[Objective] Due to the low fault localization accuracy and efficiency of traditional fault self-healing methods, a fault self-healing method for distribution main stations based on the decision tree and multi-agent system (MAS) was proposed to improve the fault handling capability of distribution systems. [Methods] A hierarchical multi-agent technology was adopted to construct a fault self-healing system for distribution main stations, which included the feeder agent and node area agent. The distribution network data were collected and the gradient boosting decision tree (GBDT) algorithm was employed in the node area agent to complete fault localization, and the fault data were transmitted to the feeder agent. In the feeder agent, data were summarized, the influence of important load recovery sequence, transfer margin, and line loss was comprehensively considered to build a fault self-healing optimization model, and the model was solved via the multi-agent evolutionary algorithm to obtain the optimal fault self-healing recovery scheme for distribution main stations. [Results] Based on the IEEE-29 system, experimental analysis was conducted on the proposed method, and the results show that the accuracy of the GBDT fault localization algorithm is nearly 97％ after 150 iteration. The important load recovery amount, network loss, transfer capacity margin, and fault self-healing time of this method are 100％, 90.58 kW, 11.26 kW, and 2.79 s respectively. The self-healing recovery rate exceeds 91％, and the highest self-healing control operation complexity is no more than 5, all of which are superior to other comparative methods. [Conclusions] The GBDT fault localization algorithm can achieve more ideal accuracy and efficiency, and the proposed method can recover all important loads in the shortest time, ensuring minimal network loss. Additionally, the proposed method has relatively stable self-healing ability, which can better coordinate new energy generation, quickly adapt to the rapid development of new power systems, and achieve high-quality power supply. Aiming at traditional fault self-healing methods suffering from problems such as large workload and poor accuracy caused by centralized processing modes, the proposed method constructed a fault self-healing system for distribution main stations based on MAS, achieving fast and accurate fault detection and recovery via the distributed collaboration of the operating status of each node. Compared to the decision tree algorithm, the GBDT algorithm gradually improves analysis accuracy by fitting the residuals of the previous round in each round of iteration to construct a new learner. It is applicable to fault localization at the level of distribution main stations and provides accurate data support for fault self-healing. Compared with traditional optimization methods, the GBDT algorithm adopts the multi-agent evolutionary algorithm to solve the fault self-healing optimization model. By assigning the target to each agent for execution, the optimization efficiency is improved, and the excellent solutions of all agents are summarized to obtain the final solution, ensuring the global optimal effect.

[Objective] Miniature current transformers are a kind of device adopted for current measurement, and their core components are composed of a primary winding, a secondary winding, and a magnetic circuit system. The primary winding is directly connected in series with the measured current circuit, mainly undertaking the task of inducing the magnetic field of the measured current, while the secondary winding is connected to measuring instruments or protective devices, and is employed to output a signal proportional to the primary current. The magnetic circuit system is composed of high-performance magnetic materials, such as high-permeability ferrites or nanocrystalline alloys, which have excellent magnetic properties and can effectively guide and concentrate magnetic fields, ensuring that the transformer can maintain stable performance in complex electromagnetic environments. However, in practical applications, due to the significant nonlinear characteristics of the excitation winding of the miniature current transformer in the saturation region, significant errors will be generated in exciting voltage calculation in conventional linear modeling methods, which seriously restricts the measurement accuracy and stability of the transformer in high-requirement application scenarios such as smart grids. An intelligent detection method for measuring winding errors of miniature current transformers was proposed to improve the measurement accuracy of miniature current transformers and overcome existing technological bottlenecks. [Methods] In response to the nonlinear saturation characteristics of the excitation winding of miniature current transformers, a segmented linearization modeling method was developed to construct an equivalent circuit of the miniature current transformer for acquiring real-time signals of the transformer under the operation status. On this basis, the problem of insufficient applicability of linear models in the saturation region was solved and more accurate data support was provided for subsequent error analysis. A hybrid filtering algorithm combining the Sine Tapers window function and discrete wavelet transform was designed to conduct filtering processing on the acquired signals. Additionally, the Wiener filter and wavelet threshold denoising technology were combined to improve the signal-to-noise ratio, achieve precise separation of high-frequency noise and effective signals, and enhance signal quality. Meanwhile, correlation analysis was conducted on the filtered data, the principal component subspace was extracted via singular value decomposition, and statistical measures were constructed in the residual subspace. Meanwhile, the principal component analysis method was adopted to decompose the signal into the principal component subspace and residual subspace, with statistical measures and contribution rate calculations performed to achieve quantitative detection and accurate positioning of errors. Additionally, expected value operation was introduced to compensate for temperature drift, with fast transient response achieved by error fluctuation modeling, and real-time monitoring and intelligent detection of measurement winding errors realized via combining statistical changes. [Results] The experimental results show that the error detection method for miniature current transformers based on multi-spectral adaptive wavelet filtering and principal component space decomposition proposed in this paper has significant advantages over the traditional methods. Its signal acquisition results have a higher degree of agreement with the voltage current characteristic curve, and show extremely high accuracy in ratio and angle difference detection. [Conclusions] By deeply integrating multidisciplinary technologies, the key technical difficulties in error detection of miniature current transformers are solved, which can achieve highly accurate detection and fast positioning of measurement winding errors of miniature current transformers and improve the measurement stability and safety of power system operation.

A crucial role is played by insulators in providing insulation and mechanical support during the operation of overhead transmission lines. However, in complex and ever-changing natural environments, the surfaces of insulators are highly susceptible to the adhesion of pollutants such as dust and salt, which can result in uneven contamination, seriously alter the electrical performance of insulators, affect their DC flashover characteristics, and increase the risk of flashover accidents on the line, thus causing serious consequences such as power outages, and posing a huge threat to the safe and stable operation of the power system. Therefore, this paper aims to analyze the direct current (DC) flashover characteristics of unevenly contaminated insulators. [Methods] Both the uneven contamination during actual operation by creating an unevenly contaminated insulator model, and the DC flashover process of unevenly contaminated insulators in overhead transmission lines based on electromagnetic field theory were simulated. The two key parameters of non-soluble deposit density (NSDD) and salt deposit density were selected to comprehensively explore the influence of contamination on the flashover characteristics of insulators. Specifically, the content of insoluble contaminants on the insulator surface was reflected by NSDD, while the content of soluble contaminants was reflected by salt deposit density. The adhesion of contaminants on the insulator surface under different contamination levels was simulated by adjusting the wind speed, as the deposition rate and distribution of contaminants on the insulator surface can be affected by wind speed. Additionally, the DC flashover voltage of insulators with different combinations of NSDD and salt deposit density in different wind speed conditions was recorded. [Results] Uneven contamination can cause significant changes to the electric field distribution on the insulator surface, leading to non-uniformity in the originally even electric field. Meanwhile, the too high local electric field strength can increase the possibility of flashover. NSDD formed by non-soluble contaminants can reduce the risk of flashover to a certain extent. This is because although NSDD changes the roughness of the insulator surface, they themselves are non-conductive, thus to some extent hindering the conduction of currents and increasing the flashover voltage. In contrast, soluble contaminants of salt deposit density are prone to form conductive channels in humid environments, thus significantly reducing the insulation performance of insulators and flashover voltage. Analysis was further conducted on the experimental data to verify the accuracy of the analysis results, and the results show that the fitting coefficient is high, indicating sound agreement between the obtained analysis results and the actual situation. [Conclusions] By simulating the influence of wind speed on the contamination situation of the insulator surface, the actual operating environment gets truly reflected, providing new ideas and methods for studying the DC flashover characteristics of unevenly contaminated insulators. Important theoretical basis and technical support are provided by this study for the selection of insulators for overhead transmission lines, the development of anti-contamination measures, and the safe and stable operation of power systems. The incidence of power outages caused by insulator flashover can be reduced and the power supply stability can be improved.

[Objective] Traditional grey models are widely applied to short-term load prediction due to their sound adaptability to small-sample and information-poor data. However, when handling complex electricity consumption data featuring both exponential growth and linear trends, they suffer from inherent limitations, such as insufficient prediction accuracy, sensitivity to data noise, and weak generalization capability, thus making it difficult to meet the demands of modern refined power management. Given the shortcomings of traditional grey models, a comprehensively improved prediction framework was proposed to significantly enhance the accuracy and practicality of electricity consumption behavior prediction, thereby providing more reliable data support for intelligent management of power systems. [Methods] In the data preprocessing stage, the standard deviation method was adopted to identify and remove outliers, while the linear interpolation method was applied to fill missing values in electricity consumption data with dense collection cycles. During the stage of analyzing user consumption behavior, the K-means clustering algorithm was employed to process load curves, and the elbow method was utilized to determine the optimal number of clusters, identifying user groups with similar consumption patterns. In the stage of prediction model building, an improved grey model was proposed to integrate the traditional grey model with a linear regression model for building a fused grey-linear regression model. In the fused model, sequences were generated via accumulation, and fitting was conducted by employing the combined equation, with the parameters estimated via sequence transformation and the least squares method. Meanwhile, the fused model was utilized to predict the residual sequence, and the Fourier transform was introduced for spectral analysis and noise reduction. A Fourier basis matrix was constructed, and related coefficients were solved by adopting the least squares method to correct the original predicted values. [Results] Validation based on the actual data from 205 users in a specific region demonstrates that the improved model successfully identifies four typical electricity consumption patterns by clustering analysis. The proposed improved grey model was compared with the three baseline models of the traditional grey model, the grey model+linear model, and the grey model+residual correction model. The results show that the improved model exhibits significantly lower mean absolute error (MAE) and mean absolute percentage error (MAPE) than the other three models across all user categories and prediction time points. Its advantage is particularly pronounced during the initial prediction periods, indicating that the model is more suitable for short-term load prediction. [Conclusions] Clustering, linear compensation, and Fourier-based residual correction are integrated in the improved grey model. The classification foundation is provided for refined user management by K-means clustering. The traditional model's lack of linear fitting capability is effectively compensated for by linear regression, while noise and systematic errors are significantly reduced by Fourier-based residual correction. A substantial improvement in the model's accuracy and generalization capability is led to by the combination of the three elements. The model demonstrates excellent performance in short-term load prediction, holding practical significance in real-time electric power dispatch, demand response, economical energy usage, and cost reduction. The improved model is mainly applicable to short-term electric power load prediction, and future research will explore the integration with machine learning or the introduction of more factors to enhance its ability for medium-to-long-term electric power load prediction.

