Reflood Phase Research Articles

Fast prediction of key parameters in FEBA using the COSINE subchannel code and artificial neural network

Numerical techniques have emerged as an essential tool for operators and designers to preemptively acquire key parameters in accidents analysis. However, due to insufficient experience, it is difficult for them to obtain satisfactory numerical results. Moreover, the uncertainty analysis and quantification necessitate the simulation of a substantial number of samples, which requires a significant amount of computational time. Therefore, the development of a fast prediction model becomes imperative. In this work, a prediction model based on the in-house COSINE subchannel code and Multi-Head Perceptron (MHP) is developed. The COSINE subchannel code is employed to provide data sets for training neural networks. Firstly, the numerical results of COSINE subchannel code are compared with experimental data to ensure the accuracy of data sets. Secondly, input features for neural networks are selected by evaluating the impact of input parameters on numerical results, and a series of simulations is carried out to generate data sets. Then, a comparative analysis was conducted between the Multi-Layer Perceptron (MLP) and Support Vector Regression (SVR) models, and the MLP model performs better. Subsequently, the MLP was compared with the MHP, demonstrating the advantage of MHP model. Based on this, the predictions are conducted using the MHP model and the distribution of key parameters is compared with that obtained by COSINE subchannel code. The results illustrate that developed MHP model is an efficient tool for predicting key parameters during the reflooding phase.

Reliable prediction of peak cladding temperature during top reflooding processes: A neural network-based perception-interaction approach

The reliable prediction of peek cladding temperature is crucial for heat transfer and reactor security when large break loss of coolant accident (LBLOCA) happens. This study proposes a neural network-based Perception-Interaction system to effectively predict and monitor the peak cladding temperature during reflooding phase after LOCA. The system employs a narrow rectangular channel for conducting top reflooding experiments, wherein the peak temperature is predicted by the Back Propagation Neural Network (BPNN) algorithm. The system’s primary objective is to accurately predict the peak cladding temperature of various test conditions. First, the special Counter-Current Flow Limitation (CCFL) phenomenon is identified based on the different heat exchange patterns inside the test section. Then the experimental data is collected by LabVIEW and calculated by MATLAB to accurately predict the quench velocity and peak temperature at a specific location by conducting various test conditions. The final experiment results indicate that the built Perception-Interaction system in this article has a high prediction accuracy, with a relative error within 2% between the predicted peak temperatures and actual temperatures.

Research on thermal hydraulic characteristics of reflooding process based on core multi-channel model in system analysis code

There are four main stages in PWR Loss of Coolant Accident (LOCA): blowdown, re-irrigation, re-flooding and long-term cooling. The re-flooding stage is a critical phase to avoid core meltdown and achieve long-term cooling. RBHT reflooding experiment was simulated with the RELAP5 and COSINE program respectively.Thermal-hydraulic characteristics in re-flooding phase were analyzed using multi-channel model. Quench time is mainly influenced by system pressure and coolant subcooling. Later quench will result in a higher cladding temperature. Comparatively, the peak cladding temperature is less affected by the power of the heating rod and the re-flooding rate, which fluctuates within 7 %. The higher re-flooding rate allows quenching to occur as fast as possible and improves cooling efficiency. However, higher coolant sub-cooling and re-flooding rates may cause greater thermal stresses on the core and endanger core integrity. Based on the analysis of the re-flooding rate on the peak cladding temperature and quench time, the optimal re-flooding rate interval of 0.099–0.105 m/s was obtained.

Machine Learning Applications and Uncertainty Quantification Analysis for Reflood Tests

The reflooding phase, a crucial recovery process after a loss of coolant accident (LOCA) in reactors, involves cooling overheated fuel rods with subcooled water. Its complex nature, notably in its flow regime and heat transfer, makes prediction challenging, resulting in high uncertainty and computation cost. In this study, we utilized the data assimilation (DA) technique to enhance the prediction of reflooding phenomena and subsequently deployed machine learning models to predict the accuracy of the safety and performance analysis code (SPACE) simulation. To generate the dataset for the machine learning model, we employed the sampling method for highly nonlinear system uncertainty analysis (STARU), providing a high-quality dataset for a complex problem such as a reflooding simulation. In this dataset, the physical models were assimilated under their selected uncertainty bands and utilized the effective sampling approach of STARU, generating the high-quality output and efficient enhancement of SPACE predictions. Consequently, the implemented machine learning model can be used to enhance model development and uncertainty quantification (UQ) analysis using the system code.

