Abstract Heatwaves are becoming more frequent under global warming, leading to concurrent load surges and intensified fluctuations in light-industrial electricity consumption. This heatwave-induced nonstationarity weakens the accuracy of conventional short-term load forecasting approaches. To capture both heatwave-driven interuser coordination and dynamic spatiotemporal dependencies, we propose a heatwave-gated adaptive spatiotemporal graph neural network (HWG-STGNN). Specifically, gated temporal convolutional network blocks with causal dilated convolutions model long-range temporal dependencies; a learnable adjacency matrix adaptively characterizes spatial correlations among multiple users; and a heatwave gating module adjusts feature contributions according to meteorological intensity, enabling the model to respond to varying heat-stress levels. We evaluate the proposed method using 15-min multiuser load data from a southern Chinese city (2018–2019). Results show that HWG-STGNN achieves competitive overall performance across different forecasting settings, with more pronounced and consistent gains during heatwave intervals. On heatwave samples, the mean absolute error and mean absolute percentage error are reduced by approximately 18% and 14%, respectively, with root mean square error also improved. These findings indicate that incorporating extreme meteorological information into spatiotemporal forecasting frameworks is an effective way to enhance accuracy and robustness in high-temperature, strongly nonstationary conditions.

Background: In the large-scale steel industry, significant power load variability, especially during processes like steel smelting, poses challenges to power system safety. Although there is an abundance of research and patents related to load forecasting, studies and patents specifically addressing large industrial load forecasting are sparse. Hence, accurate ultra-short-term load forecasting becomes particularly crucial. Objective: This study proposes an innovative method for ultra-short-term load forecasting to improve prediction accuracy during peak periods and mitigate risks in high-load conditions. Methods: We introduce an LSTM-XGBoost model enhanced by a random forest network and an improved grey wolf optimization algorithm (IGWO) for feature selection and parameter optimization, respectively. Results: Compared to other advanced models, our method demonstrates superior performance across key indicators such as MAPE (1.93%), RMSE (220.81), and R2 coefficient (0.99), and the prediction error is lower during both peak and off-peak periods. For instance, the proposed model achieved a MAPE improvement of over 25% compared to traditional models. Validation with data from multiple time periods confirms the model's accuracy and robustness. Conclusion: The proposed forecasting method effectively tackles load fluctuations in the steel industry, supporting safe and economical power system operations. Future research will aim to further improve peak identification accuracy and enable continuous adaptive learning.

Multi‐energy system (MES) is crucial for the development of smart cities. This paper summarises the research progress and achievements from an engineering case study in China which aims at enhancing MES energy efficiency. Key theories and technologies were tested at an MES demonstration site in Hunan, China, using both software and hardware systems which improve significantly in energy efficiency, carbon savings, environmental and economic benefits. Specifically, the optimal operation modes proposed in this paper achieves a best‐case scenario of 9.97 tons of CO 2 savings, 26.4% lower than conventional systems. The findings provide crucial guidance for policymakers and industry stakeholders which aims to implement sustainable energy solutions in future urban planning.

Operating under shading conditions is one of the most significant negative phenomena that photovoltaic (PV) arrays face, severely affecting power generation. Shading arrays can produce multiple local maximum power points (LMPP) and a single global maximum power point (GMPP). Therefore, it is crucial to reconfigure the shading modules within the array to extract the GMPP. This highlights the importance of studying the uniform dispersion of shading across the PV array surface. This paper proposes an optimized configuration method for shading PV arrays based on the multi-objective pied kingfisher optimizer (MOPKO). The primary objective of the proposed MOPKO is to provide the optimal structure of the switching matrix to minimize the row current difference and column swapping within the PV array. The advantage of this strategy lies in its ability to perform a more realistic dynamic reconfiguration process. The method is validated on a 9 x 9 PV array with ten shading cases. Additionally, the results of the MOPKO scheme are compared with TCT, INGO, and Triple X Sudoku based on evaluation metrics such as fill factor (FF), mismatch loss (ML), and performance ratio (PR). Results show that reconfiguration of PV arrays with MOPKO always obtains the highest PR under ten different shading conditions. PR has considerably raised with 13.4 %, 3.5 %, 20.7 %, 15.1 %, 12.2 %, 10.2 %, 13.5 %, 6.5 %, 3.2 %, and 2.2 % compared to TCT. The results of the analysis verify the advantages of the proposed MOPKO in solving the problem of multiple peaks in the P-V characteristic curve and in achieving high power levels.