[Objective] Power supply side faults in distribution networks are prone to cause three-phase current unbalance, which may further lead to regional power supply interruption and threaten the safe and stable operation of the power grid. Traditional load transfer strategies usually only focus on power balance, without fully considering the impact of negative sequence current distribution on system recovery, which tends to result in disadvantages such as low recovery efficiency and frequent switch operations. To address this problem, aiming at the scenario of power supply side faults in distribution networks, a regional maintenance load transfer method was proposed to achieve efficient grid recovery after faults and optimize load loss, network loss, and the number of switch operations. [Methods] The symmetrical component method was adopted to calculate the negative sequence current in the fault composite sequence network. By comparing the phase difference between the fault phase voltage phasor and the negative sequence current and setting a threshold value, the area to be restored and power outage area were accurately located. After the area division, load transfer was carried out. A multi-objective load transfer optimization model was established, which took the minimization of combined load loss, the number of tie switch operations, and additional network loss as the objectives, with six constraints including topological structure and load controllability set. In view of the conflicts among multiple objective functions, a hybrid algorithm combining genetic algorithm and heuristic search algorithm was used for solution, which could effectively avoid falling into local optimal solutions. Based on the topological analysis and judgment results, available tie switches and preliminary load transfer schemes were searched. Combined with various influencing factors, the hybrid algorithm was applied to obtain the optimal transfer path and generate the best load transfer scheme. By adjusting the states of tie switches, the load was transferred from the area to be restored to the power outage area, which thereby restored power distribution and supply. [Results] To verify the effectiveness of the proposed method, a simulation test was carried out on a 10 kV distribution network in a certain area. The test results show that the average load loss of the proposed method is only 0.215 3 kW, and the additional network loss is relatively small. The minimum number of tie switch operations is only 9 times, and the voltage of each node is generally maintained at a high level, which ensures power supply reliability. [Conclusions] By accurately analyzing the negative sequence current distribution, reasonably dividing the fault-affected areas, and realizing load transfer based on multi-objective optimization, the proposed method effectively avoids the performance trade-off phenomenon in traditional strategies. This method can balance power supply reliability, economy, and rapid recoverability, providing a new technical path for handling power supply side faults in distribution networks, which has significant engineering practical value and popularization prospects. Future research will further combine with dynamic topological reconfiguration and intelligent prediction technologies to enhance its adaptability in complex fault scenarios.

[Objective] Mainly relying on manual inspection, traditional power inspection methods have such problems as low efficiency, high cost, and great danger, and it is difficult for them to meet the requirements of modern power systems for efficient, safe, and intelligent inspection. In recent years, the rapid development of unmanned aerial vehicle (UAV) technology provides a new solution for power inspection. UAVs have the advantages of strong flexibility, wide coverage, and relatively low cost, which can effectively improve the inspection efficiency and reduce the manual inspection risk. However, UAV power inspection systems still face many challenges in practical applications, especially in terms of precise positioning, navigation, and data transmission under complex environments. As a global navigation satellite system independently developed by China, the BeiDou Navigation Satellite System (BDS) has the characteristics of high precision, high reliability and global coverage, providing powerful technical support for UAV power inspection systems. [Methods] By introducing BeiDou satellite technology, a UAV power inspection system that could maintain high precision and stability in complex scenarios was designed to improve the monitoring and maintenance efficiency of power equipment. The core of this paper is to reconstruct the hardware framework of UAV power inspection systems for the deep integration with BeiDou satellite technology. Based on the hardware framework reconstruction, a software algorithm based on the PPP-RTK function model of BeiDou satellite positioning was designed. This algorithm could obtain high-precision position information of UAVs in real time, thus effectively overcoming the influence of complex environments on inspection precision. By implementing this technical route, the stable and precise inspection of UAVs in complex scenarios was realized. In the research process, targeted reconstruction was conducted on the hardware framework to ensure that UAV could stably receive and efficiently process BeiDou satellite signals, and the system performance was fully verified by employing a large amount of experimental data. [Results] The experimental results show that in complex scenarios, the proposed UAV power inspection system can significantly improve the precision and stability of inspection results, and effectively reduce the influence of environmental factors on inspection quality. The effectiveness and superiority of the UAV power inspection system integrated with BeiDou satellite technology in complex scenarios were verified. [Conclusions] By conducting the reconstruction design of the hardware framework and software algorithm optimization, the inspection ability of UAVs in complex environments are notably improved, thus providing more reliable technical support for the monitoring and maintenance of power equipment. The innovation of this paper lies in the introduction of BeiDou satellite technology to the UAV power inspection system, and the realization of high-precision positioning based on the PPP-RTK function model, thereby effectively solving the inspection problem of traditional systems in complex scenarios. This paper not only improves the precision and stability of UAV power inspection but also provides a new technical path for the intelligent and precise monitoring and maintenance of power equipment, which has theoretical and practical significance.

[Objective] Firmware security in power Internet of Things (IoT) devices is crucial for ensuring the stable operation of critical infrastructure. However, existing vulnerability detection methods suffer from limited accuracy and adaptability due to complex firmware characteristics and reliance on a single analysis dimension. To address these issues, a multi-granularity vulnerability detection method for smart power IoT firmware suitable for the ubiquitous IoT background was proposed to improve the comprehensiveness and accuracy of vulnerability detection. [Methods] First, an i2vBi model was designed to map address space operands into eight classes to control the loading base address range, thus accurately generating instruction word vectors. The Softmax function was used to calculate contextual word probabilities, a maximum likelihood estimation model was trained, and instruction vectors were aggregated through a bidirectional long short-term memory (BiLSTM) network to obtain basic block embedding vectors containing forward and backward semantic information. Second, basic block embeddings were used to construct attribute control flow graphs to extract fine-grained structural features within functions. Furthermore, the principal neighborhood aggregation (PNA) algorithm was adopted, combining multiple aggregators and node-degree-based scalers to adaptively aggregate node neighbourhood information, generating more expressive graph embedding vectors and achieving function-level meso-granularity feature extraction. Subsequently, a convolutional neural network (CNN) and a self-attention mechanism were used to extract local pattern features of function execution order from graph embedding vectors, and these sequential features, together with attribute control flow graph features constructed from basic block embeddings, were input into a multilayer perceptron for fusion to form the final comprehensive feature vector. Finally, a semantic analysis dimension was introduced, in which known vulnerable functions were transformed into natural language text. A semantic embedding model based on bidirectional encoder representations from transformers (BERT) was used for masked modeling and mean pooling to generate semantic vectors. The cosine similarity between the semantic vectors and the comprehensive feature vectors of target functions was computed, and multi-granularity vulnerability detection based on semantic similarity was achieved by setting a threshold. [Results] To verify the effectiveness of the proposed method, experiments were conducted on a dataset containing real power IoT firmware images. The experimental results show that the AUC value of the proposed method remains stable between 0.85 and 0.95, which is significantly higher than that of comparative methods, demonstrating excellent overall classification performance. The Kappa coefficient lies in the high range of 0.85-0.95, indicating a high degree of consistency between detection results and actual conditions. The Hamming distance remains at a low level, indicating that false positive and false negative rates are effectively controlled, and prediction results are more accurate. [Conclusions] The proposed method effectively overcomes the limitation of a single feature dimension by integrating multiple levels of features, including instructions, basic blocks, function control flow, and semantics. This method not only significantly improves the accuracy and robustness of vulnerability detection but also exhibits better environmental adaptability due to its understanding of code semantics. The research results provide a reliable technical approach for automated and intelligent security analysis of smart power IoT firmware and have positive significance for enhancing the overall security and stability of power IoT systems.

