Annals of Nuclear Energy最新文献

Laser induced fluorescence (LIF) is characterized as a non-insertion and whole-field measuring technology, which is commonly used in the thermal–hydraulic experiments. The conventional method for LIF technology is performed on the basis of the theoretical correlations for the fluorescence. A new method for the LIF base on the calibration tests and empirical fitted polynomial was proposed. The mock refueling water tank and mock reactor pressure vessel (RPV) were employed to verify the new method. The precise comparison between the conventional method and new method was carried out. The results indicated that the deviation in temperature measurement was within 0.5 °C for the improved method, and it was within 1.0 °C for the conventional method. The concentration measurement error was within 5.67 % for the improved method, and it was about 7.10 % for the conventional method. The improvement in measurement accuracy brought by new methods of great importance for the thermal–hydraulic experiments.

The main aim of the research was to design, implement and investigate an Artificial Neural Network (ANN) to predict the behavior of selected parameters of a nuclear reactor core. The studied core was a typical power-generating Pressurized Water Reactor (PWR). The PARCS v3.2 nodal-diffusion core simulator was used to generate training and validation data. The ANN was implemented using Python 3.8 code with Google’s TensorFlow 2.0 library. The effort was based to a large extent on the process of automatic transformation of generated data, which was later used in the process of the ANN development. Various ANN architectures were studied to obtain better accuracy of prediction. In this study, a special focus was put on the prediction of the fuel cycle length for a given core loading pattern. In addition, a conversion of the input data was applied, allowing for very good accuracy of the cycle length prediction ($>99$%).

This paper introduces the updated emergency core cooling scheme for Tsinghua High Flux Reactor. The corresponding Relap5 model is established and key designs are studied by simulating the corresponding loss-of-coolant accidents. These key designs specifically involve ACC internal pressure, ACC coolant temperature, ACC liquid fraction, ACC location, emergency pumps, check valves at pressure vessel inlets, the design of pressure vessel outlet, connection between the dry pool and the reactor pool, and operation pressure. Through analysis, some general conclusions are obtained, which can be used to guide the design of emergency core cooling system in the near future. Critical heat flux is predicted by Sudo correlations during loss-of-coolant accidents. According to minimum departure from nucleate boiling ratio, core safety margin is evaluated. By studying these key designs, the current scheme can be better understood and more safety margins can be found.

SMART, a conceptual small modular reactor (SMR), uses uranium dioxide (UO₂) as fuel but is prone to damage in high-temperature scenarios like station blackout (SBO) accidents. This study analyzes the neutronic aspects of SMART with alternative fuels: UN (natural N), UN (enriched ¹⁵N 99 %), UB₂ (enriched ¹¹B 100 %), and U₃Si₂ using MCNP6 with the ENDF/B-VII.1 nuclear data library. The ¹⁵N 99 % enriched nitride fuel has the highest k_eff while improving safety by reducing radial power fraction and improving core cycle length with higher fissile content. Enriched ¹⁵N nitride fuel, boride fuel, and silicide fuel are viable substitutions for oxide fuel. Reducing the enriched ¹⁵N level in nitride fuel can manage excess reactivity at the beginning of the cycle. Despite different burnup levels, the neutron flux distribution, radial power peaking factor, and effective delayed neutron fraction (β_eff) show minimal variation among the fuel types.

Fuel performance codes, such as Transuranus, predict fuel behavior and are used to ensure the safe operation of nuclear reactors. These codes are moderately time-consuming and affordable in many applications but may be limited in others, primarily when many fuel rods must be evaluated simultaneously. This work presents how the temporal neural network techniques, Temporal Convolutional Networks, and a Fourier Neural Operator can be combined to form a deep heterogeneous joint architecture as a surrogate model for fuel performance modeling in time-critical situations. We train the model using realistic power histories and corresponding outputs generated using the fuel performance code Transuranus. The ultimate result is a surrogate model for use in time-critical situations that take milliseconds to evaluate for thousands of fuel rods and have a mean test error of unseen data around a few percent.