Analysis of LBLOCA of APR1400 with 3D RPV model using TRACE

It is very difficult to capture the multi-dimensional phenomena such as asymmetric flow and temperature distributions with the one-dimensional (1D) model, obviously, due to its inherent limitation. In order to overcome such a limitation of the 1D representation, many state-of-the-art system codes have equipped a three-dimensional (3D) component for multi-dimensional analysis capability. In this study, a standard multi-dimensional analysis model of APR1400 (Advanced Power Reactor 1400) has been developed using TRACE (TRAC/RELAP Advanced Computational Engine). The entire reactor pressure vessel (RPV) of APR1400 has been modeled using a single 3D component. The fuels in the reactor core have been described with detailed and coarse representations, respectively, to figure out the impact of the fuel description. Using both 3D RPV models, a comparative analysis has been performed postulating a double-ended guillotine break at a cold leg. Based on the results of comparative analysis, it is revealed that both models show no significant difference in general plant behavior and the model with coarse fuel model could be used for faster transient analysis without reactor kinetics coupling. The analysis indicates that the asymmetric temperature and flow distributions are captured during the transient, and such nonuniform distributions contribute to asymmetric quenching behaviors during blowdown and reflood phases. Such asymmetries are directly connected to the figure of merits in the LBLOCA analysis. Therefore, it is recommended to employ a multi-dimensional RPV model with a detailed fuel description for a realistic safety analysis with the consideration of the spatial configuration of the reactor core.

Study on the influence of flow blockage in severe accident scenario of CAP1400 reactor

Deformed fuel rods can cause a partial blockage of the flow area in a subchannel. Such flow blockage will influence the core coolant flow and further the core heat transfer during the reflooding phase and subsequent severe accidents. Nevertheless, most of the system analysis codes simulate the accident process based on the assumed flow blockage ratio, resulting in inconsistencies between simulated results and actual conditions. This paper aims to study the influence of flow blockage in severe accident scenario of the CAP1400 reactor. First, the flow blockage model of ISAA code is improved based on the FRTMB module. Then, the ISAA-FRTMB coupling system is adopted to model and calculate the QUENCH-LOCA-0 experiment. The correctness and validity of the flow blockage model are verified by comparing the peak cladding temperature. Finally, the DVI Line-SBLOCA accident is induced to analyze the influence of flow blockage on subsequent CAP1400 reactor core heat transfer and core degradation. From the results of the DVI Line-SBLOCA accident analysis, it can be concluded that the blockage ratio is in the range of 40%–60%, and the position of severe blockage is the same as that of cladding rupture. The blockage reduces the circulation area of the core coolant, which in turn impacts the heat exchange between the core and the coolant, leading to the early failure and collapse of some core assemblies and accelerating the core degradation process.

Simulation of high temperature fuel rod bundle behavior during the QUENCH-06 experiment using CINEMA code

The CINEMA code has been developed to secure the ability to analyze severe accidents in nuclear power plants with a Korean domestic computer code. As part of the code validation, the current work simulated the QUENCH-06 experiment with the CINEMA. Calculation results from CINEMA were compared to the experimental results and numerical results of SPACE and SCDAP/RELAP5. CINEMA overestimated the bundle temperatures near the test section entrance, as observed in the SPACE and SCDAP/RELAP5 calculation due to an underestimation of the thermal conductivity of argon-filled ZrO2 fiber insulation. The hydrogen generation rate was overestimated by the CINEMA because of the high bundle temperature at the upper part, where CINEMA is unable to model extreme heat loss through the shroud without insulation. The CINEMA results demonstrated good agreement with the experimental data, validating that the CINEMA is capable of simulating high temperature fuel rod behavior before and during the reflood phase.