Accurate wind turbine power prediction (WPP) is especially important for grid dispatching and grid integration. However, the high randomness and intermittency of wind resources make wind turbine (WT) have complex spatial–temporal dynamics and hierarchical characteristics. Most existing studies rely on combinatorial methods, which cannot accurately predict WT power. To realize accurate WPP, this paper proposes a novel model called multi-scale spatial–temporal interaction network (MSTINet). The model constructs a hierarchical spatial–temporal feature extraction framework to process subsequences with different scales to fully extract the hierarchical characteristics of WT. In each layer, MSTINet utilizes the interactive learning strategy that allows each subsequence to have both local and global views. The interactive learning strategy, temporal and spatial feature extraction module are combined to fully extract local and global information and complex spatial–temporal dynamics. Subsequently, we use the slight convolutional block attention module to highlight dynamic changes in the sequence and capture deeper spatial–temporal dynamics. The experiment results show that MSTINet achieves a maximum R2 of 0.9799, a 15.74

Background: The electricity demand is continuously increasing. However, various institutions, enterprises, and individuals exhibit many irregularities in their electricity usage, leading to significant wastage of electricity. To achieve effective energy management, researchers are attempting to analyze and regulate users' electricity demands by monitoring their load usage through Non- Intrusive Load Monitoring (NILM) technology. The accuracy of load identification in this technology will greatly impact the results of load monitoring. Although there are currently many articles and patents related to NILM, they utilize a large amount of computational resources and require high sampling rates from devices, yet the results are still unsatisfactory. Therefore, it is necessary to improve the accuracy of load identification in data with relatively low sampling frequencies. Objective: To improve the accuracy of load identification with low sampling frequency data, this paper proposes a typical scenario load identification method based on feature fusion and transfer learning. Methods: This method adopts the fusion of current and power factor angles to provide abundant identification information for NILM, effectively reducing the situation of single-feature overlap of different loads. By inputting the fused feature data into GoogLeNet and utilizing transfer learning for training, not only is the accuracy improved, but also the training time and the requirement for the sampling rate of training data are greatly reduced. In addition, selecting typical scenario loads can monitor loads in a targeted manner, reduce the waste of computing resources caused by irrelevant loads, and more effectively guide electricity usage strategies. Results: The proposed load identification method was tested on the low sampling frequency dataset used in this paper. It achieved an overall load identification accuracy of 94.61% across three scenarios, improving accuracy by 3% to 7% compared to other models. Conclusion: The simulation results indicate that this method achieves high load identification accuracy at low sampling frequencies. It also exhibits good generalization ability. This method not only reduces the performance requirements for monitoring equipment but also enhances monitoring efficiency.

The modeling of dynamics in energy devices and pipeline networks reflects the real states of multi-energy flows, which is significant for realizing accurate optimal dispatch of integrated energy system (IES). In this paper, an approach with data-driven dynamic energy hubs (DDEH) and thermal dynamics of pipeline networks (TDPN) is proposed to describe the energy dynamic response process of IES. Based on the efficiency characteristics of energy conversion and storage devices, data-driven deep neural network (DNN) is adopted to excavate input-output relationship of the variable efficiency devices, which helps establish DDEH in time domain. To address the problem of nonlinear introduced by DNN, the nonlinear activation function in DDEH is equivalently converted into mixed integer linear model. At the same time, the TDPN is developed by bilateral characteristic line method (BCLM), which quantifies the time delay and loss of pipeline networks. TDNP demonstrates the network transportation dynamics in optimal dispatch, and the virtual energy storage effect of pipeline networks are analyzed. Case study from a community IES verifies the proposed approach effectively improve dynamic modeling accuracy of devices and pipeline networks, and have superiority in providing precise, reasonable and highly efficient optimal dispatch scheme.

Photovoltaic (PV) power has become a crucial solution to the escalating energy crisis. Among the various implementations, Rooftop PV power generation systems (RPVPGS) are predominant in PV buildings. However, RPVPGS will face challenges such as reduced output power due to array fault or shading, leading to fluctuations in Building-Integrated PV (BIPV) power generation. This paper attempts to solve this problem by proposing a novel multivariate reconfiguration method based on the improved northern goshawk optimization algorithm (INGO). The aim is to find the optimal state of RPVPGS under various conditions. In this paper, extensive simulations were conducted on the experimental platform to assess the feasibility and effectiveness of the proposed method. It is worth noting that INGO outperforms existing technologies such as Arrow SoDuku and Zig-zag for the evaluation metrics mentioned in the article. Furthermore, rigorous simulation experiments were conducted on the semi-physical platform to validate the proposed approach. The power enhancement percentage deviation was between +0.1% to +0.2%. These results unequivocally demonstrate that the INGO-based multivariate reconfiguration method accurately reconfigures RPVPGS, ensuring the efficiency and stability of BIPV systems.

