Space nuclear reactors employing heat pipes and rotating drums offer a compact, mechanically simple architecture for multi-megawatt power systems, yet their conceptual design is tightly constrained by coupled neutronic behavior, shutdown-safety requirements, heat-transport limits, and geometric integration considerations. This work develops an integrated safety-physics modeling and multi-objective optimization framework for a space heat-pipe reactor (SHPR) to simultaneously maximize deliverable thermal power and minimize total system mass under rigorous feasibility screening. The evaluation framework integrates: (i) an approximate neutronics model with a drum-based shutdown formulation encompassing two distinct shutdown conditions; (ii) steady-state heat-pipe operating-limit model accounting for capillary, sonic, and other transport constraints; and (iii) a geometry-mass model that enforces drum-installation clearances and determines reflector sizing consistent with criticality requirements. A dedicated multi-objective optimizer (MOPIMO) is benchmarked against MOEA/D, NSGA-III, MOPSO, and MOTOC under identical evaluation budgets and multiple random seeds, using hypervolume (HV) and inverted generational distance (IGD) to assess convergence, diversity, and robustness. Under a fixed computational budget, MOPIMO shows superior convergence and diversity of the feasible Pareto set (highest HV and lowest IGD), improving the robustness of trade-off exploration in the proposed conceptual design workflow. For reporting purposes, we select a non-endpoint compromise solution from the feasible Pareto set (a centrally located point identified using crowding distance in the objective space), yielding 8.203 MWth at a total mass of 4261.7 kg. The selected compromise design satisfies the shutdown criteria for both the nominal shutdown scenario (S0) and the design-basis single-control-drum failure shutdown scenario (S1) and respects all heat-pipe operating limits; the binding constraints are explicitly identified to indicate the governing feasibility mechanisms. The consistency among multi-algorithm fronts, HV/IGD statistics, and physically interpretable boundary solutions demonstrates the soundness and practical relevance of the proposed framework and provides a reproducible reference configuration for engineering applications.

-Accurate open channel flow measurement is essential for managing water resources, optimizing irrigation, and forecasting floods. Existing soft sensing methods have limitations, including an over-reliance on stage-discharge relationships, oversimplified 3D velocity distributions, and weak integration between Computational Fluid Dynamics (CFD) and neural networks. To address these issues, this paper presents a new method that combines CFD simulations with Machine Learning (ML). The study focuses on wide, shallow trapezoidal channels. Highresolution flow field datasets are generated using CFD numerical simulations, which are built and calibrated with field-measured data from China's Renmin Canal. An error correction mechanism that integrates the standard k-epsilon and Renormalization Group (RNG) k-epsilon turbulence models reduces the mean absolute percentage error (MAPE) to 0.66%. Based on the characteristics of vertical velocity distribution, key features-including velocities at the inner zone (y/H = 0.2), outer zone (y/H = 0.6), and surface (y/H = 1.0)-are selected as inputs for a back propagation (BP) neural network. This network is optimized using the golden jackal optimization (GJO) algorithm. Tests using data from the Renmin Canal and validation against ten channels in the Qiongxia Irrigation Area show that the GJO-BP model delivers better prediction accuracy (test set MSE: 0.805) and greater robustness (test set MAPE: 2.039%) than conventional Genetic Algorithm (GA)-BP and Particle Swarm Optimization (PSO)-BP models. The proposed solution is cost-effective and highly precise, reducing the number of required field measurement points by more than 60%.

Cable harness routing is a critical component in the layout design of complex electromechanical systems in nuclear power plants, where its accuracy and safety directly impact the operational reliability of nuclear island buildings. However, current mainstream metaheuristic algorithms suffer from inherent limitations in solution quality control and physical interpretability. Meanwhile, traditional edge-driven multi-commodity network flow integer programming models face computational infeasibility due to the combinatorial explosion of decision variables, particularly struggling to meet the stringent engineering constraints of radiation shielding, safety clearances, and hierarchical routing in nuclear power scenarios. To address these bottlenecks, this study proposes a node-driven Binary Integer Linear Programming model (BILP) combined with a Problem Decomposition and Workspace Approximation Algorithm (PD&WA), achieving a balance between mathematical rigor and computational feasibility for nuclear power cable routing problems. Validation through industrial-scale nuclear island building scenarios in nuclear power plants (involving 10 groups of cables with different voltage classes and protection categories) shows that under equivalent time budgets, the proposed strategy reduces the number of decision variables by 99.6 % compared to the original model. The objective function values are decreased by 81.9 %, 51.2 %, and 79.5 % relative to the original model, literature benchmarks, and metaheuristic algorithms, respectively. Theoretical analysis and experimental data confirm that the near-globally optimal solutions generated by this framework within finite time can be directly applied to nuclear power cable layout design, establishing a novel paradigm that integrates mathematical rigor and computational feasibility for highdimensional constrained path planning problems in nuclear reactor systems.