Status of benchmark exercises in the frame of the NUGENIA QUESA project

Severe accidents in nuclear power plants with loss of cooling or loss of coolant and core uncovery can lead to air ingress into fuel assemblies. If air or a steam–air-mixture comes into contact with cladding material, the exothermal character of zirconium oxidation and nitride formation leads to enhanced cladding oxidation, heat-up and degradation up to core melt and fission product release. Two experiments, QUENCH-18 and CODEX-AIT-3 were performed in the frame of EC projects ALISA and SAFEST to fill knowledge gaps on relevant phenomena. Both tests are accompanied by pre- and post-test benchmarks within the NUGENIA project QUESA. The pre-test benchmarks were completed, while the post-test benchmarks are currently running. Results of the pre-test benchmarks indicate that participating codes can qualitatively predict the experimental observations. Generally, the range of the predictions are wider for CODEX-AIT-3 while the trend is close except one for QUENCH-18. The specification of both test procedures was additionally supported by parameter studies. First post-test benchmark results for QUENCH-18 show that code predictions are in good agreement during the pre-oxidation and mixed steam/air phase. In the subsequent reflooding phase with intensive melt release, relocation and oxidation, notable differences between codes and simultaneously deviations from experimental results are observed. In the paper the main findings of the QUENCH-18 test are discussed, but the focus will be on the pre-test analyses of CODEX-AIT-3.

Experimental investigation on quenching behavior during reflooding in a 3 × 3 dual-cooled annular rod bundle

The internally and externally cooled annular fuel is an innovative fuel geometry proposal for advanced PWR, which could provide a substantial increase of power density while maintaining or improving safety margins. The quenching behavior of annular fuel during reflood phase in LOCA is more complicated than cylindrical solid fuel, owing to the geometrical difference and the coupling effect of dual-cooling. However, the investigation focusing on quenching characteristics in annular fuel geometry is very little. In the present study, an experimental study on quenching behavior of bottom reflood in a 3 × 3 dual-cooled annular rod bundle was carried out under various test conditions (reflood velocities from 0.025 to 0.15 m·s−1, inlet subcooling from 10 to 80°C, peak cladding temperature from 300 to 900°C, and linear power density from 0 to 1.58 kW·m−1). Self-designed indirect-heating annular rods were employed to provide an axial uniform power distribution in this experiment. It was found that the quenching process on external and internal surfaces of annular rod bundle was almost synchronous. The parametric effects on quenching behavior, i.e. precursory cooling rate, quench front propagation velocity (Vqf) and minimum film boiling temperature (Tmin), were investigated in detail. Earlier occurrence of quenching downstream the grid spacer was also observed under low inlet subcooling and high peak cladding temperature conditions. In addition, several correlations for Tmin were assessed against the present results. Kim correlation was in good agreement with the present data, with the mean absolute errors of 7.81% and 7.55% for outer and inner channels respectively.

Physics-informed machine learning-aided framework for prediction of minimum film boiling temperature

Minimum film boiling temperature (TMFB) is a crucial parameter in post-critical heat flux (CHF) condition by determining the collapse of stable vapor film on overheated surface. Although many correlations have been suggested with a wide range of experimental data to predict the TMFB considering various parameters, they have failed to accommodate the universal quenching data due to limited regression capability addressing effects of multivariate parameters. Even though several data-driven machine learning models (DDML) showed improved prediction performance, their performance are dramatically degraded in extrapolation conditions, which are unincluded in the training process. To overcome the inherent problems of conventional correlations and DDMLs, physics-informed machine learning-aided frameworks (PIMLAF) for TMFB were developed in this study by combining the conventional correlations (Henry model and Groeneveld-Stewart correlation) and machine learning techniques (multi-layer perceptron and random forest) to overcome their limited regression capability and ‘black-box’ characteristics. For the interpolation condition, it was observed that the Groeneveld-Stewart correlation provides better baseline for TMFB than Henry model, and random forest (RF) model has advantage discovering the pattern between TMFB and input variables with tuning capability. Through benchmark studies corresponding to extrapolation condition of the training dataset, it was concluded that hybrid approach of RF model and Groeneveld-Stewart correlation exhibits the best performance on TMFB prediction for both interpolation and extrapolation conditions compared to standalone machine learning models, Groeneveld-Stewart correlation, and MLP-based PIMLAF. The suggested RF-based PIMLAF is expected to enhance the TMFB prediction performance during reflood phase of loss of coolant accidents in light water reactors.

