the updated data from interviews with operators and on-site surveys [147]. Efforts have been made to improve the model of the MAAP code. The model enhancement items to improve the simulation capability for molten corium behavior in the accidents at the Fukushima-Daiichi nuclear power plants were validated with the Phenomena identification and ranking table (PIRT) [148]. Figure 4 shows system subcomponents for the MAAP which was used for the simulation of the accident progression of Fukushima-Daiichi unit 3 to construct the PIRT.

Figure 4. System subcomponents for the PIRT [148].

Benchmark Study of the Accident at the Fukushima-Daiichi Nuclear Power Station (BSAF) Project by OECD/NEA was launched in November 2012 [149]. The project is hosted by JAEA and 17 organizations from eight countries simulating thermal-hydraulic behavior of the Fukushima-Daiichi, units 1 through 4 with the SA codes. From Japan, the CRIEPI with the MAAP 5.0, Institute of Advanced Energy (IAE) with the SAMPSON, the JAEA with the THALES-2, and the Nuclear Regulatory Authority with the MELCOR 2.1 codes participate in the project. The accident analyses for units 1, 2, and 3 with the SAMPSON have been reported in the Journal of Nuclear Technology [150–152]. Participants of the BSAF project reported the knowledge obtained from the comparison of their results of SA analyses [153].

Model comparison between the SAMPSON and MELCOR was carried out by Itoh et al. [154] for the SA analysis of the Fukushima-Daiichi unit 1. The results are comparable to MAAP-MELCOR crosswalk study by EPRI [155] because the same initial conditions were applied for the accident analysis.

The Fukushima accident analysis has been conducted at Japanese universities as well. Sensitivity study of the melt behavior for unit 1 with the MELCOR and the MPS method was conducted at Waseda University [156]. Figure 5 shows the dryout debris bed configuration by the MELCOR for preceding calculation and computational domain with the MPS method. Analysis to clarify the Reactor core isolation cooling system (RCIC) on the accident progression of unit 2 was conducted with the RELAP/ScdapSIM at the University of Tokyo [157,158], with the TRAC-BF1 code at the University of Fukui [159].

Figure 5. Computational domain and nodalization in the MPS method [156].

Works on the source term analysis with SA codes were also reported. The short-term FP release was evaluated with the SAMPSON code to estimate accumulated FP in the contaminated water [160]. Iodine release from unit 3 was evaluated with the THALES2 for the progression of SA coupled with KICHE for the iodine chemistry in aqueous phase [134,161]. Uncertainty and sensitivity analyses for nuclear reactor SA source terms were done using MELCOR for an accident sequence similar to that of unit 2 [161,162].

For assisting the possible estimation on debris distribution, thermal hydraulic PIRT and source term PIRT were developed by the Research Expert Committee on Evaluation of Severe Accident of AESJ [163]. In order to apply the lessons learned from Fukushima-Daiichi accident, further enhancement of defense in depth (DID) was imposed for ABWRs in Kashiwazaki-Kariwa Nuclear Power Station [164]. For improved DID, external event features as well as multiple failure of the safety systems were considered. TEPCO is planning to restart operations of Kashiwazaki-Kariwa plants in 2019.

One of the remaining questions from the accident for Mark-I type BWR is the considerable difference in PCV pressure behaviors that were recorded in units 2 and 3 of Fukushima-Daiichi. For both units, RCIC systems were operating to supply makeup coolant into the reactor pressure vessel (RPV). Mizokami [165] pointed out that RCIC in unit 2 discharged two-phase steam due to the malfunction of SRV, and the one in unit 3 heated up the suppression pool water to generate thermal stratification. It is expected that such phenomena may have caused the increase in PCV pressure. Furthermore, it is expected that the SA will degrade the suppression pool's heat sink and filtration functions [166]. In order to investigate this phenomenon, thermal stratification experiment was carried out at a scaled-down suppression pool chamber at University of Tokyo [167]. The effect of steam flow rate on different thermal stratification behaviors was investigated in their study.