[Objective] The size and needs of the visually impaired groups cannot be ignored. The application of tactile sensing in the field of assisted reading is particularly noteworthy. This technology can not only be integrated into robots or prosthetic systems, but also provide an effective braille reading tool for blind or visually impaired groups. Therefore, research on vision aids and braille recognition technology with information interaction competency is significantly valuable for offering technical support for visually impaired groups. [Methods] Based on biomimetic principles, the function of biomimetic hair was simulated. With magnetic iron-gallium wire as biomimetic hair and Hall elements as receptors at the hair roots, a biomimetic electromagnetic tactile sensor was designed according to the size of braille dots. Based on magnetization intensity and magnetic induction intensity theories as well as mechanical equations, the relationship curve between the sensor's applied force and output voltage was deduced. A dynamic characteristic testing system was constructed, which consisting of a signal generator, a power amplifier, a vibration exciter, data acquisition card, a computer and a direct current stabilized voltage supply. The dynamic characteristics of the tactile sensor were tested. [Results] Test results show that the tactile sensor can convert applied force into an electrical signal within the range of 0-1.5 N. Within the applied force range of 0-1.5 N, the output voltage gradually increases with higher force. When the contact force is less than 0.5 N, the two approach a linear relationship. The sensor exhibits high stability in output voltage under an applied force of 0.2-1.4 N at a frequency of 1 Hz. Under an applied force of 1.0 N at 1 Hz, its sensitivity is 34.5 mV/N. When the applied force is 0.5 N at 1 Hz, the response time and recovery time are 20 ms and 18 ms, respectively. The designed biomimetic electromagnetic tactile sensor was applied to establish a braille recognition system consisting of a two-finger robotic hand, a motor-driven slide, a data acquisition card, and a computer. The correspondence between braille letters and the output voltage waveform was determined by scanning the braille dots. [Conclusions] The output characteristics of the developed biomimetic electromagnetic tactile sensor were tested. The experimental results show good agreement with the calculated values, indicating that the calculated model can describe the relationship between the applied force and the output voltage. The designed tactile sensor features high stability, high sensitivity, and fast response speed, making it suitable for detecting both static and dynamic applied forces. The braille recognition system was used to determine the voltage waveform corresponded to the braille letters. It is pointed out that voltage waveform peak count, peak intensity, and peak initiation time can serve as the criteria for recognizing the braille letters, demonstrating that the braille recognition system can recognize braille letters. The research results can provide new braille recognition tools and technical pathways for visually impaired groups, allowing for the deep integration of tactile sensors and recognition technology to build assistive technology support systems for visually impaired groups.

[Objective] A substation cost prediction model based on an improved long short-term memory (LSTM) network, spatio-temporal long short-term memory (ST-LSTM) network, was proposed to address the limitations of existing substation cost estimation methods in terms of prediction accuracy and computational efficiency, thus improving the cost prediction accuracy and efficiency. Based on building information modeling (BIM) data, a dual-stream memory conversion mechanism and ZigZag spatio-temporal memory flow are introduced in the model to effectively capture and learn the complex dynamic features in spatio-temporal data, thereby achieving unified modeling of the short-term spatial detail changes and long-term temporal dynamic evolution process. [Methods] Firstly, BIM cost data were preprocessed, including data cleaning, standardization, and time series division, to ensure data integrity and usability. Then, the ST-LSTM network model was built. By improving the triple gating mechanism of traditional LSTM networks, the ZigZag spatio-temporal memory flow and dual-stream memory conversion mechanism were introduced to enhance the model's ability to extract and fuse spatio-temporal features. At the model training phase, the grid search method was adopted to optimize the number of hidden layer neurons, mean square error (MSE) was employed as the loss function, and the Adam optimizer was combined to complete the updating of model parameters. In the experiment, the actual BIM cost data from 105 substations were selected and divided into the training set, validation set, and test set in a 3∶1∶1 ratio for model training and performance evaluation. [Results] By carrying out multiple rounds of simulation experiments, the prediction performance of the ST-LSTM network model, particle swarm optimization (PSO) algorithm, and traditional LSTM network model was compared and analyzed. Additionally, the mean absolute percentage error (MAPE), root mean square error (RMSE), and Pearson correlation coefficient were adopted as the evaluation indexes. The results show that the highest accuracy of the ST-LSTM network model is approximately 95％ in short-term prediction and more than 90％ in long-term prediction, and its overall average prediction accuracy exceeds 90％. Thus, this model significantly outperforms the PSO algorithm and traditional LSTM network models. In terms of computational efficiency, the average running time of the ST-LSTM network model is 1.1 s, slightly longer than 0.5 s of the PSO algorithm but shorter than 1.2 s of the traditional LSTM network model. However, the values are all within an acceptable range of engineering applications. Further analysis reveals that the ST-LSTM network model demonstrates more significant prediction advantages when dealing with large-scale datasets characterized by complex spatio-temporal features. [Conclusions] The substation cost prediction method based on the ST-LSTM network model can effectively extract and fuse multidimensional spatio-temporal features, significantly improving the accuracy and overall computational efficiency of short-term and long-term cost prediction. Compared with the PSO algorithm and traditional LSTM network models, the ST-LSTM network model has a significant advantage in prediction performance, but its computational complexity is relatively high, thereby putting forward higher requirements for computing resources and training time. Future research will focus on model structure optimization and computational complexity reduction to enhance the model's application feasibility and promotional significance in engineering practices.

[Objective] Against the backdrop of the global energy transition and the “dual carbon” targets, energy demand in cold regions has risen sharply due to natural conditions such as low temperatures, snowfall, and permafrost, which adversely affect the efficiency, service life, and operational safety of energy storage systems. To develop new energy storage technologies suitable for cold regions, it is urgently needed to establish a comprehensive performance evaluation system covering the entire service lifecycle to support industrialization and green energy utilization. This study aimed to establish an evaluation framework and conduct a technical comparison to promote the rational deployment of different energy storage pathways, thus addressing regional energy security and sustainable development requirements. [Methods] This study focused on four major categories of energy storage technologies:electrochemical, mechanical, thermal, and hydrogen-based energy storage. It systematically analyzed five evaluation indicators:economic performance, technical performance, safety and reliability, environmental friendliness, and lifecycle energy efficiency. In terms of economic performance, total capital cost and levelized cost analysis were employed, with a low-temperature correction factor introduced. In terms of technical performance, low-temperature adaptability, energy efficiency, and cycle life were analyzed. In terms of safety and reliability, a multi-dimensional safety grading system was constructed using fault diagnosis methods. In terms of environmental friendliness, the entire chain from raw material acquisition to manufacturing and power plant construction was considered. Lifecycle energy efficiency was measured using the energy return on investment (EROI) and energy storage on investment (ESOI) indicators. Through horizontal comparison and vertical analysis, this study revealed the differential impacts of cold environments on the performance and cost of various energy storage technologies. [Results] The results show that lithium-ion batteries have reduced lifespan and safety under low temperatures. Flow batteries have long lifespans and high cycling stability, but electrochemical energy storage carries the risk of thermal runaway. Compressed air energy storage offers significant advantages in long-term energy supply and cost controllability, and its lifecycle energy efficiency is higher than that of electrochemical energy storage, but is constrained by geological conditions. Flywheel energy storage has a fast response speed but high manufacturing and maintenance costs at low temperatures. Sensible heat storage and chemical heat storage in thermal energy storage systems show potential for seasonal regulation in cold regions. Hydrogen energy storage, with its high specific energy and multi-energy coupling characteristics, demonstrates unique value in inter-seasonal peak regulation and microgrid applications, but faces challenges such as high-pressure leakage and hydrogen embrittlement. [Conclusions] Future efforts should intensify low-temperature adaptation measures and strengthen subsidies for temperature control, promote the development of multi-energy complementary energy storage dispatch centers, enhance overall system resilience, and establish a standard system for energy storage in cold regions covering the entire construction, operation, and end-of-life recycling, so as to achieve a balance between economic efficiency and safety and promote the healthy development of the energy storage industry in cold regions.

[Objective] In power systems, load data analysis is crucial for power grid dispatching, planning, and management. However, with the deepening of the complexity and intelligence of power systems, power load data exhibit characteristics of high dimensionality and sparsity, which poses significant challenges to traditional data analysis methods in terms of processing efficiency and ability to capture the intrinsic information of load changes. An efficient unsupervised data mining algorithm was proposed in this paper, which was aimed at improving the processing efficiency and information extraction capability of high-dimensional sparse power load data. [Methods] Firstly, a feature ranking method based on information entropy was adopted to determine feature importance. Data initialization was completed by calculating mutual information and conducting centralization and standardization. Features with the maximum mutual information were selected to expand the feature set, and feature subsets were screened by calculating relevant information entropy. The subset screening process was optimized using support vector machine (SVM)classifier as the benchmark model, and an improved particle swarm optimization algorithm was introduced for secondary feature selection. Meanwhile, the SVM classifier was used to complete the preliminary feature screening. Secondly, the principal component analysis (PCA) was introduced for dimensionality reduction. The sample matrix was centralized, and the covariance matrix was established. Eigenvalues and eigenvectors were obtained, and eigenvectors were selected to construct a new matrix to achieve dimensionality reduction. Finally, an autoencoder network based on unsupervised learning was introduced to conduct unsupervised mining. In the encoding stage, input data were converted into feature representations. In the decoding stage, data recovery was completed. Through steps such as data setting, clustering execution, data point screening, data balancing processing, and model training to obtain a classification interface, hidden feature extraction and network adjustment were realized. [Results] When the algorithm in this paper is applied, the Rand index values all exceed 0.60, indicating high clustering accuracy. In 60 iterations of experiments, the maximum memory overhead ratio is about 8.3％, demonstrating the algorithm's high efficiency in computing resource utilization. Compared with other traditional methods, this algorithm can achieve higher processing efficiency and better mining results when dealing with high-dimensional sparse power load data. [Conclusions] The unsupervised mining algorithm performs excellently in the analysis of high-dimensional sparse power load data. By reducing computational complexity through feature selection and dimensionality reduction, and mining nonlinear features with the autoencoder network, it significantly improves the accuracy and efficiency of data mining, and has strong applicability and feasibility. Its innovation lies in integrating multiple methods such as information entropy-based feature ranking, SVM, improved particle swarm optimization, PCA, and autoencoder network to form a complete system from feature processing to data mining. This system can not only effectively address the challenges in mining high-dimensional sparse power load data but also provide a new and effective means for load data analysis in power systems, which is of great significance for promoting the intelligent development of power systems.