Study on the effect of flow blockage due to rod deformation in QUENCH experiment

During a loss-of-coolant accident (LOCA) in the pressurized water reactor (PWR), there is a possibility that high temperature and internal pressure of the fuel rods lead to ballooning of the cladding, which causes a partial blockage of flow area in a subchannel. Such flow blockage would influence the core coolant flow, thus affecting the core heat transfer during a reflooding phase and subsequent severe accident. However, most of the system analysis codes simulate the accident process based on the assumed channel blockage ratio, resulting in the fact that the simulation results are not consistent with the actual situation. This paper integrates the developed core Fuel Rod Thermal-Mechanical Behavior analysis (FRTMB) module into the self-developed severe accident analysis code ISAA. At the same time, the existing flow blockage model is improved to make it possible to simulate the change of flow distribution due to fuel rod deformation. Finally, the ISAA-FRTMB is used to simulate the QUENCH-LOCA-0 experiment to verify the correctness and effectiveness of the improved flow blockage model, and then the effect of clad ballooning on core heat transfer and subsequent parts of core degradation is analyzed.

A three-dimensional PWR LBLOCA simulation using the CUPID code

As the need for simulations of multi-dimensional behavior in the safety analysis is increasing, efforts have been devoted to developing a three-dimensional module for the conventional system analysis codes. After having developed a one-dimensional system analysis code, MARS, the Korea Atomic Energy Research Institute (KAERI) began to develop the CUPID code to address the need for a multi-dimensional analysis. Recently, the reactor vessel module, such as a package of interfacial heat and momentum transfer model, wall friction and wall-to-fluid heat transfer model, were implemented in the CUPID and named CUPID-RV module. The final goal of this study was an analysis of the Large Break Loss of Coolant Accident (LBLOCA) at a rod-by-rod level, and the study focused on the validation of LBLOCA thermal-hydraulics phenomena and the application to a three-dimensional LBLOCA simulation. In this paper, we introduce the validations using Edward pipe, UPTF, and FLECHT-SEASET experiments related to the LBLOCA blowdown, refill, and reflood phases and a three-dimensional LBLOCA simulation for the APR1400 PWR reactor pressure vessel using simple boundary conditions at an assembly-by-assembly level as an application.

Improvement of the heat transfer enhancement model considering the droplet-wall heat transfer downstream of the flow blockage in the reflood phase

When a flow blockage is formed by fuel swelling in the reflood phase of a large-break loss-of-coolant-accident (LOCA) in a pressurized water reactor (PWR), flow separation and reattachment occur due to the flow area change downstream of the blockage. These phenomena increase turbulence and vorticity in the downstream, resulting in the enhancement of heat transfer by steam and liquid droplets. The existing heat transfer enhancement model has considered the enhancement factor for the single-phase heat transfer only. For the assessment of the existing model, it was implemented in the CUPID code in which a two-fluid three-field model is applied. The existing model was evaluated using the FEBA flow blockage test without flow bypass. As a result, the existing model greatly over-predicted the rod temperature downstream of the blockage. To improve this, a new enhancement model was proposed that considers the wall heat transfer by the droplet deposition. The heat transfer enhancement factor was derived based on the data from the abrupt pipe expansion experiments that measured heat and mass transfer coefficients after flow separation point. For accurate calculation of the amount of droplets approaching the fuel rod, an improved droplet deposition model was developed using the droplet deposition experimental data for dispersed droplet and annular-mist flow. The new model implemented in the CUPID code was evaluated using the FEBA flow blockage tests with different pressure, flooding rate, and blockage ratio. It was shown that the new model gave good predictions for the temperature of the early cooled rod bundle downstream of the blockage.

A sensitivity study of physical models using in RELAP5 code based on FEBA experimental data

In the thermal-hydraulic safety analysis, simulation results using thermal-hydraulic codes depend mainly on modeling the physical phenomena built-in the codes. These models are the equations, and empirical formulas developed based on matching them to experimental data or based on the assumptions, simplifications for solving theoretical equations. Therefore, it is recommended that these physical models need to take into account the uncertainty they cause. The sensitivity study is performed to investigate the influence of physical models on the calculation results during the reflood phase after the loss of coolant accident. It is allowable to choose the physical models that have the most significant influence on the calculation results. This study conducted a sensitivity analysis of physical models in RELAP5 code based on experimental data measured on the FEBA test facility. Sixteen physical models have been selected for sensitivity analysis to find the most important models that influence the calculation results. Based on two criteria, the maximum cladding temperature and the quench time, the sensitivity analysis results show that four physical models significantly impact the calculation result. Four chosen physical models are considered further in the next step of their uncertainty evaluation.