Three hydrogen explosions during Fukushima accident brought attention to the presence of hydrogen and thermal-hydraulic behavior in the reactor building during the SA [168]. New regulatory requirements that were established in 2013 contain clear countermeasures against hydrogen explosions in the PCV and reactor building. Researches on hydrogen are mainly subdivided into three kinds, namely hydrogen generation, hydrogen distribution analysis, and hydrogen combustion, respectively. For estimating the vapor concentration in reactor building during station black out (SBO) scenario, lumped parameter method was suggested by Kondo et al. [169] to calculate the transient behavior of temperature and vapor concentration in a BWR operating floor. The proposed lumped parameter model showed 280,000 times faster calculation time than the three-dimensional CFD analysis. It was also identified that the evaporation model was affected by the thermal state of spent fuel pool.

3.3. New safety systems development

In order to reduce the risks and prepare countermeasures against SA, studies on new safety systems are actively conducted in Japan, especially in the areas of filtered containment venting system (FCVS) and steam injector (SI) developments [170–181].

The purpose of the FCVS is to filter aerosol particles as well as radioactive iodine and cesium to reduce the release of radioactive particles and to prevent the primary containment vessel (PCV) from over-pressurization in case of SA. However, filtration performance of the FCVS with varying boundary conditions and prolonged operation period are yet unknown.

Investigation of the self-priming phenomena and hydrodynamic phenomena within venturi scrubber section for FCVS was carried out at University of Tsukuba and JAEA [171,172]. At TEPCO, FCVS performance tests were conducted at various inlet flow conditions using aerosol particles with varying size range and their facility is depicted in Figure 6 [173]. The system is composed of scrubber and metal filter sections, and aerosol particles are filtered as they pass through these media. According to the study, inertial deposition was identified as the governing mechanisms to filter aerosol in FCVS. To characterize aerosol particles in FCVS system, they suggested the use of aerodynamic diameter. In addition, decontamination factor was introduced to characterize the FCVS performance. Decontamination factor showed a tendency to increase with respect to aerodynamic diameter of aerosol particles. For the aerosol particles of aerodynamic diameter greater than 0.4 micron, decontamination factor was greater than 1000. Moreover, decontamination factor in the water scrubber section showed increasing tendency with respect to Stokes number.

Figure 6. Schematic of the FCVS system [173].

At CRIEPI, full height FCVS test facility of 8 m operable at 0.8 MPa and 170 °C inlet steam flow was utilized with particles such as aerosol, iodine, and organic iodine [174,175]. In their study, iodine decontamination performance was assessed using bubbling effect.

Ishii et al. [176] and Narabayashi et al. [177] at Hokkaido University utilized silver-doped zeolite (AgX) as a filtration media, which is reported to remove 99.99% of CH₃I contained in the ventilation line. It was reported that the sorption capacity of AgX for CH₃I was 0.21 g/g (AgX) at 24 °C, and high humidity in gas phase tended to decrease the sorption capacity for CH₃I. It was also reported that the higher temperature in the sorption column led to increase in the sorption capacity. Further research and development of FCVS for improved filtration performance as well as to identify stable operation mode to avoid possible flow instabilities are encouraged.

Another new safety system, SI, is a passive jet pump which utilizes direct contact condensation between steam and liquid to discharge subcooled water at higher pressure than the inlet pressures. Due to its simple design and stable operation, it has a potential to be applied as a passive core coolant injection system to make up water in RPV or other safety systems such as isolation condenser (IC) and passive containment cooling system (PCCS). In the past, central steam-jet-type SI has been widely investigated, but recent findings show that central water-jet-type SI equipped with overflow port shows reliable start-up and stable operation. Comprehensive review of the SI research up until 2014 is given by Takeya et al. [178]. In Japan, SI researches are mainly carried out at universities to investigate the maximum attainable discharge pressure, stable operation mode, local temperature and velocity distribution within the unit [179–181]. There still remain various issues to be solved, such as reliable heat transfer and hydrodynamic models for direct contact condensation between steam and water jets, relationship between flow instability and SI's operation mode, and operation at higher inlet steam pressure. Furthermore, fundamental investigation of the two-phase flow dynamics at the diffuser section, which is a crucial part to raise the SI's discharge pressure, is suggested as part of the future studies.