[Objective] Dry-type air-core reactors (DARs) are prone to inter-turn short circuits during operation. Characterized by low detectability and limited real-time observability, such faults pose a serious threat to the safe and stable operation of power grids. Although widely adopted in engineering, detection methods based on magnetic flux density amplitude exhibit limitations. Specifically, they are insensitive to inter-turn short circuits in inner windings, and detection accuracy is significantly degraded by magnetic field interference from adjacent equipment when multiple DARs operate in parallel. To enhance fault identification capability and achieve early warning, a DAR inter-turn short circuit detection method based on the magnetic dield direction angle (MFDA) was proposed. [Methods] Based on finite element simulation technology, the surrounding magnetic field distribution of a single-phase DAR under normal operation and inter-turn short circuits at different locations was analyzed. By arranging multiple monitoring points around the DAR, magnetic flux density data were collected, and the variation patterns before and after the fault were analyzed. Based on the principle of magnetic field superposition, the temporal variation patterns of MFDA at each monitoring point near the faulty phase in a three-phase DAR were calculated and analyzed. By comparing the MFDA data of a single DAR with those of multiple parallel DARs, the interference level of adjacent normally operating DARs on fault detection was quantitatively evaluated, and the placement of monitoring points was optimized. [Results] Simulation results indicate that after an inter-turn short circuit occurs in a single-phase DAR, both the amplitude and direction of the surrounding magnetic flux density undergo significant changes. Compared with the amplitude-based detection method, MFDA exhibits higher response sensitivity to inter-turn short circuits and can accurately capture the changes in electromagnetic characteristics caused by faults. By analyzing the temporal variation of MFDA at each monitoring point before and after the fault and establishing a fault decision threshold, inter-turn short circuits in single-phase DARs can be diagnosed. For the scenario of three-phase DARs operating in parallel, arranging monitoring points near the faulty phase and on the symmetric perpendicular bisectors of the normal phases can effectively mitigate magnetic field interference from adjacent DARs. By combining the temporal characteristics of MFDA with the fault decision threshold, short-circuit faults can still be diagnosed. [Conclusions] The proposed MFDA-based detection method can effectively address the limitations of the magnetic flux density amplitude detection method in terms of sensitivity and anti-interference capability, providing a reliable technical approach for real-time monitoring of DAR operating status and fault early warning. Based on the simulation results, the proposed method shows strong potential for engineering applications and provides practical support for improving the operation and maintenance of power grid equipment and ensuring the safe and stable operation of power systems.

[Objective] As the digital transformation of substations accelerates, traditional evaluation methods have limitations in the accuracy and adaptability of reliability analysis, and the reliability of substations is of great significance for the stable operation of the power system. To this end, a reliability analysis method based on the improved dynamic Bayesian network (DBN) for a 110 kV digital substation model was proposed to enable real-time monitoring and accurate evaluation of the system's status. [Methods] Firstly, statistical analysis of the key parameters was conducted, such as failure rates of various types of equipment and components within the substation, thus constructing the basic data for reliability evaluation. Secondly, DBN was introduced as the modeling tool. The network structure was dynamically adjusted and redesigned in response to environmental factors such as temperature, humidity, and load fluctuations to enhance the model's adaptability to a non-stationary operating environment. Finally, fault tree analysis (FTA) was employed to identify logical relationships of system-level faults, and the results were systematically mapped into DBN to build a probabilistic reasoning model with both hierarchical and causal characteristics. By adopting probabilistic reasoning to compensate for information gaps, reasoning robustness and accuracy can be improved by this method under incomplete information or missing data. [Results] Experiments conducted on the 110 kV digital substation model show that the area under the ROC curve of the proposed method is the closest to 1, indicating that the analysis results are the closest to the actual values. Meanwhile, it exhibits the lowest error rates and stronger stability in the reliability analysis of the three substations, with the accuracy, precision, recall, and F₁ scores being 0.891, 0.875, 0.904, and 0.889, respectively. Thus, its overall performance is better than the comparative methods. [Conclusions] The proposed method exhibits significant advantages in terms of accuracy, stability, and adaptability. By integrating the structured modeling capabilities of FTA with the adaptive reasoning mechanism of DBN, it effectively overcomes the limitation of insufficient evaluation accuracy of traditional methods in dynamic environments and information deficiency conditions. This method not only achieves dynamic quantification of reliability indexes for substation digital models but also provides a reliable theoretical support and practical tool for system status monitoring and intelligent operation and maintenance, with promising engineering application prospects.

[Objective] Under the impetus of the carbon peaking and carbon neutrality goals, the high proportion of renewable energy grid connection has weakened system inertia and damping characteristics, posing a serious threat to grid stability. Although grid-forming virtual synchronous generator (VSG) control technology can actively provide inertia support for the grid, the complex nonlinear characteristics of renewable energy systems cause traditional VSG control to face risks of instability in angular frequency and output voltage under non-ideal operating conditions, resulting in suboptimal control performance. To address this issue, this paper proposed a grid-forming converter control strategy based on passivity-based control, overcoming the limitations of traditional linear control methods to enhance system dynamic response performance, interference resistance, and robustness. [Methods] This paper adopted a nonlinear control design framework. According to VSG control principles, the rotor motion equation of a synchronous generator was employed to achieve active frequency regulation, and the excitation system was used to achieve reactive voltage control. A Hamilton system model incorporating grid-forming control was built. Based on the core principles of passivity-based control theory, the Hamilton model was mathematically transformed into a dissipative Hamilton standard form with port characteristics. This model inherently possesses advantages for stability analysis, providing a theoretical foundation for controller design. In addition, based on the system′s stable operation requirements, the desired equilibrium operating point was set. To accelerate the dissipation of system energy toward the desired equilibrium point, effectively suppress oscillations, and enhance convergence speed, a damping term was introduced. Ultimately, the active and reactive control laws applicable to grid-forming converters were derived, achieving global asymptotic stability of the nonlinear system. [Results] Simulation test results show that under non-ideal conditions such as power change, grid voltage imbalance, short-circuit fault, and load variation, the grid-forming converter control strategy based on passivity-based control significantly outperforms traditional VSG control in terms of system angular frequency stability. The amplitude of frequency fluctuations is significantly reduced, and the time required to recover to the desired value is greatly shortened. The overshoot of the output voltage is reduced, the regulation process is smoother, and the voltage stabilizes faster, effectively enhancing voltage stability. [Conclusions] The proposed grid-forming converter stability control strategy establishes a nonlinear design framework based on the dissipative Hamilton model and designs a control strategy through energy shaping and damping injection. The derived control laws exhibit strong robustness and can accommodate complex operating conditions without requiring a precise system model. It effectively addresses the stability deficiencies and slow dynamic response of traditional VSG control when faced with system nonlinearities, overcoming the limitations of fixed linear control parameters. This provides a control strategy with strong interference resistance and fast dynamic response for renewable energy grid-connected systems, supporting the stable operation of new power systems.

[Objective] In the power system, transmission lines serving as a key component are often prone to failures caused by natural disasters such as lightning strikes due to their remote location and long distance. Lightning strokes on transmission lines in China account for about 50％ of the total accidents, which can cause tower insulation flashover, abnormal voltage, and even power supply interruption. Additionally, they can also damage electronic equipment and cause significant economic losses. An analysis of flashover voltage characteristics of 110 kV line insulators was conducted to improve the design of transmission lines and enhance their lightning resistance, thus strengthening the resistance of transmission lines to lightning strikes. [Methods] Based on the electric field theory and the energy conservation principle, the energy conservation equation of electrons in the electric field was constructed. Combined with the motion equations of three types of particles (positive attribute particles, negative attribute particles, and neutral particles) and Poisson′s equation, a mathematical model for lightning strokes on insulators was built. By adopting this model to calculate the lightning current overvoltage, the standard waveform of the current of the lightning impulse transmission line was obtained. At the same time, the relationship between the lightning resistance level of transmission lines and flashover voltage of insulator strings was analyzed. By analyzing the characteristics of lightning flashover voltage of insulators via overvoltage, the current value and lightning current overvoltage at the lightning strike point were obtained. Additionally, silicone rubber insulators for 110 kV lines were selected for experiments to simulate different parameters such as the call height, grounding resistance, lightning current waveform, and insulation distance, and analyze the influence of each parameter on the voltage waveform. [Results] The experimental results show that the voltage of the insulator gradually flattens after 12 μs when the insulator is subjected to different waveform lightning strikes. Specifically, the double exponential voltage peak is the largest, the oblique angle model has the best effect on voltage control, and the voltage waveform tends to be more stable under the insulation distance greater than 5 m. Compared with other methods, the voltage waveform analyzed by this method is closer to the actual situation, with an error of less than 1 kV and higher accuracy. [Conclusions] Under the tower height of 25 m, the insulator is the most sensitive to lightning impulse response, and the peak voltage increases with the rising grounding resistance. The oblique angle lightning current model has the best effect on voltage control, and the insulation distance greater than 5 m can improve the insulator′s ability to resist lightning impulse. In this study, mathematical models were combined with electric field theory, the influence of various parameters on insulators was comprehensively considered, and a more comprehensive and accurate voltage waveform and mathematical equation for insulator strings was established. The results provide scientific basis for the design of transmission lines, and hold engineering and reference significance for improving the lightning resistance of transmission lines and ensuring the safe and stable operation of power systems.

