To enhance the reliability and accuracy of predictions, a lithium battery RUL prediction model based on Bidirectional Long Short-Term Memory Network(BiLSTM) and Attention mechanism is constructed. Firstly, the model takes the key performance parameters during the charge-discharge process of lithium batteries as the inputs, and utilizes the bidirectional memory characteristics of the BiLSTM network to fully capture the long-term and short-term temporal dependencies in the battery performance degradation process and effectively excavates the degradation trend features hidden in the data. Secondly, the Attention mechanism is introduced to assign higher weights to the key time-step features that affect RUL prediction, strengthening the model's ability to focus on important information and thereby outputting accurate RUL prediction values. Validation and analysis on four sets of public CALCE lithium battery datasets show that the R² values of the BiLSTM-Attention model reach 98.77%, 99.64%, 99.50%, and 98.47% respectively. Compared with other prediction methods, this method exhibits superior prediction performance.

Aiming at the problems commonly existing in distribution networks with high-proportion new energy integration, such as low voltage, over-voltage, unbalanced voltage, line overload and excessive line loss, a flexible energy storage optimized control method for active distribution networks oriented to edge-terminal coordination is proposed. Firstly, based on the terminal-edge-cloud concept technology of the distribution internet of things, a coordinated operation architecture of the terminal equipment layer and the edge computing service layer relying on communication services is constructed, and the scheduling information interaction mechanism between the multi-source devices at the distribution network terminal and the intelligent integrated terminal platform is clarified. Secondly, considering that battery energy storage can adaptively adjust the battery grid-connected capacity according to the grid demand, a mathematical model of modular series battery energy storage is established. On this basis, taking into account the charging and discharging costs, operation and maintenance costs and voltage regulation costs of battery energy storage, a coordinated optimized control method for active and reactive power of flexible energy storage in active distribution networks is proposed. Finally, case studies on the IEEE 33-node system verify the effectiveness of the proposed method in enhancing the operational economy and voltage quality of distribution networks.

Distributed energy storage systems have key advantages such as modular design flexibility, bidirectional power regulation, ease of installation, and flexible revenue models, which provides effective solutions for power flow overrun, reverse power flow, and three-phase voltage imbalance in distribution grids. This paper analyzes the application scenarios of distributed energy storage and summarizes the mainstream system architectures of distributed energy storage systems. For the power conversion units within energy storage systems, the current status and advantages/disadvantages of DC/DC and DC/AC converters are discussed. On this basis, the control strategies for DC/DC and DC/AC converters in energy storage systems are analyzed. Finally, it provides an outlook on the future trends of distributed energy storage.

Based on the importance of state of charge(SOC) to the battery management systems(BMS), a Grey Wolf Optimizer enhanced SVD-based Multi-Innovation Unscented Kalman Filter(GWO-SVD-MIUKF) for SOC estimation is proposed. The method combines Grey Wolf Optimizer(GWO) with a singular value decomposition-based MIUKF(SVD-MIUKF). Both the parameter identification and the filter structure are optimized. SVD is used to reconstruct the covariance matrix in MIUKF to improve nu?merical stability, while GWO is used to identify model parameters and dynamically adjust the es?timation window, enhancing adaptability and convergence. Experiments are conducted on the public INR18650-20R dataset from the University of Maryland under various typical conditions. Results show that the proposed method achieves high estimation accuracy, with SOC error controlled within ap?proximately 0.20%, and demonstrates good convergence performance.

Aiming at the problems of increased volatility and reduced accuracy of calculation results under wide operating conditions in both existing time and frequency domain identification methods, a quadratic iterative parameter identification method based on the construction of increasing rank condition is proposed to improve the parameter identification ability under wide operating conditions and variable power grid conditions. Firstly, in view of the problem of insufficient adaptability of the existing time-domain methods to the change of power grid operation conditions, the interaction characteristics of the grid-forming energy storage converter and the power grid are mathematically abstracted, and a closed-loop identification model is constructed for single-machine and two-machine/multiple-machine forms. Secondly, to address the problem that the existing frequency domain methods tend to fall into local optimization during iterative computation, a quadratic iterative solving method based on increasing the rank of state equations is proposed, which improves the global optimization performance by applying the increasing-rank condition, and improves the consistency of the identification results under a wide range of operating conditions. Finally, the effectiveness of the closed-loop identification model is verified by simulation. Resultsshow that the proposed method can obtain more accurate identification results under variable operating conditions and power grid scenarios.