3.4. LWR safety research

In this section, LWR safety-related researches after Fukushima-Daiichi accident in Japan are reviewed. A number of LWR safety studies using system analysis codes were reported by several institutions. OECD/NEA Rig-of-Safety Assessment (ROSA) Project experiment using the Large Scale Test Facility (LSTF) was conducted simulating a PWR loss of feedwater (LOFW) transient with assumptions of high-power natural circulation due to failure of scram and total failure of high-pressure injection system. Post-test analysis with the RELAP5 code for the experiment simulating PWR LOFW transient was performed to well understand the observed phenomena [182]. Analyses with the RELAP5/MOD3.2 code were performed for PWR small-break LOCA simulation test with LSTF [183]. The RELAP5 code was also utilized for the tests simulating cold-leg small-break loss-of-coolant accidents with 2-inch diameter break using an early steam generator (SG) secondary-side depressurization with or without release of nitrogen gas dissolved in accumulator water [184]. The code predicted most of the overall trends of the major thermal-hydraulic responses.

The second OECD/NEA Rig-of-Safety Assessment (ROSA-2) Project was performed with the LSTF, which simulated thermal-hydraulic responses during a PWR cold-leg intermediate break loss-of-coolant accident (IBLOCA). RELAP5 post-test analysis was performed for the experiment [185]. At the University of Tokyo, to understand the dynamics of reflooding, assessment and validation for the RELAP/SCDAPSIM have been performed for the QUENCH tests data [186].

Studies to improve a correlation of countercurrent flow limitation (CCFL) for reactor system codes are carried out. Steam–water experiments were conducted in an inverted U-tube simulating SG U-tube at 0.1–0.14 MPa and Wallis-type CCFL correlation for SG U-tube in PWRs was derived [187]. The effects of fluid properties on CCFL were evaluated based on experimental results for air–water, steam–water, and air–glycerol water solution [188]. Research on CCFL for hot leg and pressurizer surge line in PWRs were also conducted to obtain more accurate correlation based on experimental data and computations [189–191].

Investigation of the heat transfer characteristics of seawater became increasingly important after Fukushima-Daiichi accident, where seawater was directly injected into RPVs immediately after the accident. Experimental works have been carried out at JAEA for quite some time, and comparative studies to investigate heat transfer characteristics of seawater, artificial seawater, and NaCl solutions were reported [192–194]. Uesawa et al. [194] conducted pool boiling and forced-convective boiling experiments using artificial seawater. It was found that the deposition of calcium sulfate on heating surface became noticeable when seawater concentration around the heating surface is lower than 11 wt%, and the deposition caused temperature excursion. Such temperature excursion phenomenon was enhanced for forced-convective boiling on vertical upward flow case, and higher sea salt concentration at the lower part of the test section promoted temperature excursion at much lower heat flux than that of pool boiling case. It is worthy to note that recent TEPCO report revealed that the seawater injected into RPVs on the accident day was not reached into the vessel due to the complex piping system and inability to control the flow system under SBO situation. Preparedness to secure the emergency coolant injection lines is crucial, and must be remembered as one of the lessons learned from the Fukushima accident.

Fundamental thermal-hydraulic researches focusing on reactor component and system are also conducted by various researchers. Possible utilization of molten metal to seal the water leakage was conducted using gallium and Wood's metal [195]. As the leakage of contaminated water has been becoming an issue at Fukushima-Daiichi, such methodology may be a practical approach to solve water leakage problem [195]. New air cooling system for ABWR under natural convection conditions was presented by Mochizuki and Yano [196]. The system utilizes 240 finned heat transfer tubes and their numerical calculation demonstrated that two of these systems could remove 3926 MWt decay heat.

Several experimental works on the steam separator of LWR for improved operational performance were conducted at Hitachi Research Laboratory [197] and Kobe University [198,199]. The aim of the former study was to reduce the pressure drop without decreasing the steam–water separation performance. In order to achieve the goal, swirl-vane section was relocated from diffuser section and its curvature was modified. Their results showed 30% improvement in pressure drop from the conventional swirl-vane. For the latter studies performed at Kobe University, a steam separator with three pick-off rings was utilized to characterize the two-phase swirling flow behavior. Pressure drop and the liquid-separation characteristics of the modified swirlers with six and eight vanes were also examined [198,199].