Velocity field and flow redistribution in a ballooned 7×7 fuel bundle measured by magnetic resonance velocimetry

During a loss of coolant accident (LOCA), blocked sub-channels may appear due to the swelling of the fuel rods’ cladding, which results in flow redistribution during the reflooding phase. For this reason, special attention has been paid to the effect of fuel rods ballooning on the thermal-hydraulics in LOCA conditions. Due to the practically impossible physical or optical access to blocked sub-channels, no experiment so far has performed precise three-component velocity field measurements in the presence of ballooned regions. In this study, we used magnetic resonance velocimetry (MRV) to obtain three-component velocity fields of water flow within two 7 × 7 fuel rods bundles built mainly in plastic, one regular and one containing sixteen ballooned fuel rods with 90% blockage ratio and 240 mm blockage length. We present herein results with 50 lpm water flow rate, which corresponds to a Reynolds number of 1936. With the regular bundle, the performance of spacer grids’ mixing vanes to homogenize the flow was notable. With the ballooned bundle, we observed transverse velocities upstream of the ballooned zone that are as intense as the bulk mean velocity. Furthermore, there are substantial decreases in the axial velocity within blocked sub-channels up- and downstream of the ballooned zone, reaching near-zero and even negative values downstream, indicating flow recirculation. Although the flow is highly affected by the ballooned zone, the mixing spacer grid placed downstream remarkably homogenized the flow and effects of the flow redistribution disappeared. Finally, with the present ballooned bundle configuration, about 90% of the flow that should pass through blocked sub-channel deviates towards less resistant regions, which suggests a predominant geometric effect on the flow redistribution.

Preliminary study of uncertainty qualification methods based on the simplified LB LOCA model for PCT estimation

Uncertainties for predictions of thermal hydraulic problem are caused by the complex physical origin. Uncertainty qualification methods are designed to meet the requirement of regulation to ensure acceptable safety margins with enough confidence. Various methods are proposed for the industrial applications to achieve the proposed level of probability and confidence. However, it is not enough to obtain more statistics information with limited calculation by the full evaluation model. The present paper is based on a simplified model of large break loss of coolant accident for peak clad temperature (PCT) estimation (Catton et al. (1990)). The simplified model is based on physical foundation with sufficient uncertainty parameters and trend indication. The higher PCT during blow down phase and reflooding phase was set to be the target output parameter. The numerous parameters and distribution in the simplified model were set to the source of input uncertainty. The effects of PIRT, sample number, uncertainty distribution were studied. The results provided qualitative and quantitative results for uncertainty methods and would be beneficial for the determination of uncertainty qualification methods in engineering design.

Modeling of the droplet entrainment rate in the post-dryout regime for the analysis of a reflood phase

Many droplets are generated during the reflood phase of a large-break loss-of-coolant accident (LOCA) in a pressurized water reactor (PWR). The droplets greatly contribute to the cooling of the reactor core and, thus, have a great influence on the temperature behavior of the fuel rods. This requires an accurate model for droplet entrainment rate, especially for the reflood phase. However, the existing droplet entrainment models in thermal-hydraulic system codes cannot properly capture the entrainment phenomena in the post-dryout regime during the reflood process. In this study the observation results of visualization experiments for the post-dryout regime were utilized for modeling the droplet entrainment rate. Through the experiments, it was confirmed that droplets are mainly detached from the core liquid jet and the liquid sheet. We suggest a new droplet entrainment rate model based on the observation results and instability analysis. The new model is implemented in the CUPID code and evaluated using a set of the FEBA and the FLECHT SEASET reflood experiments. The results are compared with those of the existing droplet entrainment models. It was shown that the new model predicts the peak clad temperature (PCT) and the quenching time(QT) very well.