[Objective] Transmission lines operating in regions with strong lightning activity are highly vulnerable to lightning strikes. Double-circuit lines on the same tower feature compact structures and significant electromagnetic coupling effects, resulting in consistently high lightning fault rates. Existing lightning protection measures rely heavily on statistical experience and cannot effectively distinguish different types of lightning faults such as shielding failures and back flashovers, making precise protection difficult. Consequently, line tripping accidents remain a recurring problem, posing a serious threat to the safe and stable operation of the power grid. To address this issue, this study proposed a deep learning-based lightning fault identification method with high-accuracy automatic identification of shielding failure and back flashover, providing effective technical support for differentiated lightning protection in transmission lines. [Methods] A lightning fault simulation model for 220 kV double-circuit transmission lines on the same tower was developed using the electromagnetic transient simulation software ATP-EMTP to obtain overvoltage response data under different lightning current amplitudes and grounding resistance conditions. To address the non-stationarity and mode mixing of lightning signals, ensemble empirical mode decomposition (EEMD) was introduced, where Gaussian white noise was added to suppress mode mixing. The first four intrinsic mode functions (IMFs) were extracted to preserve the major characteristic components. Subsequently, frequency slice wavelet transform (FSWT) was applied to compute multi-band energy ratios, which, together with lightning current amplitude and grounding resistance, formed a multidimensional feature set. In terms of classification modeling, a CNN-LSTM-Attention deep learning architecture was proposed:CNN extracted spatial features, LSTM modeled temporal dependencies, and the Attention mechanism focused on critical information, thus enabling effective fusion and identification of complex signal features. [Results] Experimental results demonstrate the excellent performance of the proposed method in distinguishing shielding failure from back flashover. The overall identification accuracy reaches 98.6％, with both precision and recall exceeding 98.5％, and an F₁-score of 0.99. Compared with benchmark models such as SVM and CNN, the proposed method exhibits a clear advantage in identification accuracy. Results from 10 independent comparative experiments show an average accuracy of 99.7％ and a variance of 0.000 93, fully verifying its stability and reliability. [Conclusions] The lightning fault identification method based on EEMD-FSWT feature extraction and the CNN-LSTM-Attention fusion model effectively characterizes the time-frequency features of lightning signals of double-circuit transmission lines on the same tower, achieving high-accuracy differentiation between lightning shielding failure and back flashover. This method not only improves the accuracy and timeliness of fault diagnosis but also provides important data support for formulating differentiated lightning protection strategies. The research results have significant engineering application value and strong potential for wide application in reducing lightning-induced line tripping and ensuring the safe and stable operation of power systems.

[Objective] Relay protection equipment plays a crucial role in the operation of power systems. However, with the long-term operation of the equipment, aging and damage are inevitable, which will cause abnormal displacement data in relay protection equipment. If these abnormal data cannot be properly processed, the safe and stable operation of power systems will be affected. Therefore, how to effectively handle the abnormal displacement data of relay protection equipment has become a problem to be urgently solved. [Methods] An autonomous controllable fault-tolerant storage algorithm was proposed. Firstly, the state of the relay protection equipment was evaluated by this algorithm, with the potential disturbance factors analyzed. On this basis, the model predictive control (MPC) technology was adopted to predict the possible abnormal data. Based on the dynamic model of the system, decisions were made by MPC in advance via predicting the future state of the system. Then, the predicted abnormal data were corrected to restore the data to their original state. Meanwhile, the data elasticity theory and granularity rate were employed to calculate the compensation storage intensity. The data elasticity theory helped to measure the tolerance ability of the system in the face of faults, and the granularity rate was related to the data refinement degree. The accuracy and integrity of the data were ensured by the combination of the two. During the study, an experimental environment based on the above-mentioned algorithm was constructed, and the algorithm was tested by simulating the abnormal displacement data generated by the relay protection equipment in different working conditions. [Results] The algorithm′s efficacy was verified by the experiment. The fault-tolerant rate of the algorithm is above 0.89, which means that under a large amount of abnormal data, most of the data errors can be successfully handled by the algorithm. The proportion of memory required for storage is less than 20 MB, indicating that the algorithm occupies fewer memory resources for data storage. Under the data number of 10 000, the data transmission number is only 401, which reflects the high efficiency of the algorithm in data transmission. [Conclusions] It can be concluded from this study that the data fault-tolerance ability of relay protection equipment can be effectively enhanced by the proposed autonomous controllable fault-tolerant storage algorithm. The accuracy and integrity of the data are ensured by accurate prediction, correction of abnormal data and a reasonable storage strategy, thereby enhancing the ability of relay protection equipment to cope with equipment aging and damage. The reliability of relay protection equipment can be improved by the application of this algorithm in power systems to further ensure the safe and stable operation of power systems. The innovation of this study lies in the combination of MPC, data elasticity theory and the granularity rate to develop a brand-new fault-tolerant storage algorithm. This method of comprehensively adopting multiple technologies has unique advantages in handling the abnormal displacement data of relay protection equipment. By adopting the proposed algorithm, the data processing ability of relay protection equipment can be improved, and the risk of power system failures caused by data abnormalities can be reduced, which is of great significance for ensuring the safe and stable operation of power systems.

[Objective] Under the background of new power system construction, the randomness and fluctuation of the system are significantly exacerbated by the large-scale integration of high-proportion renewable energy and widespread popularity of flexible loads. Coupled with the continuously expanding power grid scale, control variables increase sharply, thus posing a severe challenge to the control strategies of traditional power grid voltage and tidal currents. The electrical distance between nodes is mainly relied on the existing power grid partitioning methods for reactive partitioning, which is difficult to adapt to the operation requirements for new power systems with drastic source-load changes. To this end, a power grid partitioning optimization method comprehensively considering multiple factors was proposed to lower the overall control difficulty of power grids under the penetration of high-proportion new energy and improve the autonomous operation capability of partitioning. [Methods] The core of this study is to build a set of partitioning index system and optimization model, and break through the traditional partitioning′s limitation of only focusing on the topological association. Additionally, the tightness of internal electrical connections and the degree of source-load matching were creatively considered, with the reactive partitioning indexes based on electrical distance and active partitioning indexes based on source-load matching constructed respectively. On this basis, the optimization model of power grid partitioning was built to minimize the reactive partitioning index, thus aiming to maximize the electrical tightness inside the partitioning and simplify reactive power and voltage control. Meanwhile, the key constraint that the active partitioning indexes satisfied the requirements was employed to limit the frequent interaction of active power between partitions, reduce the violent fluctuation of net loads within partitions, and ensure the source-load balance within partitions. At the same time, a bionic joint optimization algorithm was proposed, in which the global search ability of genetic algorithms and fast local refinement ability of firefly algorithms were fully used to efficiently solve the built nonlinear complex optimization model, improve the optimization speed, and avoid falling into the local optimal solutions. [Results] The standard IEEE 39-node system was adopted to verify the case example. The simulation results show that by adopting this algorithm, the source-load matching degree within partitions can be significantly improved, the fluctuation of net loads between and within partitions can be reduced, and unnecessary tidal current interaction can be decreased. Additionally, the difficulty of reactive control in the system can be lowered, the electrical tightness of nodes within the partitions can be enhanced, and the voltage and reactive regulation process within the partitions can be simplified by employing the algorithm. The proposed firefly-genetic bionic joint optimization algorithm exhibits excellent solution performance and can obtain the optimized partitioning scheme rapidly and effectively. [Conclusions] There are two main innovative points in this study. Firstly, the optimization objective of reactive control based on electrical distance and the constraint of active balance based on source-load matching are integrated in the power grid partitioning model, which overcomes the defect of insufficient adaptability of traditional methods to source-load changes. Secondly, an efficient and robust firefly-genetic bionic joint optimization algorithm was proposed to solve the partitioning model, thus effectively improving the optimization speed and accuracy. This algorithm provides a new technical way to solve the problem of partitioning operation control under the complex network structure of new power systems, and holds theoretical and practical significance for improving the safe and stable operation of power grids and promoting efficient consumption of new energy.

[Objective] The influence of photovoltaics (PV)-assisted current and extraction current on conventional relay protection hinders the effective functioning of relay protection equipment. A multi-objective setting method for relay protection in distribution networks suitable for conditions with high permeability of distributed PV was proposed to address this problem, which is aimed at enhancing the rapidity, sensitivity, and selectivity of protection, and ensuring economic viability and practicality, thus effectively safeguarding the power grid security and supporting the widespread access of distributed PV. [Methods] The influence of PV-assisted current and extraction current on the protection configuration of the distribution networks was analyzed, and the problem of unwanted operation and refuse operation of distribution network protection caused by PV access was avoided by introducing distance protection and instantaneous current protection as the protection criteria. A multi-objective optimization model with the optimal parameters of protection rapidity, sensitivity and selectivity was built, and the particle swarm optimization (PSO) algorithm was improved by adopting the dynamic splitting operator to make the solution of the protection setting meet the practical application requirements. [Results] High-permeability distributed PV results in unwanted operation or refuse operation of distribution network protection, which is effectively avoided by introducing distance protection and instantaneous current protection as the protection criteria. The multi-objective optimization protection configuration model was built, and the evaluation indexes of the overall protection effect of a certain area were formed, with the solution of the protection setting completed based on PSO algorithm. Finally, the overall evaluation of the protection effect under high-permeability PV access was realized, with the rapidity, sensitivity, and selectivity of protection improved. [Conclusions] The results show that the combination of distance protection and instantaneous current protection can effectively avoid the influence of the PV-assisted effect on the conventional instantaneous current protection. The protection performance can be effectively improved by the proposed multi-objective optimization scheme. Under the equilibrium strategy, the rapidity, sensitivity, and selectivity increase by about 82.2％, about 3.8％, and about 33.1％, respectively. The innovation of this study is that the combination of distance protection and instantaneous current protection was adopted to form the protection criteria, thus avoiding the problem of unwanted operation and refuse operation of the distribution network protection due to PV access. Additionally, a multi-objective optimization scheme for protection settings was constructed, and PSO algorithm was improved by employing the dynamic splitting operator, thereby avoiding the limitations of PSO algorithm and improving the reliability and applicability of protection settings.