Fundamental studies on rod bundle geometry are actively carried out in various institutions. Development of high-efficiency grid-spacer using CFD was reported by Ikeda [200]. The numerical results on rod-gap cross-flow velocity values were benchmarked with rod-embedded laser Doppler velocimetry (LDV) system, and satisfactory agreement was observed. It was suggested that careful designs on springs, dimples and mixing vanes (MVs) are crucial to control the rod-gap cross-flow. Hosokawa et al. [201] also developed small submersible LDV probe to measure two-component liquid velocity in a 4 × 4 rod bundle. Kawahara et al. [202] experimentally investigated the effects of a grid spacer with MV equipped with two different vane angles on annular flows in 16 mm ID channel. Measurement of the pressure drop across grid spacer, and entrainment fraction and deposition rate at downstream of the spacer were conducted. MV with 20 degree angle showed reduced pressure drop across the spacer compared to the one with 30 degree angle, while maintaining liquid droplet re-deposition rate. For predicting the interfacial shear term or averaged void fraction value in rod bundle geometry, drift–flux correlations are utilized in nuclear thermal-hydraulic system analysis codes. Ozaki et al. [203] developed new distribution parameter for rod bundle geometry which includes the effects of radial and axial power distributions and unheated rods on void fraction. Validity of the model was assessed with other existing drift–flux correlations using FRIGG and NUPEC data for different bundle types with varying unheated rod locations and numbers [204]. In the paper, recommended drift–flux correlations for different bundle geometry types were suggested.

Thermal-hydraulic phenomena involving flow-turning element, such as T-junction [205] and elbow section [206–208], continue to be an important issue among Japanese thermal-hydraulic communities. Qian et al. [205] carried out simulation work on thermal fatigue phenomena due to the temperature fluctuations generated at T-junction. CFD/Finite Element Analaysis (FEA) coupling analysis, where dynamic Smagorinsky model (DSM) for the LES SGS turbulence model, a hybrid scheme for calculation of the convective terms in conservation equation, and direct calculation of heat transfer between a fluid and a structure through thermal conduction, were utilized, and good agreement against experimental data was observed. Fujisawa et al. [206] utilized a test section equipped with elbow, orifice and straight pipe to investigate the non-axisymmetric wall-thinning phenomena. In the past, swirling flow observed in the pipeline caused pipe rupture in Mihama nuclear power plant and clear understanding of flow-accelerated corrosion (FAC) continues to be a crucial research topic. In their experimental work, it was found that magnitude of the swirling intensity plays a major role, and non-axisymmetric feature of the flow disturbances created at pipe-turning element is magnified with the presence of swirl generated at the orifice section. Researches on flow-induced vibration (FIV) continue to be an important issue. As summarized by Miwa et al. [207], linkage between two-phase flow and FIV is not yet clearly understood, and many empirical-based models to predict peak frequency and amplitude were developed for axial-flow and cross-flow orientations. For internal two-phase FIV, predictive models for wavy-stratified two-phase flow regime was developed by Miwa et al. [208] by treating wavy gas–liquid interface as a source of collisional force. Further studies on two-phase FIV need to be conducted to understand the relationship between secondary/swirling flow and vibration phenomena for improved piping system integrity and plant safety.

3.5. Next-generation reactors

In this section, researches and developments on next-generation reactors related to reactor thermal-hydraulics are introduced. Active research works have been continued for the high-temperature gas reactor (HTGR), fast reactor (FR), and molten salt reactors (MSRs) since before the Fukushima-Daiichi accident. Fundamental works on next-generation reactor including super-fast reactor and MSR are reviewed as well.

3.5.1. High-temperature gas reactors

Researches and developments on HTGR have been carried out for a number of years in Japan using high-temperature engineering test reactor (HTTR). HTTR in JAEA is the first HTGR in Japan with a thermal power of 30 MW and maximum coolant (helium) outlet temperature of 950 °C [209]. Not only as a power generation, due to its utilization of high-temperature coolant, HTGR can be potentially utilized for hydrogen production by coupling with a chemical plant using iodine–sulfur process (IS process), and a set of safety requirement to satisfy the regulations was presented by Sato et al. [210]. Various safety evaluation tests using HTTR have been conducted at JAEA including thermal-load fluctuation test, reactivity temperature coefficient measurement and loss of forced cooling (LOFC) test [211]. New advanced reactor named naturally safe high-temperature gas-cooled reactor (NSHTR) concept was also developed at JAEA after the Fukushima accident [212].