Partial shading of solar photovoltaic (PV) panels can significantly affect the performance of solar PV arrays. Various reconfiguration techniques have been explored in recent years. Still, their applicability to actual PV power generation is controversial due to the number of electrical switches, physical locations, interconnections and complexity. This study proposes an adaptive two-step staircase (A2SS) static reconfiguration method. The technique is experimentally validated in several conditions and compared with the conventional TCT connection, single-step staircase (1SS) static reconfiguration method, Arrow soduku, modified odd-even-prime (MOEP) and two-step staircase(2SS) static reconfiguration method. For the eight shading cases of LN, LW, LD, Ran, Cen, Cor, CD, and Plus at SET#1, after reconfiguring the PV array using A2SS, the power has a significant improvement of 17.6%, 17.0%, 13.4%, 13.4%, 20.6%, 20.2%, 3.1%, and 0.82% than TCT. In the four shading cases of Lr. C, Lr. O, Lr. T, and Lr. U at SET#2, the power showed a significant improvement of 11.8%, 9.2%, 10.7%, and 15.8% compared to TCT. It also has the best performance in various reconfiguration techniques, which are mentioned. In addition, the A2SS reconfiguration method can be better applied to various sizes of PV arrays. By optimizing the shading distribution and adjusting the row irradiance deviation, the power stability of PV power generation is improved while maximizing energy efficiency.

To ensure the timely detection of safety hazards in overhead transmission lines with railroad conductors and improve the accuracy of night insulator defect detection, this paper proposes the DPYOLOv5 algorithm with dark and light channel enhancement optimization. It improves the night insulator image quality by introducing the dark and light channel enhancement algorithm, builds a lightweight network by combining the DP-BS module, and adds the Shuffle Attention module to enhance the feature extraction and ensure detection accuracy. At the same time, the EC-Loss loss function is used to optimize the prediction frame adjustment, accelerate the model convergence, and improve detection efficiency and accuracy. The simulation results show that the insulator dataset processed by DP-YOLOv5 has an accuracy of 95.3%, a recall of 94.8%, an average accuracy of 95.5%, and FLOPs of 219.3. Compared with YOLOv5, the mapped value is improved by 0.9%, the F1 is improved by 1%, and the model parameter and FLOPs are reduced by 48.8% and 50.8%, respectively.

Smart community (SC) serves as a vital hub for diverse human activities, and the integrated energy technology plays an important role in maintaining its normal operation and realizing advanced functions. Integrated energy technology consists of multiple energy technologies throughout the whole life cycle and the complete energy transformation process. Although there are some literature reviews on energy technologies and SC, in most of the existing works, these two fields are studied separately, let alone works on the technical and practical application of integrated energy technology in SC. To fill the gap, a compact survey on the integrated energy technology in SC is given in this paper. The main energy issues and corresponding solutions in SC are introduced, including addressing uncertainties, reducing energy loss, taking advantages of interactions among multiple energy entities, dealing with various faults, and evaluating the operating performance of SC. Then, a number of worldwide SC projects are compared regarding energy technology, with a typical demonstration project in China elaborated in detail. Finally, the future research directions of integrated energy technology in SC are discussed. This survey aims at giving reference for future investigation of integrated energy technology in SC to promote smart human livelihood and reduce carbon emissions.

The power transformer is the most important and critical component in the power grid of the electrical system. Its safe and stable operation is of great significance for the reliable transmission of renewable energy generation and the reliable power supply to end users. With the continuous development of new types of power systems, the load of the power system undergoes drastic changes, resulting in increased volatility and instability, which leads to issues such as overload, harmonics, and short circuits in transformers. Therefore, in order to accurately assess the operating status of transformers and promptly identify any existing conditions, a method is proposed in this paper. This method uses an optimal cloud entropy parameter calculation method and a variable weighting method to optimize the gray-cloud evidence model. Through a novel multi-source information fusion method, the results of various test items are fused to obtain the state awareness results. Meanwhile, different evaluation indicators are selected for different levels of renewable energy penetration and compared with other methods. The results are validated through examples, demonstrating that the method proposed in this paper can accurately reflect the operating status of transformers and has good scalability.