Effects of newly developed entrainment model for a thermal hydraulic system code (SPACE) based on breakup of large droplets in reflood

The accurate prediction of the fuel rod temperature from a peak cladding temperature to quenching is required for a thermal–hydraulic system code in the reflood phase within a pressurized water reactor. Although conventional codes have well predicted the wall temperature trends, inaccuracies with the hydraulic parameter such as the steam flow rate still remains. To predict reasonable results that reflect an actual phenomenon, we implemented the breakup of a large droplet generated in an inverted slug flow. We suggested a new method to estimate the entrainment for an Eulerian system code (SPACE). The new method was successfully verified and compared to the analytic results. After the modification, we obtained improvements on the steam flow rate, the wall temperature, PCT times, and quenching times for FLECHT-SEASET. For RBHT, it showed mixed results with a faster decrease of the wall temperature, which implies the possibility of a transition for the reflood phenomena.

Wall heat transfer in the inverted annular film boiling regime

Inverted annular film boiling (IAFB) is one of the post critical heat flux (post-CHF) regimes, which may occur during the core reflooding phase following large-break loss of coolant accidents in water-cooled nuclear reactors. This paper first analyzes wall heat transfer characteristics and summarizes convective wall heat transfer coefficient correlations developed for the IAFB regime in the literature. Parametric effects of the inlet subcooling, mass flux, and pressure on the wall Nusselt number in the IAFB regime are studied. It is found based on the experimental data in the literature that the system pressure has negligible effect on the wall Nusselt number while the effects of the mass flux and inlet subcooling are complex and depend on the flow conditions. The modified Bromley model and laminar flow model with an enhancement factor for the IAFB regime, currently used in COBRA-TF and TRACE, respectively, are compared with Stewart’s experimental data under subcooled and low-quality conditions in the IAFB regime. Both of these two models under-predict the heat transfer coefficients calculated from Stewart’s experimental data for high flow and high subcooling conditions. Based on the analysis of the parametric effects, two new wall heat transfer correlations are proposed by correlating the Nusselt number and non-dimensional vapor film thickness using Stewart’s experimental data and benchmarked against the experimental data in the literature. These two new correlations cover ranges of the system pressure from 2.0 to 9.0 MPa, mass flux from 215 to 923 kg/m2-s, and inlet subcooling up to 50 °C.

Emergency core cooling system performance criteria for Multi-Layered Silicon Carbide nuclear fuel cladding

Multi-layered Silicon Carbide (SiC) cladding has been proposed to replace the current Zircaloy (Zr) cladding in existing nuclear power plants. Due to its ceramic property, SiC may have a different failure mechanism from that of metallic Zr clad. This characteristic should be considered in designing the plant’s safety system performance, such that the fuel’s integrity is reasonably assured during accident scenarios.This paper aims at optimizing the Emergency Core Cooling System (ECCS) performance criteria in mitigating LBLOCA accidents at a Pressurized Water Reactor (PWR) using the uranium dioxide fuel and duplex or triplex SiC cladding structures. Clad failures by brittle fractures and thermal degradation were modeled in a fuel analysis code. This fuel model was coupled with the RELAP5 plant model by adapting the fuel’s mesh configuration and formulating the effective heat transfer coefficients. A model to quantify the likelihood of clad functional failure as a combination of individual layer fractures at various axial positions was proposed. Updated SiC material properties were gathered from recent reference studies. Clad reliability estimates using the proposed model were compared with the estimates obtained using other models. The results were used to relax the criterion of required actuation timing of ECCS, such that the risk of the new plant design is less or equal to that of the initial plant configuration.Triplex clad was more reliable than duplex clad and Zr, reducing the conditional core damage probability by three orders of magnitude in the reference LBLOCA scenario. Compressive stresses due to irradiation swelling played the determining role in the vulnerabilities of each clad design. The duplex clad was vulnerable to pressure-loading stresses while triplex clad was prone to fail due to thermal shocks experienced during the LOCA reflood phase. Additionally, it was concluded that the most vulnerable axial fuel rod location in the reference LBLOCA scenario was located at the highest axial power peaking location.A sensitivity study on the ECCS timing density was conducted. The results showed that ECCS actuation timing could be extended from the initial technical specification, without increasing the fuel damage likelihood. The time extension was distributed in the order of several hundreds of seconds, which was comparable with time extensions offered by FeCrAl and Cr-coated-Zr clad designs in LBLOCA scenarios. This time extension may potentially reduce the failure probabilities of EDGs and pumps, hence improving the plant safety further.