The HTGR possesses various advantageous features and some of the notable works are briefly reviewed. One of the advantageous features of HTGR is that in case of loss of core cooling and reactor reactivity control, it can spontaneously stabilize itself. Evaluation of the CO concentration and heat generated by graphite oxidation in HTTR were analytically demonstrated [212]. Depressurization test considered as the design basis accident was investigated for very high-temperature reactor (VHTR) [213]. Understanding of the air ingress behavior during a pipe rupture accident is crucial for VHTR, since the core graphite structures may be possibly oxidized. From the natural circulation experiment and 3D-CFD analysis performed on a reverse U-shaped channel, it was found that the natural circulation of air can be controlled by helium gas injection, and the method can possibly be applied as one of the preventative measures against primary pipe rupture accident.

In addition, Hamamoto et al. [214] investigated on the helium gas purification control for HTGR using the newly established analytical model embedded into primary helium purification system (PHPS) [214]. Plant's stability and reliability of facility components and long-term high-temperature operation using HTTR in the high-temperature test operation mode of 950 °C for 50 days were carried out by Shimizu et al. [215]. Sato et al. [216] investigated the feasibility of core heat removal of HTGRs under depressurized loss of forced circulation (DLOFC) without active or passive safety systems. Takamatsu [217] carried out 3D numerical simulation on LOFC test under 30% reactor power (9 MW) without scram on HTTR using STAR-CCM+. For the removal of decay heat during loss of forced cooling accident, HTTR is equipped with reactor vessel cavity cooling system (RCCS), which is capable of removing heat by passive cooling. Tsuji et al. [218] conducted CFD analysis to evaluate the heat removal performance of RCCS of HTTR. Takatmatsu and Hu [219] proposed a new RCCS design that consists of continuous closed regions at exterior of RPV and cooling section which is in contact with ambient air.

3.5.2. Fast reactors

Following the Fukushima-Daiichi accident, few FRs-related researches on MONJU and Japan sodium-cooled fast reactor (JSFR) were reported. While MONJU facility was decided to be shutdown in late 2016, research works carried out to understand the liquid metal fast reactor system have essential values in view of reactor thermal-hydraulics and new reactor development for next generation. In this subsection, notable research works published in major nuclear engineering journals after the Fukushima accident are introduced. While a large number of sodium-related experiments were performed in Japan in the past few decades, recent studies are more focused on code development and investigation of the fluid/structure interaction.

Understanding of thermal stratification characteristic is crucial to ensure the structural integrity of the liquid metal fast breeder reactors (LMFBRs). Thermal stratification in the upper plenum of MONJU was investigated using a commercial CFD code FLUENT [220–221].

Thermal fatigue caused by thermal mixing phenomena in JSFR was assessed using the MUGTHES code [222]. The MUGTHES code couples unsteady thermal-hydraulics and heat conduction in structure to simulate thermal mixing phenomena between fluid and structure.

Development of the SERAPHIM (Sodium–Water Reaction Analysis: Physics of Interdisciplinary Multi-Phase Flow) code has been carried out by JAEA to evaluate the possible failure of SG tubes in sodium-cooled fast reactors (SFRs). The code is capable of calculating compressible multicomponent multiphase flow with sodium–water chemical reaction. Numerical simulation results of water vapor discharging into sodium pool under unstructured mesh-based numerical method was conducted by Uchibori et al. [223], and good predictive capability of fluid temperature distribution around the SG tube was confirmed.

3.5.3. Supercritical reactors

The supercritical-water-cooled reactor (SCWR) was proposed as one of the six Generation-IV reactors, which has advantageous features including high thermal efficiency and simplified plant system. It is operated at 22.1 MPa, and hence no phase change will be observed during the normal operation. In Japan, researches and developments of supercritical-pressure LWR (Super LWR) have been conducted at the University of Tokyo and Waseda University for a number of years as summarized by Oka and Mori [224]. In addition, researches on supercritical-pressure light water-cooled fast reactor (Super FR) have been carried out at Waseda University.

For Super FR, startup characteristics using recirculation system was studied by developing detailed startup procedures and using time-dependent thermal-hydraulic analysis code [225,226]. Due to its high flow-to-power ratio, it was shown that the maximum cladding surface temperature (MCST) was not sensitive to the change of inlet coolant temperature, gap volume, and flow rate. Sutanto and Oka [227] also conducted studies on passive safety systems of a Super FR, consists of IC, automatic depressurization system, core make-up tank (CMT), gravity-driven cooling system (GDCS), and PCCS. Analysis of the anticipated transient without scram (ATWS) with the Super FR equipped with single flow pass core was also carried out by Sutanto and Oka [228].