Acoustic thermometry is widely used in engineering temperature monitoring, but sparse acoustic rays caused by constrained transducer deployment hinder its high-precision upgrade. This study aims to address the core challenge of high-fidelity temperature field reconstruction under sparse measurement conditions. A novel method integrating sparse dictionary learning and physical constraints is proposed: the sparse dictionary captures spatial structural priors of temperature fields, while physical governing equations embed temporal evolution laws, jointly breaking through data dimensionality limitations. Simulation and practical verification (nuclear power plant eight-channel ultrasonic device) show that the method achieves excellent spatiotemporal consistency, accurately reproducing time-varying temperature field evolution; validated with actual nuclear power plant data, the proposed method improves the accuracy of the output average temperature by 2-7 K compared with existing methods. This work provides a new technical approach for high-precision temperature measurement under sparse conditions, enhancing thermal parameter safety monitoring in complex engineering scenarios.

Seismic signals generated by granular geohazards (e.g. rock avalanches, debris flows) contain critical information for hazard assessment and early warning systems. A key challenge is the real-time inversion of grain size distribution (GSD), which governs flow dynamics and impact mechanisms. This study develops a novel high-frequency (≥1 Hz) seismic model for falling granular geohazards, combining elastic impact theory, Green’s function and energy dissipation from bed deformation and inter-granular collisions. Model validation is conducted through laboratory experiments involving single-grain impacts, dual-grain impacts and multi-grain assemblies with varying GSDs, supplemented by field observations. Results of single-grain impacts with diameter d show that the centroid frequency (fcen), elastic energy (Wel) and total power spectral density (PSDT) of seismic signals scale proportionally with d-0·5, d5, d3, respectively. These scaling laws are further supported by field observations. Comparative analysis of different GSDs reveals that conventional effective diameters exhibit limited correlations with seismic signal characteristics. In contrast, the newly proposed GSD parameter GSDe, which incorporates fractal dimension (Df) and effective diameter, demonstrates robust and scale-independent relationships with seismic signals. The new theoretical model shows strong agreement with measured seismic signals, outperforming existing physical models in prediction accuracy. The findings provide a theoretical and practical framework for real-time GSD inversion from seismic signals, with direct applications in geohazard monitoring and mitigation strategies.

Acoustic tomography (AT) is a non-contact technique for reconstructing internal temperature distributions from the relationship between sound velocity and temperature. However, conventional static reconstruction methods based on synchronous sampling suffer from the failure of the “frozen field” assumption in unsteady temperature fields and cannot adequately handle the spatiotemporal asynchrony of multipath signals. To address this problem, this paper proposes a spatial-feature-learning-based reconstruction framework for dynamic temperature fields. The method organizes transit-time data into tensor form, models acoustic propagation paths using graph structures, and employs non-negative matrix factorization to decouple spatiotemporal features. A graph-specific spatial-basis learning mechanism is further introduced to extract reusable low-rank features under fixed transducer layout and fixed path correspondence, enabling frozen reuse across unseen temperature-field backgrounds within the same graph structure. Numerical simulations show stable reconstruction performance across multiple dynamic moments, with overall temperature-field errors maintained within 2%–4%. In a field case using three valid TOF frames from the Chongqing Qineng Power Plant boiler furnace, the thermocouple references were located 500 mm below the acoustic measurement plane. The paired field-test records indicated an apparent non-coplanar reference discrepancy of 0.86% ± 0.16%. Under this limited reference configuration, the proposed method produced physically plausible temperature fields and lower RMS discrepancies than comparison methods, supporting preliminary engineering feasibility rather than strict full-field accuracy validation.