[Objective] With the development of artificial intelligence (AI) and 5G technology, data centers are called important infrastructure facilities for future development. However, there is a prominent contradiction between the high energy consumption characteristics of data centers and current low-carbon development needs. The optimization of relying solely on clean energy power supply faces bottlenecks, such as difficult waste heat recovery and low efficiency. In response, a large-scale hydrogen production capacity configuration method was proposed to study waste heat recovery in data centers, offering a new approach for the green low-carbon development of data center energy supply systems and enhancing green hydrogen production efficiency. [Methods] Firstly, the energy consumption structure of the data center was analyzed. The mathematical model of the output heat energy of the water-cooled data center was constructed, and the mathematical model of the influence of the electrolyte temperature change on the efficiency of electrolytic hydrogen production was established. The electrolyte temperature rise coefficient was proposed as the coupling node of the data center and electrolytic hydrogen production, which laid the foundation for the subsequent establishment of the waste heat water coordination mechanism. Secondly, the data center′s energy consumption characteristics and hydrogen needs were analyzed. Using waste heat recovery and heating electrolyte theory, the data center+clean energy+green hydrogen operation architecture and the matched mode were built, and a dynamic supply-demand balance model was established. Based on clean energy output and data center operations, various electrolytic hydrogen production modes were formulated. Finally, considering the data center load characteristics, clean energy output fluctuations, hydrogen energy market demand, and other factors, a hydrogen production capacity configuration model was constructed with system economy, carbon emissions, and renewable energy consumption rate as optimization objectives. The multi-objective optimization method based on the improved timing difference algorithm and particle swarm optimization algorithm was designed, and simulation analysis was carried out with Matlab. [Results] The data center+clean energy+green hydrogen coordinated operation mode can reduce the annual electricity energy consumption by 2.59％ under the typical day scenario. The pricing method of the auxiliary peak shaving market can guide the system to operate according to different objectives and can deal with the economy of the system to varying degrees and low carbon demand. This study achieves the structural transformation of the data center energy system through multi-dimensional technological innovation. [Conclusions] The proposed large-scale electrolytic hydrogen production capacity configuration method meets waste heat utilization needs of data centers and reduces reliance degree of the energy supply system on fossil energy. It offers a new technical path for creating a new energy system of “adjustable load-energy storage-energy supply” in data centers, supporting digital transformation of China′s economy and the steady advancement of its “dual carbon” goals.

[Objective] During verifying settings of relay protection equipment in substations, traditional methods mainly rely on manual verification or simple program verification. The manual verification accuracy varies, with relatively low verification efficiency. Simple program verification improves its efficiency to some extent, but there is room for further accuracy enhancement. To this end, a setting verification method for relay protection of intelligent substations based on deep learning was proposed. [Methods] Firstly, an improved convolutional recurrent neural network (CRNN) was employed to identify relay protection settings. Specifically, the convolutional neural network (CNN) was adopted to convert text images into feature sequences, followed by leveraging the recurrent neural network (RNN) to identify the feature sequences. Finally, the identification results were transcribed by adopting a dictionary-based connectionist temporal classification (CTC) loss function to obtain the setting text information. On this basis, the RNN module was enhanced by utilizing a convert gate unit, thus building a bidirectional convert gate long short-term memory (Bi-CGLSTM) model to achieve adaptive adjustment of data weights. Then, the setting verification was carried out by combining Chinese word segmentation technology. A complete dictionary of setting names was constructed, with the Levenshtein distance algorithm adopted to calculate the similarity between the text to be verified and the standard text. Additionally, an improved forward maximum matching algorithm was applied to match the setting text, thus completing the one-by-one setting verification of relay protection equipment in substations. [Results] 240 relay protection setting sheets from a power supply company that cover ten common equipment models were selected as experimental samples to validate the feasibility and effectiveness of the proposed method. The training parameter setting of the deep learning model was as follows:the iteration count of 100, learning rate of 0.001, and the Adam optimizer for adjusting weights and biases. The experimental results show that the identification accuracy of the improved CRNN model exceeds 97％, while the verification accuracy of the proposed method reaches 97.07％, with relatively shorter verification time and better overall performance than that of other comparative methods. [Conclusions] The identification accuracy of the setting text of substation relay protection in the context of big data can be effectively enhanced by the improved deep neural network. Additionally, verification accuracy can be ensured and verification efficiency can be significantly enhanced by the combination of the Levenshtein distance algorithm and the improved forward maximum matching algorithm. Powerful technical support is provided by the proposed method for the intelligent operation and maintenance of intelligent substations.

[Objective] With the high proportion integration of renewable energy sources such as wind and solar power into the grid, their inherent intermittency and volatility pose significant challenges to the voltage stability at the distribution end of the power grid. In particular, at the end of regional power grids, the uncertainty in wind-solar-load output increases the risk of rapid voltage drops, which may lead to equipment damage or even cascading failures. Existing studies have notable shortcomings in areas such as handling prediction errors in wind-solar-load output and multi-objective collaborative optimization. For example, the full-pure embedding sensitivity analysis method fails to adequately consider the influence of prediction errors, while the source-grid-load coordinated control framework ignores the interference of prediction errors on collaborative outcomes. To address these issues, this paper proposed a novel fast regulation algorithm for low voltage. By quantifying the uncertainty in wind-solar-load output, a multi-objective optimization model was developed that balances safety, performance, and cost. The model aims to achieve rapid and stable voltage regulation at the distribution end of the grid, thereby improving the reliability and adaptability of high-proportion renewable energy integration into the power grid. [Methods] The Collaborative Genetic Algorithm (CGA) was used as the core solution method. Firstly, precise probability density function models were established to account for the randomness in the output of wind power, photovoltaics, and load output. The output of wind power was quantified by combining the Weibull distribution of wind speed with the normal distribution of prediction errors. Photovoltaic output was associated with light intensity and photoelectric conversion efficiency, incorporating prediction errors. Load output was represented by a probability density function reflecting its volatility. Based on this, a low-voltage regulation model was developed with optimization objectives balancing safety, performance, and cost. The safety indicator quantified the total power loss at the distribution end of the grid, the performance indicator included the overall network loss and voltage deviation, while the cost indicator calculated the total lifecycle cost. Through integer-based mixed coding schemes and dynamically adjusted crossover and mutation probabilities, the algorithm effectively optimized the population and output the optimal solution that satisfied voltage stability margin requirements. [Results] Based on actual grid data from a region in Guangzhou, simulation experiments validate the effectiveness of the proposed algorithm. In terms of uncertainty handling, the proposed algorithm shows a significantly higher correlation between wind and photovoltaic output predictions and actual data compared to other traditional methods. This is due to the algorithm modeling output power prediction errors as random variables, which more accurately reflects the uncertainties in real-world systems. Regarding voltage regulation, when fluctuations in wind-solar-load output and increased load lead to voltage drops, the algorithm quickly and effectively restores node voltages to normal levels. Its performance outperforms traditional methods, such as those based on steady-state grid models and the double-loop voltage-current control algorithm. In terms of static voltage stability margin, the proposed algorithm maintains a high voltage stability margin of over 0.8 across various test scenarios, demonstrating strong voltage regulation capability. Furthermore, while ensuring voltage stability, the algorithm also considers the economic and performance efficiency of grid operations. [Conclusions] The fast regulation algorithm for low voltage effectively addresses the issue of low-voltage instability at the distribution end of power grids with high integration of renewable energy by deeply combining the wind-solar-load output uncertainty modeling and multi-objective optimization. This algorithm innovatively introduces probability density functions to quantify prediction errors, which significantly improves the accuracy of wind-solar-load output predictions. By using CGA for coordinated optimization of safety, performance, and cost targets, the algorithm achieves rapid dynamic voltage regulation. Experimental results show that the proposed algorithm outperforms traditional methods in terms of regulation speed, stability margin, and economic efficiency, which provides reliable technical support for intelligent grid control with high proportions of renewable energy integration. The research results not only have significant theoretical value but also demonstrate great potential in practical engineering applications. Future exploration will focus on voltage coordination control strategies across multiple time scales to continuously enhance the stability and economic efficiency of grid operations.