Safety analysis of Super LWR with double tube water rods core was designed by Tamiya et al. [229]. Subchannel analysis of the Super LWR using turbulent mixing rate law was conducted by Wu and Oka [230]. Compared to typical LWR case, it was shown that turbulent mixing rate was enhanced for supercritical pressure fluid (SPF) and such difference led to the difference in MCST prediction. Comparative studies between the PWR model and the SPF model were carried out under different hydraulic parameters across the assembly, varying pin power distribution, mass flow rate, and number of water rods.

3.5.4. Molten salt reactors

Researches and developments of MSR were first initiated at Oak Ridge National Laboratory in 1950s. Due to its advantageous features including high thermodynamic efficiency, high neutron economy, large power density, and so on, continued efforts on researches and developments are carried out as one of the Generation IV reactors. Analysis of the passive decay heat removal system for MSR targeting for thorium-based FIJI-233Um of 450 MWth and fluoride-salt-cooled high-temperature reactor (FHR) has been reported following the Fukushima-Daiichi accident.

MSR FUJI-233Um, designed for Th-233U fuel cycle, consists of core with hexagonal graphite blocks and standard fuel salt for the primary loop is ⁷LiF–BeF₂–ThF₄–UF₄, which possesses the melting point at around 457 °C. The fuel salt is heated up around 697 K and its heat is transferred to the secondary coolant salt of NaBF4–NaF which eventually generates supercritical steam. The overall thermal efficiency of FUJI-233Um is reported to be over 44% [231]. Ishiguro et al. [231] conducted feasibility study of the passive decay heat removal system for FUJI-233Um, and utilization of the salt drain tank as part of the system confirmed the safe cooling of fuel salt.

The FHR is a new concept of nuclear power reactor, where its design concept has been developed mainly in the US and China. In FHR, liquid salt called Flibe (Li₂BeF₄) with melting point of 459 °C and boiling point of over 1400 °C is used as a coolant. For the fuel, graphite-matrix coated-particle fuel is utilized. Similar to SFRs, FHR operates at low pressure and high temperature. Its design inlet and outlet temperatures are 600 and 700 °C, respectively. As a result, decay heat removal systems are similar to that of SFRs. However, due to the high melting temperature of fluoride salt compared to that of sodium (98 °C), safety system may potentially overcool the reactor in case of accidents. Hence, development of the operation procedure, including identification of the grace time, is necessary for the accident management viewpoint. Performance evaluations of passive decay heat removal system involving natural circulation called direct reactor air cooling system (DRACS) and reactor vessel air cooling system (RVACS), as well as beyond design basis accident (BDBA) studies, were conducted at Hokkaido University [232–234].

4. Conclusions

The present article reviewed ongoing nuclear engineering researches since the Fukushima accident to the present date at Japanese research institutes, universities, and corporations, focusing on the core nuclear engineering disciplines including reactor physics and thermal-hydraulics. In reactor physics area, studies and works conducted in six areas including (1) works relevant to the Fukushima Daiichi accident, (2) nuclear data, (3) deterministic methods for reactor analyses, (4) uncertainty quantification for neutronics parameters, (5) particle transport and criticality simulation with Monte Carlo method, and (6) developments relevant to future nuclear systems were reviewed. For thermal-hydraulics, five areas including (1) SA researches, (2) Fukushima-Daiichi (1F) accident analysis, (3) development of new safety systems, (4) LWR safety research, and (5) development of next generation reactors, reported by research institutes, industries, and universities, were reviewed based mainly on the publications from major nuclear engineering journals.

As is evident from the present review article, the nuclear community in Japan strives for the advancement of nuclear technology by adopting lessons learned from the Fukushima accident. Roadmaps created by the reactor physics and thermal-hydraulic divisions of AESJ should play effective roles; however, the contents should be comprehensively reviewed and updated by experts frequently to avoid future obsolescence. It is authors’ sincere hope that nuclear community in Japan will continue to strive for the improvement of the safety and reliability of nuclear power plants as a top priority to truly overcome the Fukushima accident.