To address the safety issues of 280 A·h lithium iron phosphate batteries for energy storage, a comparative analysis of the effects of polyvinylidene fluoride (PVDF) and polymethyl methacrylate (PMMA) as separator adhesive layer materials on the safety performance of batteries under overcharging, external heating, and needle-puncture abuse conditions is performed. The results show that batteries with PVDF adhesive layers exhibit superior safety characteristics under all three abuse conditions: the activation time of the safety valve is delayed by 0.1~7.5 minutes, and the voltage drop time is postponed by 0.1-10.5 minutes. Following thermal runaway, the maximum temperature is decreased by 11℃, 1.5℃, and 16.7℃, respectively, while the maximum temperature rise rate is reduced by 0.11~3.18℃/s. This study provides critical experimental data for separator material selection and safety optimization in energy storage batteries.

To enhance the economy of offshore wind power and make fuller use of offshore wind power resources, high-voltage direct current transmission and DC aggregation technologies related to offshore wind power DC transmission is proposed. Combining different connection methods and different system boosting times, the topology of offshore wind power DC aggregation and boosting is selected to meet the current development trend of offshore wind power, and the characteristics of each topology in the DC aggregation system under typical offshore wind power scenarios are analyzed. The construction cost, energy loss and reliability of each topology are studied in conjunction with typical case studies. Finally, the most suitable DC aggregation scheme configuration for large-scale offshore wind farms is then derived to improve the comprehensive benefits of the offshore wind power DC aggregation system.

Aiming at the problems of slow convergence of speed error and serious vibration in the traditional sliding mode control of interior permanent magnet synchronous motor (IPMSM), a IPMSM control method combining a novel second-order super-twisting observer（NSTO） with a model-free non-singular terminal sliding mode control (MFNTSMC) algorithm is proposed. First, a mathematical model of IPMSM containing unknown parameters and load disturbances is established, and an MFNTSMC algorithm is designed. Then, based on the hyper-local model, a dual closed-loop MFNTSMC structure is developed to achieve global fast convergence of the system state errors. To further improve speed tracking performance, a novel second-order super-twisting sliding mode disturbance observer is designed to observe system disturbances in real time for disturbance compensation. Finally, simulation and experimental results demonstrate that compared with traditional methods, the proposed approach offers significant advantages in suppressing vibration, improving convergence speed and robustness.

The total electric field in HVDC wall bushing undergoes significant changes under corona discharge conditions. Therefore, an engineering-applied ±100 kV SF₆ gas-insulated DC wall bushing is selected as the study object. Using a field mill-type electrostatic field meter, the total electric field during both polarization under +90 kV DC voltage and subsequent depolarization after voltage removal are measured. The grounding tips are installed near the HVDC wall bushing flange to induce corona discharge, and the influence of tip lengths on the total electric field is analyzed. The results show that under normal conditions, the total electric field decreases rapidly initially and then stabilizes after 7200s with increasing polarization time, due to the accumulation of space charges on the surface. During depolarization, the total electric field decays approximately exponentially with a time constant of about 3×10⁴ s. As the tip length increases, the corona discharge transitions from smaller discharge magnitudes with higher frequencies to larger magnitudes with lower frequencies, resulting in a faster decrease rate and a lower stabilized value of the total electric field in the HVDC wall bushing. This study provides data support for elucidating the space charge distribution pattern on surface of HVDC wall bushings.

Post-disaster power system restoration is challenged by the dual uncertainties of unit start-up times and load recovery amounts, which poses substantial threats to restoration security. Therefore, a source-load coordinated restoration strategy based on Information Gap Decision Theory(IGDT) robust optimization is proposed. Firstly, the possible fluctuation ranges of uncertain parameters are mathematically characterized. Subsequently, a multi-objective uncertain optimization model is established with the goals of maximizing both units restoration output and load-weighted restoration amount. Relying on the IGDT theory, this uncertainty model is transformed into a deterministic optimization problem that meets minimum restoration performance requirements, which is then efficiently solved using the Non-dominated Sorting Genetic Algorithm Ⅱ (NSGA-Ⅱ). The proposed approach achieves coordinated cross-time-step restoration of sources and loads, providing effective support for operators to develop restoration plans that balance security and efficiency. Case studies on the New England 10-machine 39-bus system validate the robustness and effectiveness of the strategy.