The scale of mapping units significantly affects the accuracy and reliability of landslide susceptibility assessment. However, existing landslide susceptibility studies lack a clear determination of the appropriate slope unit scale, and the impact of different slope unit configurations on the modeling process and model interpretability has not been thoroughly investigated. This study conducted an empirical analysis using extensive real-world landslide data from the core area of the Three Gorges Reservoir region, comprehensively investigating the effect of slopeunit scales on the landslide susceptibility assessment. Initially, a geospatial dataset comprising 3594 historical landslide events and 22 initial condition factors was compiled. Subsequently, 30 different slope unit schemes of varying scales were generated by the r.slopeunits tool. For each scheme, the dataset was randomly divided into training and testing subsets with a 7:3 ratio and modeled using random forest model. This study reveals the significant impact of slope unit scales on hyperparameter optimization, factor selection, and model interpretability. The results highlight that: (1) appropriate slope unit scale can improve the quality of input variables, thereby enhancing the generalization ability and interpretability of landslide susceptibility assessments, reducing the risk of overfitting. (2) Finer and more concentrated slope units do not always lead to better results; they may excessively rely on distance metrics, resulting in overly conservative high susceptibility classifications in landslide susceptibility models. This study provides valuable insights into selecting the appropriate slope unit scale for landslide susceptibility assessment.

The reactor coolant flow of the primary loop is one of the key thermal-hydraulic parameters in nuclear reactor operation, and its measurement accuracy directly relates to the safety and stability of the plant. In pressurized water reactor (PWR) nuclear power plants, elbow flow meters are typically used to measure the coolant flow. However, the transition section of the primary loop pipe contains a non-uniform, highly dynamic flow-heat coupling field, leading to certain uncertainties in coolant flow measurement. To address the complex operating conditions at the transition section of the steam generator outlet, this study constructs an orthogonal curvilinear coordinate system adapted to the elbow geometry and, based on tensor analysis and the Navier-Stokes equations, derives a Radial Pressure Gradient Equation (RPGE). A mechanistic analytical framework is established to identify and decompose the sources of wall pressure difference in elbows and to evaluate their contributions to measurement uncertainty quantitatively. CFD numerical simulations are further conducted to validate the applicability and computational accuracy of the proposed model. Results indicate that under the uniform-density flow assumption, discrepancies between RPGE predictions and CFD results remain within 0.15 % across all investigated operating conditions. Source-term decomposition reveals that the convection term and the primary flow term constitute the dominant contributors to uncertainty, each accounting for approximately 11 % of the wall pressure difference. Nonetheless, these two contributions partially cancel each other numerically, resulting in a total uncertainty consistently maintained at less than 1 %. Under variable-density flow conditions, discrepancies between RPGE and CFD results remain within 2 %. The primary impact of the nonuniform temperature field is the increased dispersion of the quantified uncertainty intervals of individual source terms. Compared with the conservative empirical uncertainty range typically adopted in engineering practice (-3 %-3 %), the uncertainty intervals derived from the RPGE framework are reduced by approximately 35 %-40 % on average across different operating conditions. The proposed analytical approach provides an interpretable theoretical basis and a systematic quantitative tool for tracing and evaluating uncertainty in elbow flow meters operating under complex flow-heat coupling environments.

Sensors are critical components for information perception and operational monitoring in nuclear power plants, and their operating conditions are directly related to system safety and reliability. Existing sensor health monitoring methods based on prediction residuals often struggle to learn robust temporal representations when abnormal samples are scarce, making them susceptible to anomalous signal propagation and degraded fault detection and localization performance. To address this issue, this paper formulates robust prediction under sparse fault perturbations as an optimization problem in open fault scenarios and proposes a Variable-Aware Temporal Variational Autoencoder (VAT-VAE) prediction framework. Through the collaborative design of adaptive input modulation and structured perturbation training, the proposed model suppresses the influence of abnormal input channels and enhances robustness against sparse fault disturbances, thereby generating stable health reference signals. Experimental results based on real nuclear power plant operational data and simulation datasets demonstrate that the proposed method outperforms existing baseline models in signal reconstruction robustness, multi-sensor fault detection, and fault localization.

