Wei Song

The ability to represent fluid dynamics across regime transitions remains a fundamental challenge in physics-informed machine learning. In cylinder wake flows, the Hopf bifurcation at the critical Reynolds number Rec≈52 induces a transition from a steady, nearly symmetric base state to periodic von Kármán vortex shedding, accompanied by a qualitative reorganization of the underlying solution manifold. This raises a key question of whether nonlinear fluid systems admit a unified neural representation across bifurcation-induced reorganization of the solution manifold. Here, we propose a regime-adaptive dual-branch physics-informed neural network to investigate this question. The model explicitly decomposes the solution space into steady and unsteady dynamical components, which are coupled through a continuous Re-dependent gating mechanism centered at Rec≈52, thereby reducing cross-regime interference during optimization. To ensure stable learning under continuous parameterization, we combine continuous Reynolds-number feature embeddings with a multi-stage curriculum learning protocol that progressively restores the contributions of time-resolved observations and transient physical constraints while retaining the complete Reynolds-number range throughout training. Using supervised computational fluid dynamics data at 11 Reynolds numbers over 5≤Re≤120, the framework is evaluated on six held-out in-range Reynolds numbers and two extrapolation cases, Re=2 and 126. The model accurately reconstructs the principal wake structures and key flow statistics at unseen Reynolds numbers, with particularly strong performance across the held-out in-range cases. These results show that explicitly encoding regime separation in the neural architecture improves the representation of bifurcation-induced solution-manifold reorganization and enables continuous parametric surrogate modeling across steady and periodically shedding wake regimes.

The simulation of streamflow is a complex task due to its intricate formation process. Existing single models struggle to accurately capture the stochastic, non-stationary, and nonlinear dynamics of basin streamflow in changing environments. This study combined process-driven hydrological mechanism models with data-driven deep learning models, considering various factors like hydrology, meteorology, environment, and the interconnected effects of upstream and downstream rivers, to develop an interpretable hybrid streamflow simulation model. The study collected multiple external variables to better understand the hydrologic system complexity and used the Maximum Information Coefficient (MIC) to analyze their relationship with streamflow. Subsequently, the Xinanjiang (XAJ) model with physical mechanisms was employed, alongside the TCN-GRU model integrating Temporal Convolutional Network (TCN) and Gated Recurrent Unit (GRU), for separate streamflow simulations. Furthermore, a robust integration method was adopted, realizing nonlinear ensemble through Random Forest (RF), thus establishing the hybrid XAJ-TCN-GRU model. This model exhibits promising results in simulating streamflow in four different basins in China, achieving high Nash-Sutcliffe Efficiency (NSE) values with 0.991, 0.971, 0.984, and 0.986 for the Wuding River, Chu River, Jianxi River, and Qingyi River respectively. In terms of streamflow simulation, flood simulating, and interval simulation, this model outperforms other benchmark models. Additionally, the study quantified the contributions of each hydro-meteorological variable to the long-term streamflow trend using mean absolute SHAP values (SHAPABS), Feature Importance (FI), and Permutation Feature Importance (PFI), thereby enhancing the model's external interpretability. The results of this study are of significant importance for optimizing water resource management and mitigating flood disasters.

An off-grid deep-offshore electricity-hydrogen integrated energy system (DOEH-IES) provides a practical carrier for onsite offshore wind utilization and multi-energy coordination. However, under typhoon conditions, coupled wind-wave effects and platform motion reduce the available power of floating offshore wind turbines (FOWTs), introducing a time-varying downside risk that conventional scheduling methods often fail to mitigate proactively. To address this issue, a downside-risk pre-positioning framework is developed for the DOEH-IES that converts physical FOWT power reductions into actionable scheduling decisions. First, a wind power correction model for FOWTs is established by considering wave, floating-platform motion, and air density variation. A dynamically asymmetric envelope is then constructed to extract the time-varying downside exposure during a typhoon and further develop a downside risk indicator. Subsequently, driven by the risk indicator, BESS capacity is reserved in advance of the onset of high-risk periods. Case studies based on typhoon reanalysis data from the South China Sea show that the proposed dynamic asymmetric envelope achieves risk-oriented boundary characterization without increasing weekly average envelope conservatism, with a pre-typhoon average envelope width of 2.69 MW and a maximum downside deviation of 150 MW during peak disturbance. Relative to the comparison case without risk-informed pre-positioning, the proposed framework eliminates electrical load shedding and reserve insufficiency, increases pre-disturbance available stored energy by 72.5%, and reduces natural gas consumption by 15.0% with only a 2.09% cost increase. These results demonstrate that the proposed framework can effectively translate the risk into pre-positioning actions, safeguarding supply and equipment continuity.

Sea ice segmentation based on Synthetic Aperture Radar (SAR) images has become an important technical means for polar climate change monitoring and navigation safety guarantee. However, the existing methods have limitations in the utilization of SAR polarization information and the modeling of local diversity details of sea ice, which leads to insufficient segmentation, especially in complex ice-water boundary regions. To address these issues, this paper proposes a novel Polarization-Fused Edge-Enhanced UNet (PFEE-UNet) designed specifically for sea ice segmentation from high-resolution SAR images. Specifically, we design the Cross-Polarization Channel Interaction (CPCI) module, which employs a dual interaction strategy of hierarchical inter-group cascading and symmetric cross-fusion. This approach effectively leverages the complementary features of the HH and HV polarization channels, significantly enhancing the distinction between sea ice and open water. Additionally, we present the Dense–Sparse Diversity Enhancement (DSDE) module, which combines a spatial-channel joint attention mechanism to strengthen the model’s ability to capture spatial relationships within complex ice–water structures, effectively alleviating misclassifications caused by abrupt local texture changes. Finally, we design the Selective Edge Fusion (SEF) module, which dynamically selects and integrates multi-level edge features, improving the continuity of sea ice boundaries and preserving its morphological integrity. The experimental results show that the proposed PFEE-UNet model outperforms mainstream segmentation methods on the AI4Arctic/ASIP sea ice dataset, achieving an average Intersection over Union (IoU) of 84.48%, which surpasses existing methods such as HRNet (82.52%) and DeepLabv3+ (82.40%). Additionally, PFEE-UNet was applied for end-to-end ice–water segmentation on real-world Sentinel-1 SAR scenes, demonstrating its effectiveness and robustness for practical sea ice monitoring.

Maritime target detection under complex adverse weather conditions (e.g., fog, rain, and low light) is crucial for Unmanned Surface Vehicle (USV) navigation. However, achieving high detection accuracy and efficiency remains challenging due to coupled environmental interference and limited computing resources. In this paper, we propose W-MTD, a task-specific distillation framework designed for weather-robust and lightweight maritime target detection based on knowledge distillation. Building upon the Fine-grained Distribution Refinement (D-FINE) detection model, this method constructs a dual-path knowledge distillation framework tailored for maritime scenes. Through the synergistic optimization of feature similarity constraints and decoupled distillation, it facilitates multi-level knowledge transfer from a teacher model to a lightweight student model, mitigating feature degradation caused by model compression. A multi-scenario augmentation strategy is designed to balance convergence across different weather conditions. Experiments show that W-MTD’s student model improves detection accuracy by 7.0–13.9% under three adverse weather conditionscompared to the baseline teacher model trained solely on clear weather data while maintaining comparable clear-weather performance. With only 4 M parameters and 7 GFLOPs, the student model demonstrates favorable performance and efficiency compared to other real-time detectors, indicating its potential suitability for USV deployment.