[Objective] In the operation and management of power system, the medium and low-voltage distribution network serves as a key link between power sources and users. Its operating efficiency and stability are directly related to the safety and reliability of the entire power system. Three-phase line loss, as an important indicator of the distribution network′s operational efficiency, not only reflects the energy loss during the power transmission process but also directly affects the voltage quality, power consumption, and safe operation of the power grid. However, the three-phase line loss data in the distribution network exhibit complex distribution characteristics, such as multi-modal and asymmetric features. During dynamic changes, it is difficult to accurately capture the inherent patterns and structures in the data, which reduces the accuracy of anomaly detection. Therefore, this paper proposed an intelligent method for the anomaly detection of three-phase line loss in medium and low-voltage distribution networks. [Methods] During the data collection process of the distribution network′s three-phase line loss, the data can be influenced by multiple factors such as electromagnetic interference and equipment errors, leading to the presence of significant noise and outliers. These noises not only reduce the signal-to-noise ratio but also obscure the true features of the data, thereby affecting the accuracy of subsequent analysis. Therefore, a radial basis function (RBF) neural network was used to extract features from the collected three-phase line loss data. By performing nonlinear mapping of the input data, the method effectively suppressed the interference from noise, enhancing the signal-to-noise ratio. The preprocessed data were then normalized, which further improved the completeness and accuracy of the data collection. A loop current-based method was employed to decompose the circuits in the distribution network into multiple independent loops. In each loop, the real and imaginary parts of the voltage and current were calculated. By analyzing the temporal and phase variations of these values in detail, the operating status of the circuit was thoroughly understood, and potential anomaly patterns were accurately identified. Based on the real and imaginary part values of the voltage and current on the three-phase branch circuits, a Gaussian mixture distribution model was constructed. This model used multiple Gaussian distributions to describe the complex distribution features of the three-phase line loss data, allowing for more accurate capture of the inherent patterns and structures in the data. The maximum expectation algorithm was then used to fit the normalized line loss rate and construct a hybrid Gaussian model consisting of multiple Gaussian mixture distributions. The likelihood probability function of the eigenvector was calculated, and based on a preset probability threshold, the data were determined to be anomalous or normal. If the likelihood probability was below the threshold, it was classified as anomalous; otherwise, it was considered normal. This approach enabled the identification of line loss anomaly data. [Results] Experimental results show that the proposed method can accurately identify three-phase line loss buses, reducing the risk of misjudgment and missed detections. [Conclusions] This method can promptly detect and address faults in the distribution network, which is of significant importance in improving the operational efficiency and reliability of power systems.

[Objective] To assess the security of unmanned aerial vehicle (UAV) communication environments in power applications, an effective security architecture and design scheme was proposed to address electromagnetic interference, data security, and other challenges UAV communications face during power inspections, ensuring efficient collaborative operations and secure data transmission of UAVs in complex electromagnetic environments. [Methods] Firstly, the structure of UAV collaborative wireless networks and security threats they face were analyzed, especially the influence of strong electromagnetic interference near power lines on UAV communications. Subsequently, an identity cryptography-based solution to UAV collaborative wireless networks was proposed. Additionally, by designing enhanced communication protocols and anti-interference mechanisms, stable transmission of critical data under strong electromagnetic interference was ensured. Meanwhile, the mutual authentication, signature, and identity verification mechanisms for communication data were introduced to enhance the overall security of UAV communications. [Results] Experimental results demonstrate that the proposed security assessment architecture and design scheme exhibits high data recovery rates and low resource consumption under varying numbers of UAVs, communication failure rates, electromagnetic interference intensities, and data packet sizes. In particular, the system maintains high data recovery rates and fault tolerance capabilities even under high electromagnetic interference, effectively resisting potential network intrusions and data tampering threats. [Conclusions] The proposed security assessment architecture and design scheme significantly enhances the security and reliability of UAV communications during power inspections, reducing the information leakage risk and enabling efficient collaborative operations in complex electromagnetic environments. The innovation of this study lies in combining identity cryptography and public key mechanisms to design a lightweight, efficient, and safe solution, providing effective security guarantees for UAV communication networks in power inspection tasks.

[Objective] High-voltage circuit breakers are key control and protection devices in the power system, and their reliable operation is crucial for the safety and stability of power grids. However, during long-term operation, high-voltage circuit breakers may trigger various faults due to mechanical wear, component aging, and other problems. Currently, the detection of high-voltage circuit breakers faces challenges such as diverse detection signals, great difficulty in fault detection, and low accuracy. Therefore, studying an efficient and accurate mechanical status detection method for high-voltage circuit breakers is of great significance for ensuring the safe and stable operation of the power system. [Methods] This study proposed a mechanical status detection model for high-voltage circuit breakers based on multi-modal efficient Transformer. In the data acquisition stage, vibration sensors, current sensors, and displacement sensors were comprehensively employed to synchronously acquire vibration signals, current signals, and displacement signals during the operation of high-voltage circuit breakers, thus constructing a multi-modal signal dataset. In the signal preprocessing stage, wavelet transform technology was adopted to process the acquired multi-modal signals and decompose the signals into different frequency scales, thus effectively removing noise components in the signals, enhancing fault feature signals, and significantly improving signal quality. In terms of model building, an efficient Transformer module was introduced. With its powerful self-attention mechanism, the module could effectively capture long-distance dependency relationships in signal sequences and dig deeply into complex features in multi-modal signals. Additionally, by classifying the operation status of high-voltage circuit breakers into six categories, including normal operation, failure to maintain closing, loose soft connection, single-phase contact wear, loose insulating tie rod, and opening spring fracture, accurate diagnosis of the mechanical status of circuit breakers was realized. [Results] In the simulation experiments, simulation models of different fault types of high-voltage circuit breakers were built to simulate various working conditions during actual operation and generate multi-modal signal data. Inputting the data into the proposed detection model for testing shows that the model can accurately identify different fault types. In the actual experiments, multiple high-voltage circuit breakers were selected as test objects, and multi-modal signal data were collected under their normal operation and different fault settings. The experimental results reveal that the proposed method significantly improves the detection accuracy compared with traditional detection methods while ensuring the detection speed. [Conclusions] The proposed mechanical status detection model for high-voltage circuit breakers based on multi-modal efficient Transformer effectively solves the problems of complex detection signals and severe noise interference. By leveraging the powerful feature extraction and classification capabilities of the efficient Transformer model, accurate identification of multiple mechanical faults in high-voltage circuit breakers is realized. Simulation analysis and experimental results fully demonstrate that this method performs well in both detection accuracy and speed, providing reliable technical support for the status monitoring and fault diagnosis of high-voltage circuit breakers in the power system. It helps to timely detect potential faults of the equipment, and holds application significance and broad promotion prospects for ensuring the safe and stable operation of the power system.

[Objective] With the increasing importance of power transmission and transformation projects in distribution networks, traditional cost estimation methods face challenges such as large errors and excessive time consumption, making them inadequate for modern engineering management. To enhance the execution efficiency and estimation accuracy of power transmission and transformation projects, this study proposed a novel cost estimation method based on radial basis function neural network (RBFNN) and significance cost theory. This approach aims to address the limitations of traditional methods in complex cost estimation scenarios while enhancing the robustness and adaptability of the model. [Methods] This study applied significance cost theory to screen historical project data and identify the main factors affecting cost estimation for power transmission and transformation projects. These factors were used as input features for the neural network. The method introduced radial basis functions (RBFs) to restructure the traditional artificial neural network (ANN) architecture, creating a cost estimation model specifically for power transmission and substation projects. The model processed input data using Gaussian functions, initialized the hidden layer centers with the K-means clustering algorithm, and used least squares and gradient descent methods to train the output and hidden layers. To validate the model′s effectiveness, 100 sets of data from power transmission and transformation projects were used to compare the cumulative absolute error rate and average execution time of traditional methods (unit cost method and index estimation method) with the proposed model. Moreover, SHAP value analysis was employed to quantify the impact of key factors on estimation error rates. [Results] Simulation results demonstrate that the RBFNN-based cost estimation method outperforms traditional methods in both cumulative absolute error rate and execution time. When the test sample size increases to 20, the cumulative error rate for the unit cost method reaches 440％, while the index estimation method reaches 180％, and the proposed model maintains an error rate below 110％. In terms of execution time, traditional methods require an average of 5 s, while the proposed model reduces the time to just 0.5 s. In addition, SHAP value analysis reveals that factors such as wire cross-sectional area, steel pipe poles, and the number of circuits have the greatest influence on estimation error rates, with their SHAP values significantly higher than those of other factors. This finding provides critical insights for model optimization and cost control. [Conclusions] The cost estimation method proposed in this paper, based on RBFNN and significance cost theory, effectively improves the accuracy and efficiency of cost estimation for power transmission and transformation projects. Although the method still exhibits some errors in complex construction environments, it outperforms traditional methods in overall performance, making it highly practical with strong potential for widespread application. Future research will focus on integrating regression analysis, support vector machine (SVM), and other machine learning algorithms to further optimize model precision and better handle the complexity and variability of power transmission and transformation projects.