Rapid and accurate prediction of the sectional velocity field of the channel is of great significance to the design and maintenance of open channels and the improvement of irrigation efficiency. During the water delivery process of Renmin Canal of Dujiangyan irrigation system, the water level of the main canal changes rapidly and in a large range, which is the biggest difficulty in real-time prediction of its velocity field. Therefore, based on machine learning, this paper proposes a new method to construct a real-time velocity field prediction model, which can directly predict the velocity field of the channel according to the water level. According to this method, the computational fluid dynamics (CFD) technology is used to simulate the target open channel, and a machine learning model that can adaptively optimize the characteristics of the velocity field data is designed as the velocity field prediction model, which is experimented in the main canal of Renmin Canal of Dujiangyan irrigation system. The results suggest that the predictions are in line with the general features of flow velocity distribution in open channels and have high precision. Therefore, this method is of high value for engineering application and theoretical research.

Condition monitoring is essential in industrial processes to ensure safe and efficient operations. Sensor signals, which accurately reflect the state of industrial systems, play a central role in this monitoring. However, the harsh conditions in many industrial environments, especially in nuclear power plants, increase the likelihood of sensor failures. Condition monitoring systems detect anomalies by reconstructing input data, with high reconstruction errors indicating the presence of anomalies. The Multivariate State Estimation Technique (MSET) is a widely used nonlinear, non-parametric model for condition monitoring. Traditional nonlinear models assume that training and test data come from the same distribution. This assumption can lead to significant errors when the model encounters anomalies, making it challenging to detect and reconstruct sensor states. To address these challenges, this paper introduces a self-correcting anomaly diagnosis model. Unlike traditional methods, this model establishes a dedicated data structure to store normal sensor patterns and generates a dynamic memory matrix that adapts to changes in industrial processes; The proposed method combines penalized offset projection with multi-scale estimation to mitigate the impact of anomalies on estimation results. Additionally, a variable correlation analysis method is developed to optimize input feature selection for the model. The new approach self-corrects anomalous data in a transformed signal space, achieving accurate reconstruction of sensor states. The model's performance is validated using real sensor data from a nuclear power plant system. Results demonstrate that the proposed model significantly enhances signal reconstruction and anomaly detection capabilities, even under more severe simulated conditions. Compared to traditional nonlinear models, the new method improves the metric for reducing anomaly interference by an order of magnitude. However, we did not change the calculation method of the higher-order kernel in the original method, which still faces the problem of matrix inversion.

Different slope geohazards have different causal mechanisms. This study aims to propose a method to investigate the decision-making mechanisms for the susceptibility of different slope geohazards. The study includes a geospatial dataset consisting of 1203 historical slope geohazard units, including slope creeps, shallow slides, rockfalls and debris flows, and 584 non-geohazard units, and 22 initial condition factors. Following a 7:3 ratio, the data were randomly divided into a test set and a training set, and an ensemble SMOTE-RF-SHAP model was constructed. The performance and generalization ability of the model were evaluated by confusion matrix and the receiver operating characteristic (ROC) for the four types of geohazards. The decision-making mechanism of different geohazards was then identified and investigated using the SHAP explanatory model. The results show that the hybrid optimization improves the overall accuracy of the model from 0.486 to 0.831, with significant enhancements in the prediction accuracy for all four types of slope geohazards, as well as reductions in misclassification and omission rates. Furthermore, this study reveals that the main influencing factors and spatiotemporal distribution of different slope geohazards exhibit high similarity, while the impacts of individual factors and different factor values on different slope geohazards demonstrate significant differences. For example, long and persistent rainfall can erode rock masses and lead to slope creep, increased rainfall may trigger shallow mountain landslides, and sudden surface runoff can even cause debris flows. These findings have important practical implications for slope disaster risk management.