Nomenclature
ABWR	=	Advanced boiling water reactor
ADS	=	Accelerator-driven subcritical system
AESJ	=	Atomic Energy Society of Japan
ANNRI	=	Accurate neutron-nucleus reaction measurement instrument
ANS	=	American Nuclear Society
ASMOC3D	=	Axially simplified MOC in 3-D
ATWS	=	Anticipated transient without scrap
BDBA	=	Beyond design basis accident
BWR	=	Boiling water reactor
CFD	=	Computational fluid dynamics
CMT	=	Core make-up tank
CRIEPI	=	Central Research Institute of Electric Power Industry
DCA	=	Debris coolability analysis
DID	=	Defense in depth
DLOFC	=	Depressurized loss of forced circulation
DPT	=	Depletion perturbation theory
DRACS	=	Direct reactor air cooling system
DSA	=	Debris spreading analysis
ENA	=	Extended nuclear data adjustment method
FCVS	=	Filtered containment venting system
FEA	=	Finite element analysis
FHR	=	Fluoride salt-cooled high-temperature reactor
FIT	=	Feed-in tariff
FIV	=	Flow-induced vibration
FR	=	Fast reactor
GDCS	=	Gravity-driven cooling system
GPT	=	Generalized perturbation theory
HPI	=	High-pressure injection
HTGR	=	High-temperature gas reactor
HTTR	=	High-temperature engineering test reactor
IAE	=	Institute of Advanced Energy
IBLOCA	=	Intermediate break loss-of-coolant accident
IC	=	Isolation condenser
JAEA	=	Japan Atomic Energy Agency
JSFR	=	Japan sodium-cooled fast reactor
JUPITER	=	JAEA Utility Program with Immersed boundary Technic and Equations of multi-phase flow analysis to simulate Relocation behavior of molten materials
KUCA	=	Kyoto University Critical Assembly
KURRI	=	Kyoto University Research Reactor Institute
LDV	=	Laser Doppler velocimetry
LEAF	=	Legendre polynomial Expansion of Angular Flux
LMFBR	=	Liquid metal fast breeder reactor
LNG	=	Liquefied natural gas
LOFC	=	Loss of forced cooling
LOFW	=	Loss of feedwater
LWR	=	Light water reactor
MAAP	=	Modular Accident Analysis Program
MCCI	=	Molten core–concrete interaction
MCNP	=	Monte Carlo N-particle transport code
MCST	=	Maximum cladding surface temperature
MOC	=	Method of characteristics
MSR	=	Molten salt reactor
MV	=	Mixing vane
NPS	=	Nuclear power stations
NRA	=	Nuclear Regulation Authority
NSHTR	=	Naturally safe high-temperature gas-cooled reactor
PCCS	=	Passive core cooling system
PCV	=	Pressurized containment vessel
PGC	=	Primary gas circulators
PHPS	=	Primary helium purification system
PIRT	=	Phenomena identification and ranking table
PT	=	Perturbation theory
PV	=	Photovoltaic
PWR	=	Pressurized water reactor
RCIC	=	Reactor core isolation cooling
RCCS	=	Reactor vessel cavity cooling system
RNA	=	Regressive nuclear data adjustment method
ROSA	=	Rig-of-Safety Assessment
RPV	=	Reactor pressure vessel
RVACS	=	Reactor vessel air cooling system
SA	=	Severe accidents
SFR	=	Sodium-cooled fast reactors
SG	=	Steam generator
SAMPSON	=	Severe Accident Analysis Code with Mechanistic, Parallelized Simulations Oriented toward Nuclear Field
SCWR	=	Supercritical water-cooled reactor
SERAPHIM	=	Sodium–water reaction analysis: physics of interdisciplinary multi-phase flow
SI	=	Steam injector
SPF	=	Supercritical pressure fluid
TEPCO	=	Tokyo Electric Power Company
VCS	=	Vessel cooling system
VHTR	=	Very high-temperature reactor
1F	=	Fukushima-Daiichi

Acknowledgments

One of the authors (Go Chiba) wishes to express his deep gratitude to Prof. A. Yamamoto of Nagoya University for providing a figure about the LEAF method.

Disclosure statement

No potential conflict of interest was reported by the authors.