[Objective] With the increasing proportion of new energy, traditional grid-following (GFL) control based on phase-locked loop (PLL) synchronization gradually exhibits inherent stability limitations in weak grid conditions. Meanwhile, grid-forming (GFM) control with self-synchronizing source characteristics has emerged as a hot solution. However, existing research predominantly focuses on the voltage regulation or synchronization stability of GFM control, with less attention to its frequency modulation capability and characteristics. [Methods] This paper systematically reviewed four mainstream GFM control methods, including droop control, virtual synchronous generator (VSG) control, matching control, and virtual oscillator control (VOC), explained their frequency modulation principles, and analyzed their advantages and disadvantages from the aspects of the control loop and application scenarios. On this basis, a grid-connected simulation model for new energy systems was built to conduct a simulation-based analysis of the frequency modulation response characteristics of different kinds of frequency modulation control across diverse scenarios. Finally, this study summarized challenges of GFM control in strategy optimization, parameter tuning, and multi-unit coordination, with the future development prospects pointed out. [Results] Droop control regulates the active power of generating units by responding to system frequency deviations, featuring advantages of the simple structure and strong grid strength adaptability. However, its lack of inertia support results in relatively weaker frequency modulation performance. On the basis of droop control, VSG control simulates the inertia response characteristics of conventional synchronous machines and can better suppress the change performance of system frequency. However, it faces challenges in parameter tuning, fault ride-through, and multi-unit coordination. Matching control utilizes the dynamic characteristics of DC capacitors to simulate the inertia properties of traditional synchronous machines and thus restrain change performance of system frequency, but it fails to provide sustained support in the frequency quasi-steady state. VOC generates frequency responses similar to droop control via oscillator dynamic equations that directly govern amplitude and frequency. However, it is difficult for its high output harmonics to satisfy grid connection requirements. [Conclusion] Virtual synchronous machine control has become the most promising research direction in GFM control due to its technical advantages of balancing frequency modulation performance and strong grid strength adaptability in participating in system frequency modulation. However, technical challenges including synchronization stability, fault ride-through, and coordinated control need to be tackled. In the future, in-depth research should be conducted on control strategies and parameter optimization, and multi-unit collaborated control to facilitate the large-scale application of GFM control.

[Objective] The cost of distribution network engineering is influenced by multidimensional factors such as scale and capacity, equipment and material costs, and geographical conditions. Traditional statistical methods (e.g., linear regression) struggle to handle high-dimensional nonlinear data effectively, while existing machine learning approaches, despite incorporating feature reduction techniques, still exhibit limitations. For instance, principal component analysis (PCA) sacrifices prediction accuracy for dimensionality reduction, and grey relational analysis (GRA) ignores feature interactions. Therefore, there is an urgent need for a prediction method that retains critical feature information while accounting for complex inter-feature relationships. This study integrated recursive feature elimination (RFE) with the random forest (RF) algorithm to develop a RFE-RF prediction model, aiming to resolve feature redundancy and nonlinear modeling challenges. [Methods] A technical framework of “feature selection-model construction-experimental validation” was adopted. For feature selection, the recursive feature elimination (RFE) method was employed, which iterated training models to gradually eliminate features with minimal predictive contributions, retaining an optimal feature subset. For model construction, the RF algorithm was utilized. Based on ensemble learning principles, RF constructed multiple decision trees and averaged their outputs, effectively mitigating overfitting and enhancing model robustness. RF was insensitive to noisy data and quantified feature importance, providing reliable feature ranking criteria for RFE. By embedding RFE into the RF training process, a closed-loop optimization workflow was established. [Results] Experimental validation used data from 190 distribution network engineering projects provided by a power grid company, covering 21 initial features such as voltage level, line length, and equipment costs. Categorical features were numerically encoded while preserving their original distribution characteristics. Through five-fold cross-validation and root mean square error (RMSE) optimization, the optimal feature subset was identified as 12 optimal feature subsets, including such key factors as line length, comprehensive cable price, and voltage level. Compared with traditional linear regression (LR), RF, and mutual information-based RF (MI-RF) algorithms, the RFE-RF algorithm achieves a mean absolute error (MAE) of 8.6579 and a mean absolute percentage error (MAPE) of 6.97％ on the test set, significantly outperforming other algorithms. The MAE of RFE-RF on the test set increases by only about 4.5％ compared to the training set, indicating lower overfitting risks and demonstrating that feature selection effectively enhances model stability. [Conclusion] Feature selection is pivotal for improving the accuracy of distribution network cost prediction. RFE dynamically eliminates redundant features through iterative processes, substantially reducing data dimensionality and noise interference. The RFE-RF model combines high precision with strong interpretability, reduces MAE significantly compared to traditional models, and clearly quantifies the impact weights of individual features on costs. This study marks the application of combining RFE and RF in cost prediction for distribution network engineering, addressing challenges in feature interaction and redundancy filtering and providing a new paradigm for data modeling in complex engineering systems. The model serves as a precise cost prediction tool for power grid enterprises, aiding investment decisions and cost control, thus advancing intelligent and refined construction of distribution networks. Moreover, it reveals the impact mechanism of feature selection on the generalization capability of machine learning models, offering practical references for feature optimization in high-dimensional nonlinear datasets.

[Objective] With the global energy transition and the rapid development of clean energy, the penetration rate of high-penetration photovoltaic (PV) sources in distribution networks is increasing. However, PV output power exhibits significant fluctuations and uncertainty due to factors such as solar irradiance and temperature. When a large number of such sources are integrated into distribution networks, they can cause voltage fluctuations, frequency variations, and other issues, presenting significant challenges for power outage fault prediction. Traditional fault prediction methods struggle to accurately capture fault characteristics in complex distribution networks with high PV penetration, leading to reduced prediction accuracy and efficiency, which fails to meet the stability requirements for distribution network operation. [Methods] To improve prediction accuracy and efficiency, this study proposed a fault prediction method for distribution networks with high PV penetration. First, a PV-integrated grid model was built to analyze the impact of PV sources on fault current characteristics in distribution networks. This model clarified how PV sources influence fault current magnitude and distribution under different operating conditions, providing a theoretical basis for subsequent fault zone identification. Next, potential outage zones were inferred by combining grid topology and load imbalance features. The grid topology reflected the connectivity of components, while load imbalance indicated regional load variations. By integrating these factors, the method more accurately localized the fault zone. In addition, power flow entropy was introduced to assess whether circuit loads were in a critical state. Key fault-related power flow features were then extracted from the identified zones. These features were fed into an optimized SA-SAE for training, allowing the system to automatically learn underlying patterns from large datasets and achieve precise outage prediction. [Results] Experimental results demonstrate that the proposed method achieves high prediction accuracy in fault localization for distribution networks with high PV penetration, correctly identifying fault zones (sections 3-6 of the K5-K8 lines) and fault types. Moreover, the average prediction time is only 2.236 seconds, significantly outperforming comparative methods in both accuracy and efficiency. [Conclusion] By comprehensively considering PV integration effects, grid topology, load characteristics, and leveraging power flow entropy and SA-SAE, the proposed method enables high-precision and high-efficiency outage prediction in distribution networks. This method not only enhances prediction accuracy and timeliness, reducing outage risks and economic losses, but also provides robust support for grid planning, operation, and maintenance. It ensures stable distribution network operation and facilitates large-scale integration of clean energy.

[Objective] With the increasing demand for application of unmanned aerial vehicles (UAVs) in complex scenarios such as power grid inspection and emergency rescue, the limitations of single UAV in task execution have become increasingly prominent. Multi-UAV formation can effectively improve inspection efficiency and expand operation coverage, but significant challenges remain in formation maintenance, collaborative trajectory optimization, and environmental adaptability to complex environments during practical application. An optimal control method that integrated virtual spring forces with the hp-adaptive pseudospectral method was proposed to address the difficulties of formation maintenance and path optimization during large planar maneuvers of UAV swarms, thus enhancing the stability, flexibility, and disturbance resistance of collaborative flight of UAV formations and providing technical support for high-demand scenarios for UAVs such as power grid inspection. [Methods] First, a multi-UAV system dynamics model was built, and a virtual spring mechanism was incorporated into the formation control system to realize flexible constraints and elastic self-adjustment between UAVs. By combining the virtual spring method with the traditional leader-follower method, a formation strategy that could balance rigid support and adaptive adjustment ability of formations was designed. On this basis, the hp-adaptive pseudospectral method was then applied to solve the optimal control problem of UAV formations. By discretizing state and control variables at Legendre-Gauss nodes and constructing global interpolation polynomials, the trajectory optimization problem was transformed into a nonlinear programming (NLP) problem, with constraints such as dynamics, energy consumption, and velocity combined to conduct a high-precision numerical solution. In simulation experiments, a typical four-UAV diamond formation was set up, and the algorithm′s adaptability to different terrains, wind disturbances, and mission requirements was comprehensively explored. [Results] Simulation results show that the proposed virtual spring-based hp-adaptive pseudospectral method can realize smooth formation turning and velocity control. During a 90° large maneuver, UAVs can not only satisfy multiple constraints such as path deflection and speed change, but also maintain a stable formation. Compared with traditional leader-follower and artificial potential field methods, the new method demonstrates significant advantages in position error, formation maintenance, and wind resistance. Under 10m/s strong wind, the formation stability of the proposed method exceeds 70％, showing significant advantages over its competing algorithms. 3D terrain simulations and real flight tests further validate the algorithm′s adaptability and robustness, and the method still maintains lower formation deformation rates and trajectory tracking error under multiple terrains such as hills, mountains, and canyons, with the features of reasonable energy consumption control and strong engineering practicability. [Conclusion] By innovatively integrating the virtual spring elastic constraint with the hp-adaptive pseudospectral method, an optimal control technique for UAV formation trajectory planning in complex environments was proposed. The rigidity constraint limitations of traditional formation methods are overcome, flexible maintenance and adaptive adjustment of formations are realized, and the accuracy and efficiency of collaborative trajectory optimization are significantly improved by the method. The research results provide an efficient and reliable technical path for collaborative flight of UAV swarms in demanding tasks such as power grid inspection and emergency rescue. Future studies may further increase the method′s application potential in multi-formation collaboration and complex obstacle environments, promoting the intelligent and practical development of UAV formations.