Unmanned aerial vehicle (UAV) inspection has become the mainstream of transmission line inspection, and the detection of insulator defects is an important part of UAV inspection. On the premise of ensuring high accuracy and detection speed, an improved YOLOv5 model is proposed for defect detection of insulators. The algorithm uses the weights trained on conventional large-scale datasets to improve accuracy through the transfer learning method of feature mapping. The algorithm employs the Focal loss function and proposes a dynamic weight assignment method. Compared with the traditional empirical value method, it is more in line with the distribution law of samples in the data set, improves the accuracy of difficult-to-classify samples, and saves a lot of time. The experimental results show that the average accuracy of the insulator and its defect is 98.3%, 5.7% higher than the original model, while the accuracy and recall rate of insulator defects are improved by 5.7% and 7.9%, respectively. The algorithm improves the accuracy and recall of the model and enables faster detection of insulator defects.

To enhance equalization efficiency and address the issue of traditional equalization methods overlooking temperature factors, this paper proposes a multilayer equalization circuit for both intra-group and inter-group balancing. The traditional Buck-Boost equalization topology between groups is improved by incorporating a two-way interleaved inductor structure, which helps reduce equalization idle time. An adaptive fuzzy control equalization strategy for multiple objectives is applied to the topology. The state of charge (SOC) and temperature of the battery are used as key variables for equalization, with the equalization current dynamically adjusted based on changes in the SOC and temperature. This approach improves the balance between equalization speed and temperature control, reducing equalization time while limiting battery temperature rise. A simulation model is developed using MATLAB/Simulink. The simulation results demonstrate that, compared to the traditional Buck-Boost equalization topology, the proposed topology reduces equalization time by 15.1%. Additionally, under three different operating conditions, the equalization cotnrol strategy designed in this paper improves time efficiency by over 14% compared to traditional methods, while also reducing both the maximum temperature and temperature difference.

In sensor design, electromagnetic field numerical simulation techniques are widely used to investigate the working principles of sensors. These analyses help designers understand how sensors detect and respond to external signals during operation. One popular method for electromagnetic field computation is the meshless radial point interpolation method (RPIM), where the number and distribution of nodes are critical to ensuring both accuracy and efficiency. However, traditional RPIM methods often face challenges in achieving stable and precise results, particularly in complex electromagnetic environments. In order to enhance the stability and accuracy of electromagnetic numerical calculations, a node generation and adaptive refinement algorithm for the meshless RPIM is proposed. The proposed approach includes an initial node-generation method designed to optimize the balance between computational accuracy and efficiency, as well as a dynamic error threshold and hybrid node refinement method to precisely identify and adaptively refine areas requiring additional nodes, ensuring high precision in critical regions. The proposed method was validated through its application to electrostatic fields and multi-media magnetic fields, demonstrating significant improvements in both stability and accuracy compared with conventional RPIM approaches. These findings highlight the potential of the proposed algorithm to enhance the reliability and precision of electromagnetic field simulations in sensor design and related applications.

Short-term wind power forecasting is a key enabling technology for wind farm operation optimization, power grid dispatch, and electricity market decision-making. However, existing studies often lack unified standards in data partitioning, input feature construction, and hyperparameter tuning, making fair and reproducible comparisons across models difficult to achieve. To address this issue, this study focuses on day-ahead wind power forecasting for a single wind farm and establishes a benchmarking framework with strict chronological splitting, a shared feature information set, and a consistent hyperparameter tuning budget. Within this framework, six representative models, namely Ridge, XGBoost, LightGBM, DLinear, Transformer, and PatchTST, are systematically evaluated. A two-level evaluation protocol combining a fixed hold-out split and expanding-window rolling validation is adopted to compare model performance from multiple perspectives, including overall accuracy, sensitivity to hyperparameter tuning, robustness across rolling windows, and performance under typical operating scenarios. The results show that model rankings are not fully consistent between the hold-out evaluation and the rolling-validation setting. Under the fixed hold-out split, LightGBM achieved the lowest NRMSE of 10.2326%, while Transformer obtained the lowest NMAE of 6.9944%. In contrast, under the 8-fold expanding-window rolling validation, Transformer achieved the lowest average NRMSE of 8.1684%, followed by LightGBM with 8.7344%. These results indicate that the best performance on a single test split does not necessarily imply the strongest robustness across multiple time windows. In addition, strong tree-based models remain highly competitive in this single-wind-farm forecasting task, whereas more complex deep temporal models do not always deliver stable advantages. Meanwhile, the performance gains brought by hyperparameter optimization vary substantially across models, suggesting that conclusions drawn from default-parameter comparisons are of limited reliability. These findings demonstrate that systematic benchmarking under strict temporal constraints and fair tuning conditions is essential for clarifying the comparative performance, robustness, and engineering applicability of different model families. The study can therefore provide practical guidance for model selection and deployment in short-term wind power forecasting for single wind farms.

With the rapid expansion of urban metro networks, transformer direct current (DC) bias caused by stray currents from multiple metro lines has become increasingly prominent. To evaluate and mitigate DC bias currents under complex coupling conditions, a coupling model of multiple metro lines and the regional power grid is established based on the superposition principle. Field measurements are further used to validate the model and demonstrate the necessity of considering multi-line coupling. Subsequently, dynamic operating conditions involving multiple trains are simulated to evaluate the magnitude and dynamic distribution characteristics of transformer DC bias currents in the regional power grid. To address the pronounced time-varying behavior of these currents and the limited adaptability of conventional static mitigation strategies, a dynamic iterative mitigation model is proposed for the optimized and incremental allocation of DC bias suppression devices. The results show that, under the simulation period and regional power grid conditions specified in this study, the proposed model can reasonably capture the dynamic characteristics of transformer DC bias currents. Compared with a conventional threshold-triggered strategy, the proposed method reduces the number of required suppression devices by approximately one third.