Turbulent coolant flow in a nuclear reactor's primary loop leads to temperature non-uniformity in the heat pipe section. Utilizing sound wave temperature measurement techniques mitigates calculation errors caused by temperature variations. However, complex temperature-velocity coupling induces significant bending of sound wave paths, affecting measurement accuracy. This study proposes a triangular ray tracing method to accurately track sound wave paths considering coolant temperature gradients and flow dynamics. Experimental validation using finite element simulation data from the Hua-long Pressurized Reactor primary loop demonstrates improved reconstruction accuracy by up to 0.59% compared to straight lines and by up to 0.04% compared to other ray tracing methods.

Accurate streamflow forecasting is crucial for water resources management and flood mitigation, yet it remains challenging due to the complex dynamics of hydrological systems. Conventional data-driven approaches often struggle to effectively capture spatio-temporal evolution characteristics, particularly the dynamic interdependencies among streamflow gauges. This study proposes a novel deep learning architecture, termed DynaSTG-Former. It employs a multi-channel dynamic graph constructor to adaptively integrate three spatial dependency patterns: physical topology, statistical correlation, and trend similarity. A dual-stream temporal predictor is designed to collaboratively model long-range dependencies and local transient features. In an empirical study within the Delaware River Basin, the model demonstrated exceptional performance in multi-step-ahead forecasting (12-, 36-, and 72 h). It achieved basin-scale Kling–Gupta Efficiency (KGE) values of 0.961, 0.956, and 0.855, significantly outperforming baseline models such as LSTM, GRU, and Transformer. Ablation studies confirmed the core contribution of the dynamic graph module, with the Pearson correlation graph playing a dominant role in error reduction. The results indicate that DynaSTG-Former effectively enhances the accuracy and stability of streamflow forecasts and demonstrates its strong robustness at the basin scale. It thus provides a reliable tool for precision water management.

Geophysical flows, governed by particle composition and channel slope, exhibit distinct kinematic properties and seismic responses under different flow regimes. This study examines the impact of particle composition on granular flow dynamics and seismic signal generation under various flow regimes using flume experiments under dam break conditions. By varying particle composition and flume inclination angles, we investigate the kinematic properties, seismic responses, and the relationship between flow regimes and seismic signal characteristics. The results reveal that particle composition significantly affects flow dynamics, with peak velocity exhibiting a non-monotonic dependence on particle size, and an optimal proportion of large particles maximizing mobility. Seismic signals, including peak amplitude and power spectral density, increase with larger particle sizes and steeper inclination angles, indicating a strong coupling between flow dynamics and seismic responses. A two-segment positive correlation between seismic signals and collisional stress highlights the role of flow regimes, with particle-ground impacts during intense collisional interactions dominating seismic signal generation, we then introduce a dimensionless amplitude parameter and establish a unified correlation with the Savage number across flow regimes. This study advances the understanding of granular flow dynamics and seismic signatures, providing a framework for interpreting seismic data in debris flow monitoring and hazard assessment. Future work to explore the interplay of frictional and collisional mechanisms to refine models of granular flow behavior and physical interpretation of seismic data is warranted.

In real-world scenarios, the rotational speed of bearings is variable. Due to changes in operating conditions, the feature distribution of bearing vibration data becomes inconsistent, which leads to the inability to directly apply the training model built under one operating condition (source domain) to another condition (target domain). Furthermore, the lack of sufficient labeled data in the target domain further complicates fault diagnosis under varying operating conditions. To address this issue, this paper proposes a spatiotemporal feature fusion domain-adaptive network (STFDAN) framework for bearing fault diagnosis under varying operating conditions. The framework constructs a feature extraction and domain adaptation network based on a parallel architecture, designed to capture the complex dynamic characteristics of vibration signals. First, the Fast Fourier Transform (FFT) and Variational Mode Decomposition (VMD) are used to extract the spectral and modal features of the signals, generating a joint representation with multi-level information. Then, a parallel processing mechanism of the Convolutional Neural Network (SECNN) based on the Squeeze-and-Excitation module and the Bidirectional Long Short-Term Memory network (BiLSTM) is employed to dynamically adjust weights, capturing high-dimensional spatiotemporal features. The cross-attention mechanism enables the interaction and fusion of spatial and temporal features, significantly enhancing the complementarity and coupling of the feature representations. Finally, a Multi-Kernel Maximum Mean Discrepancy (MKMMD) is introduced to align the feature distributions between the source and target domains, enabling efficient fault diagnosis under varying bearing conditions. The proposed STFDAN framework is evaluated using bearing datasets from Case Western Reserve University (CWRU), Jiangnan University (JNU), and Southeast University (SEU). Experimental results demonstrate that STFDAN achieves high diagnostic accuracy across different load conditions and effectively solves the bearing fault diagnosis problem under varying operating conditions.