During winter cold wave events, wind turbines in alpine mountainous wind farms in a certain province are prone to capacity loss caused by icing, posing significant challenges to grid dispatching and power balance. To address this issue, a regional prediction method for wind turbine capacity loss caused by icing is proposed, which takes into account both meteorological conditions of glaze ice and freezing fog, and introduces a Freezing Meteorological Index(I_FM) to quantitatively describe the external meteorological conditions of wind power ice-covered in the process of winter cold waves. A predictive model is constructed to map I_FM to wind power capacity loss due to icing. Empirical validation is conducted using multiple typical cold wave events. Results show that the proposed method demonstrates high prediction accuracy across several cold wave processes, and the average accuracy reaches 85.44% in the prediction of capacity loss due to icing in the winter of 2024-2025, indicating strong practical value and regional applicability.

Aiming at the shortcomings of traditional methods for predicting line loss in substation areas, such as inability to adjust flexibly, lack of real-time performance, and accuracy affected by meteorological conditions, a new method for predicting line loss rate in substation areas based on maximum information coefficient and BO-LSTNet is proposed. The maximum information coefficient method is used to screen meteorological information in the substation area, and input variables are filtered and cleaned, and finally the line loss data is put into the LSTNet model after Bayesian optimization. The reliability is verified through simulation experiments, and the results show that compared with previous prediction methods, this model has higher adaptability to the new type of substation line loss rate prediction, and solves the problem of low accuracy in line loss prediction in practical engineering.

The existing supply chain management of power materials involves multiple business systems. The collaboration of electronic data across business systems faces challenges such as data heterogeneity, reliance on trusted third-party platforms, and high risks of data circulation leakage across systems. In order to solve these problems, a cross-system data security aggregation scheme for power material supply chain based on additive secret sharing is proposed. The scheme relies on additive secret sharing to build linear and nonlinear secure computation frameworks, enabling secure data aggregation without trusted third parties. Moreover, a series of secure aggregation protocols are designed, including secure sum aggregation, secure mean aggregation, secure variance aggregation, four-party secure comparison and secure extremum aggregation, to solve the problem of data interoperability among cross-business systems. Rigorous security analysis and experiments show that this scheme can not only ensure the accurate aggregation of data while protecting the privacy of each business system, but also does not rely on any external trusted third party at all. It has superior performance in terms of aggregation communication overhead and efficiency, providing effective technical support for the digital management of the material supply chain.

High-quality development planning for power grid enterprises is not only a key path to realize high-quality development of the energy industry, but also an important support to promote the realization of the ”double carbon” goal and the construction of a new energy system, through the latent dirichlet allocation(LDA) topic identification model, indicators are selected from six aspects, such as safety and reliability, scientific and technological innovation, service capability, production efficiency, business performance, green development, etc., and the evaluation system of high-quality development planning for power grid enterprises is established by combining the hierarchical analysis method and BP neural network, and empirical analysis is carried out. It is found that the effectiveness of high-quality development planning of power grid enterprises is on the rise, but the annual growth rate fluctuates greatly, and compared with safety, reliability and service capacity, the degree of importance attached to scientific and technological innovation and green development is relatively insufficient. The key factors influencing the high-quality development planning of power grid enterprises should be systematically optimized, and the planning and construction of scientific and technological innovation and green development should be strengthened.

In order to achieve effective online monitoring of the state of metering voltage transformers, the characteristics of group transformer voltage ratio difference data are theoretically analyzed. An anomaly detection method for voltage transformers is proposed based on the 3σ criterion, and an online monitoring system for voltage transformers is developed, which can effectively and economically realize the identification of transformer anomalies. The field data verification demonstrated that the sample proportion of ratio difference data within the 3σ interval is 99.52%, confirming the gaussian distribution characteristics. The anomaly probability for out-of-tolerance transformers fluctuates between 40% and 80%, significantly higher than that of normal transformers.

