Sheng Wu

Accurate interpretation of wind turbine operational measurements is essential for reliable wind power forecasting and efficient wind farm operation. Modern wind turbines are equipped with supervisory control and data acquisition (SCADA) systems that continuously record key operational parameters, providing rich measurement data for data-driven modeling. However, although deep learning methods have demonstrated strong capability in capturing nonlinear relationships, most existing approaches lack physical interpretability, show limited robustness under extreme conditions, and often neglect fundamental aerodynamic constraints. To address these limitations, this paper proposes a physics-constrained graph attention network (PCGAT) for multi-horizon mid-to-long term (from 6 h to 3 d) wind turbine power forecasting using SCADA measurement data. The proposed model constructs a temporal graph in which each node represents a measurement time step and connects to its preceding neighbors to capture dynamic dependencies, while a multi-head graph attention mechanism extracts informative representations from the graph-structured time series. A hybrid loss function incorporating aerodynamic constraints is introduced to ensure physical consistency during model training. Experiments on real-world SCADA datasets demonstrate that the proposed approach achieves higher forecasting accuracy and stronger robustness than several state-of-the-art models across multiple prediction horizons.

This study proposes an improved lightweight YOLOv8n algorithm for stain detection on offshore photovoltaic panels to address limited computing capacity and low real-time accuracy in cleaning robots. The original C2f module is replaced with a C2f_GhostConv structure, significantly reducing parameters and computation while preserving key features. Additionally, an Efficient Channel Attention (ECA) mechanism is integrated to enhance critical feature channels and suppress background noise under complex lighting. A custom dataset of 690 images containing bird droppings, seaweed, and salt crystallization was used for validation. Experimental results show that the proposed YOLOv8n-Ghost-ECA model achieves a 19.1% reduction in parameters and improves mAP50 to 0.913, outperforming the original YOLOv8n in both accuracy and real-time performance. The method demonstrates strong potential for practical deployment in offshore photovoltaic cleaning systems.

As a typical electronic switching device, IGBT is widely used in various fields. Reducing the parasitic inductance parameters of IGBT is of great significance for improving the performance of power devices, enhancing system stability and reliability. To study the parasitic inductance of the internal bonding wire connection structure during the operation of IGBT modules, this paper considers the morphological modeling of bonding wires, bonding parameters, and the layout of bonding wire arrays, and proposes a new analysis model. Through mathematical calculation of the analysis model and Ansys Q3D simulation, the bonding wires with different geometric dimensions and layouts were studied and good correlation were obtained. This will reduce the impact of bonding parameters on the total parasitic inductance of the bonding wire during the pre-layout stage.

Hollow mesoporous Fe3O4 microspheres are attractive as corrosion inhibitor carriers for their structural stability, and strong superparamagnetic properties. Here, the morphology-controlled magnetic hollow mesoporous Fe3O4 nanocarriers (MHMNs) with saturation magnetization of ca. 77 emu/g were fabricated by self-template etching route. Serial characterization results showed that the sample (MHMNs-7.0), obtained by etching MHMNs with 7.0 mg/mL of the oxalic acid solution, presented the largest cavity volume and a rich pore structure, which promotes its high loading capacity of 77.6% for 2-mercaptobenzothiazole (MBT) inhibitor. The resulting material could be used as the carrier for loading various types of inhibitors, e.g., MBT, Ce3+, benzotriazole (BTA), and under negative vacuum treatment. The release behavior of the obtained composite inhibitors was detected by UV-vis spectroscopy, and the inhibition behavior of the composite inhibitors for mild steel in 3.5% NaCl solution was evaluated by electrochemical impedance spectroscopy and immersion tests. The results showed that all of the as-obtained composites presented magnetic-controlled release properties and good inhibition performance. Without the external magnetic field (no MF mode), the release of MBT from MHMNs-7.0/MBT mainly involved Fickian diffusion. Under the application of a magnetic field (MF ON mode), the release behavior was dominated by a swelling mechanism, Fickian diffusion, and diffusion between magnetic particles. The nanocomposite inhibitor had good inhibition efficiency for carbon steel in 3.5% NaCl solution with a maximum of 93.66%, and its continuous release properties enabled it to exhibit long-term corrosion inhibition performance.

As a crucial component in marine monitoring, meteorological observation, and navigation systems, studying the motion characteristics and impact loads of buoy water entry is vital for their long-term stability and reliability. When deployed, buoys undergo a complex motion process, including the impact of entering the water and a stable floating stage. During the water entry impact phase, the motion characteristics and impact loads involve interactions between the buoy and the water, the trajectory of motion, and dynamic water pressure, among other factors. In this paper, the VOF model is used to calculate the buoy’s water entry motion characteristics, and then the STAR-CCM+&ABAQUS bidirectional fluid–structure interaction (FSI) method is used to calculate the water entry impact load of the buoy under different water surface conditions and different initial throwing conditions, considering the influence of the flow field on the structure and the influence of the structure deformation on the flow field. The study finds that under the influence of wave and current impacts, changes in wave height significantly affect the buoy’s heave motions. Under different parametric conditions, due to the specific direction of wave and current impacts, the buoy’s pitch amplitude is relatively more intense compared to its roll amplitude, yet both pitch and roll motions exhibit periodic patterns. The buoy’s pitch motion is sensitive to changes in the entry angle; even small changes in this angle result in significant differences in pitch motion. Additionally, the entry angle significantly impacts the peak vertical overload on the buoy. Instantaneous stress increases sharply at the moment of water entry, particularly at the joints between the crossplate and the upper and lower panels, and where the mast connects to the upper panel, creating peak stress concentrations. In these concentrated stress areas, as the entry speed and angle increase, the maximum equivalent stress peak at the monitoring points rises significantly.

The resonant magnetoelectric (ME) effect of Fe₇₈Si₉B₁₃/Pb(Zr,Ti)O₃ (FeSiB/PZT) composites with a surface-modified Fe₇₈Si₉B₁₃ amorphous alloy has been studied. The surface-modified FeSiB can improve the ME coefficient at the resonant frequency by optimizing the magnetomechancial power conversion efficiency. The maximum ME coefficient of the surface-modified ribbons combined with soft PZT (PZT5) is two-thirds larger than that of the composites with fully amorphous ribbons. Meanwhile, the maximum value of the ME coefficient with surface-modified FeSiB ribbons and hard PZT (PZT8) is one-third higher compared with the fully amorphous composites. In addition, experimental results of magnetomechanical coupling properties of FeSiB/PZT composites with or without piezoelectric layers indicate that the power efficiency of the composites first decreases and then increases with the increase in the number of FeSiB layers. When the surface crystalline FeSiB ribbons are combined with a commercially available hard piezoelectric ceramic plate, the maximum magnetoelectric coupling coefficient of the ME composite reaches 5522 V/(Oe*cm), of which the electromechanical resonant frequency is 23.89 kHz.