In rolling bearing fault diagnosis, when an unknown fault is present, the Closed-Set Recognition (CSR) method tends to misclassify it as a known fault. To address this issue, an Open-Set Recognition (OSR) framework is proposed for rolling bearing fault diagnosis in this study. The framework is built upon a serial multi-scale convolutional prototype learning (SMCPL) network, enhanced with an efficient channel attention (ECA) mechanism to extract the most critical fault features. The extracted features are fed into the Density Peak Clustering (DPC) module, which identifies known and unknown classes based on the computed local densities and relative distances. Finally, validation is performed on the Case Western Reserve University (CWRU) dataset, the Xi’an Jiaotong University rolling bearing accelerated life test dataset (XJTU-SY), and the Paderborn University bearing dataset (PU), Germany, and the framework is comprehensively evaluated in terms of several evaluation metrics, such as normalization accuracy and feature visualization. The experimental results show that SMCPL-ECA-DPC outperforms the comparative methods of SMCPL, CPL, ANEDL, CNN, and OpenMax and has high diagnostic performance in the identification of unknown faults.

Accurate prediction of steam generator (SG) water levels under various operating conditions enhances the safety and economic efficiency of nuclear power plants. Based on existing machine learning and intelligent algorithm theories, this paper proposes an ECRBM-GRU-SSA ensemble model for SG water level prediction. The model employs an enhanced continuous restricted Boltzmann machine (ECRBM) for data augmentation and combines the prediction results of multiple gated recurrent unit (GRU) sub-models, optimized by the sparrow search algorithm (SSA), to generate the final output. To validate the model’s predictive performance, this study uses simulated SG water level data. Experimental results demonstrate that the proposed model exhibits a rational and effective structure, with predicted water levels closely aligning with actual variations. Furthermore, its evaluation metrics outperform those of other mainstream models. This model provides an effective tool for fault diagnosis and fault-tolerant control in SG-related systems, offering significant engineering value.

Flood and debris flow frequently transpire in mountainous regions, leading to ground vibrations that can be detected by geophones. The quantitative analysis of these ground motion signals provides valuable seismology-based insights. This study centers on a previously unmonitored debris flow gully that incorporates an artificial step-pool system designed for debris flow mitigation. An integrated monitoring system was implemented, consisting of meteorological, hydrological and seismic sensors, to monitor flood and debris flow events. The recorded seismic signal was analyzed by means of short-time Fourier transform, and a bedload-induced model was utilized to compute the sediment flux. The results indicated that flood and debris flow can be distinguished by the short-term average/long-term average (STA/LTA) method, with thresholds of 30 and 20 for upstream and downstream stations, respectively. The sediment fluxes and transport rates passing through the monitoring stations were evaluated. Considering the topographic variation, the sediment source was identified. And the presence of step-pools mitigated the magnitudes of debris flow from upstream to downstream stations, leading to localized erosion within the downstream reach. The results offer valuable insights for monitoring debris flow and analyzing sediment movement influenced by bed structures such as step-pools.

The accurate and rapid measurement of boiler furnace temperature distribution is vital in the energy industry for optimizing efficiency and ensuring safety. Traditional temperature field reconstruction algorithms, while effective, suffer from high computational complexity, continuous error accumulation, and parameter sensitivity. To overcome these limitations, the fast response temperature field reconstruction network (FTRN) is proposed, which directly utilizes acoustic measurement data and integrates a feature extraction encoder (FEE) with a multiscale attention-enhanced reconstruction (MAR) module. A method for generating a path-time input matrix is developed, transforming measurement data into a 3-D format that enhances the processing efficiency of FTRN. The proposed method avoids complex traditional mathematical processing and experiential parameter selection, enhancing both accuracy and speed. Simulation studies and engineering validations demonstrate the performance and industrial applicability of the proposed approach. The code is publicly available at https://github.com/Childweii/FTRN.git.