In nuclear power plants (NPPs) operations, the prediction of multi-dimensional parameters is found to help operators to grasp the condition of the system. However, majority of existing studies are focused on single-dimensional parameter prediction. In this study, a multi-dimensional parameter prediction framework of NPPs based on Long Short-Term Memory Network and Graph Convolution Network (LSTM-GCN) and a multi-model integrated parameter correlation analysis framework (PCAF) are proposed, in which PCAF is used to build a parameter correlation network for GCN, and LSTM-GCN is used to predict multi-dimensional parameter of NPPs. To verify the feasibility of the LSTM-GCN framework, multi-dimensional parameter prediction researches are conducted using data from a thermohydraulic experimental bench that simulates the operation of NPPs. Results indicate that compared to traditional prediction models, LSTM-GCN framework enhances the prediction accuracy of multi-dimensional parameter, which benefits from the ability of LSTM-GCN to utilize the temporal dependencies and spatial correlations of parameters.

The iterated fission probability (IFP) method enables assessment of adjoint flux-weighted kinetic parameters, i.e., effective kinetic parameters, in Monte Carlo (MC) simulation, an essential capability in modern MC codes. This method can be extended to calculate adjoint flux-weighted quantities within a prescribed phase-space, enabling the estimation of adjoint flux distributions. The iMC Monte Carlo code, developed at the Korea Advanced Institute of Science and Technology (KAIST), is proficient in both calculating effective kinetic parameters and adjoint flux distributions. This paper presents benchmark results verifying the code’s capabilities. Critical device configurations are considered for evaluating kinetic parameters, compared with the Serpent2 code results. Both multi-group and continuous-energy benchmarks are solved to assess IFP-based spatial- and energy-wise adjoint flux distributions, and comparison is made against deterministic transport calculations. Results show that effective kinetic parameters can be accurately estimated, and acceptable adjoint flux distributions can be obtained using the iMC code.

Due to the structural characteristics of the helical tube, centrifugal force complicates the flow boiling phenomenon inside the tube. How to accurately describe and predict the parameter distributions involved in helical tube flow and heat transfer is a concern for scientists. Based on the OpenFOAM, this paper combines the wall boiling model, the interphase model, other closed models with the Eulerian two-fluid model, analysed changes of void fraction, surface heat transfer coefficient, pressure drop with working conditions in the single-phase to nuclear boiling section in the tube. The results show that this solver and the corresponding empirical relational model have the ability to accurately simulate the boiling of the flow in helical tube; In nuclear boiling section at the working conditions of P=4–8 MPa, q = 200–350 kW/m², Re = 66827–89103, The degree of gas-phase buildup near the inner wall surface of the spiral tube decreases with the increase of Re number, increases with the increase of heat flux and pressure, and the ratio of friction pressure drop to total pressure drop decreases with the increase of Re number, heat flux, and pressure by a maximum of 1.4 %, 4.26 %, and 17.35 %. This paper can provide a reference for adding new models and developing new solvers in the OpenFOAM to simulate boiling in helical tube flows.

Molten salt reactors (MSRs) are Generation IV nuclear reactor concepts gaining significant research interest worldwide. The present work proposes a methodology for the definition and optimization of the MSR operating conditions and control strategy. The proposed methodology employs a novel 0D steady-state model, the multi-physics system-scale MOSAICS (MOlten SAlt Incompressible Calculation System) model and a Global Sensitivity Analysis (GSA) method based on Hilbert-Schmidt Independence Criterion indices. The methodology was then applied to the ARAMIS (Advanced Reactor for Actinides Management in Salt) fast-spectrum chloride-salt burner reactor. Two alternative control strategies are proposed in order to achieve target margins to the salt freezing and structure material limit temperatures during normal operation, plus a 20 % power variation per minute load-following objective. The performance of the control strategies was assessed through comparison with the natural behavior during a load-increasing transient, as well as during Unprotected Transient Over-Power (UTOP) and Station Blackout (SBO) accidents. Controlled load variation transients are observed to reduce overall temperature variation rates throughout the salt circuits, while the inclusion of a variable fuel flow from the reactor commands can further limit these fluctuations. However, the selected operating conditions exhibited insufficient margins to freezing or the materials limit temperature under unprotected accident conditions. GSA enabled the derivation of correlations to characterize the dynamic response of MSRs to the accidents. Based on the observed reactor response to the various transients, potential modifications to the ARAMIS design and control strategies are proposed.