Aiming at the limitations that the traditional substation three-dimensional temperature re?construction technology not only relies on high-precision laser radar point cloud completion and three?-?dimensional scanning technology, but also frequently encounters issues such as insufficient point cloud density and distortion of appearance representation, given the inherent characteristics of infrared images of low textures and high noise, a dual-path three-dimensional modeling system based on infrared images is proposed. This system completes high-density temperature distribution reconstruction through the cross?-domain fusion of thermal structural features and geometric priors, combined with a three?-dimensional Gaussian splatting algorithm, ultimately generating three-dimensional models of substation scenes and equipment that achieve high-density temperature distribution reconstruction while maintaining extremely high morphological fidelity. Compared with existing methods, the model accuracy is improved by about 5%. The method proposed in this paper can achieve an upgrade in three-dimensional temperature reconstruction effects for substations, enhancing the intelligent operational inspection capabilities of substations.

In response to the challenges of high-dimensional nonlinear problems and alternative node strategies that struggle to accurately reflect the actual structural requirements of the distribution network in traditional distributed generation (DG) siting and sizing planning, a DG siting and sizing method based on an improved strategy combined dung beetle optimization algorithm is proposed. Firstly, the weighted active loss objective function incorporating the load sensitivity factor(LSF) of the fusion node is constructed, and the dual-objective optimization model is formed together with the voltage deviation. Secondly, an improved dung beetle optimization algorithm that integrates Piecewise chaotic initialization, random walk perturbation, and cross strategies is designed to enhance global search capabilities and adaptability in complex environments. Finally, through the analysis of the IEEE-33 node distribution system example, a comprehensive evaluation of the optimization scheme is conducted from the perspective of the entire lifecycle, including aspects such as net benefits, investment costs, and carbon reduction benefits. The results indicate that the proposed method not only outperforms various traditional algorithms in reducing network losses and voltage deviations, but also demonstrates significant economic advantages, which validates its comprehensive superiority in both technical performance and economic feasibility.

To achieve efficient bearing fault diagnosis, a novel method based on the IEC module (integrating InceptionV2, Efficient Channel Attention, and Convolutional Block Attention Module) and a Two-stream Convolutional Neural Network(TSCNN) is proposed. First, raw vibration signals are converted into one-dimensional data and two-dimensional time-frequency images using Fast Fourier Transform(FFT) and Continuous Wavelet Transform(CWT). Subsequently, an improved TSCNN fusion model is constructed, and the obtained wavelet time-frequency images and FFT spectra are used as inputs to extract the spatial features of the time-frequency images by using InceptionV2 and ECANet-CBAM improvement module, and the resulting dual-layer feature information is fused into the Softmax layer to accomplish fault classification. Finally, comparative analysis based on a standard rolling bearing fault data set demonstrates that the proposed IEC-TSCNN method achieves superior diagnostic accuracy.

Aiming at the low working efficiency, high safety risk, difficult quality control, poor terrain adaptability and other problems of the existing manual work mode of pole erection, a multifunctional construction vehicle for distribution network pole erection is designed to integrate drilling, lifting, digging, crushing, tamping and other functions into a single vehicle. Furthermore, based on the construction specifications for distribution network pole erection and the structural and functional features of the designed multifunctional vehicle, a standardization study is carried out for distribution network equipment mechanized construction programme. A dedicated construction methodology compatible with the multifunctional vehicle is proposed, promoting the shift toward mechanized and standardized processes in distribution network construction.

Aiming at the problem of high voltage XLPE cable buffer layer ablation faults causing a big impact on the safe operation and maintenance of cables, the buffer layer ablation fault detection technology based on characteristic gases is proposed. By building the XLPE cable buffer layer simulated ablation experimental platform and cable true type experimental platform, the characteristic gases and their ablation characteristics generated by the cable buffer layer ablation faults are studied, focusing on the buffer layer moisture rate, the buffer layer thickness, and the buffer layer applied current RMS value on the buffer layer characteristic gases. The results show that the magnitude of current applied to the buffer layer is the most important factor affecting the concentration of characteristic gases, and the buffer layer is more prone to ablation under moisture conditions,with smaller suffer layer thickness being more prone to ablation. Through the practical application of two cables with buffer layer ablation faults, the results show that the analysis of the buffer layer gas composition can be used to diagnose whether or not buffer layer ablation faults have occurred in the cable segment and the severity of faults, which verifies the feasibility of buffer layer ablation fault diagnostic technology based on characteristic gases.