To obtain the operating status of equipment and loads in buildings without installation of intrusive monitoring devices, non-intrusive load decomposition and monitoring methods have become the research focus of many scholars and researchers. To improve the accuracy of non-intrusive load decomposition, a non-intrusive load decomposition based on parallel connection network and attention mechanism is proposed. This proposed method educes the depth of network by ‘parallel connection’ and reduces the risk of overfitting. In additional, the dilated residual convolutional neural network and the bidirectional long short-term memory network are connected in parallel to extract features respectively, which greatly improves the representation ability of features. The attention mechanism is introduced to eliminate redundant information, focus on important information, and improve the decomposition performance. Finally, the domestic self-assessment data set is used, and different evaluation indicators are used for evaluation and comparison with other commonly used models. The simulation results show that the decomposition accuracy of proposed method is significantly improved.

The individual cells of lithium-ion battery packs cause inconsistency in the battery packs due to production differences. The working environment further aggravates the inconsistency of the battery packs, which seriously affects the efficiency and service life. To improve the consistency of battery packs, the four-switch reconfigurable cells are proposed in this paper; Heuristic algorithms are proposed to determine the switching states to obtain a battery topology that satisfies the requirements. This minimises the mismatch between the target load voltage and the output voltage of the battery pack and maintains the State of Charge (SOC) balance between different cells. The core of the proposed reconfiguration strategy is to divide all normal battery cells into modules when output voltage requirements are met. The modules are connected in parallel within the modules and in series between the modules so that all the battery cells are kept charged and discharged at the same rate. The results show that after reconfigurable balance, the energy utilization efficiency of the battery pack reaches 86.08%. The deviation of the SOC of the battery pack is reduced by 62.15%. It indicates that the inconsistency of the battery pack has been greatly improved.

In high-energy and high-power applications, thousands of batteries are connected in series and parallel, imposing a substantial computational burden for state of charge (SOC) estimation. The second-order RC equivalent circuit model is often utilized for SOC estimation. However, this model requires the identification of numerous parameters, rendering the calculations complex and computationally intensive. Furthermore, the model often neglects the impact of temperature. To enhance the speed and accuracy of SOC estimation for numerous individual cells, an equivalent circuit model is constructed. This model incorporates temperature correlation coefficients and the electrical characteristics of lithium-ion batteries at various temperatures. Subsequently, a combined forgetting factor recursive least squares and extended Kalman filter algorithm is introduced for battery SOC estimation. The results demonstrate that the improved model significantly reduces SOC estimation time. Compared to the traditional second-order RC model, the improved model reduces the time by 37.8

Background:: The railroad catenary insulator, which is a crucial component of the catenary system and is situated between the pillar and wrist arm, is crucial for electrical conductor isolation, electrical equipment insulation, mechanical load bearing, anti-fouling, and anti-leakage. The catenary insulators will experience tarnished flash, breakage, insulation strength deterioration, and other issues as a result of the long-term outside unfavorable working circumstances. The train electrical system's ability to operate normally is greatly hampered by these problems. Although there are many patents and articles related to insulator fault detection, the precision is not high enough. Therefore, it is crucial to improve the precision of catenary insulator fault detection. Objective:: An improved region-based convolutional neural networks (Faster R-CNN)-based fault detection method for railway catenary insulators is proposed in response to the long detection time of the conventional railroad catenary insulator fault, the low precision of the catenary insulator fault detection for occlusion and truncation, the poor performance of multi-scale object detection, and the processing of class unbalance problem. Methods:: The Faster R-CNN is optimized from four perspectives: feature extraction, feature fusion, candidate box screening, and loss function, in accordance with the properties of the catenary insulator. First, to solve the problem of multi-scale catenary insulator fault detection, convolutional block attention module (CBAM) and feature pyramid network (FPN) are used to fuse the deep feature and shallow features of the image. This results in a feature map with more critical semantic information and higher resolution. After that, the weighted non-maximum suppression (WNMS) algorithm improved by distance-intersection over union (DIOU) and Gaussian weighting function is used instead of the traditional NMS algorithm, which effectively introduces the overlap of detection frames into the confidence level and makes full use of the effective information of the detection frames. Finally, the improved Focal loss is used as the classification loss, and the focusing parameter and the balance factor of the Focal Loss are adjusted dynamically to solve the problem of sample imbalance and difficult sample identification in the model better. Results:: The effects of SSD, YOLOV3, traditional Faster R-CNN and improved Faster R-CNN models are tested on the contact network insulator fault detection dataset constructed in this paper, and the experimental results show that the improved Faster R-CNN has higher precision, recall, and mAP compared to the other detection models, which reach 94.31%, 96.68% and 95.22%, respectively. Conclusion:: The results of the experiments demonstrate that this method may successfully detect the faults in different scale catenary insulators. It can effectively detect truncated, obscured faulty catenary insulators. It has higher precision and recall and provides a reliable reference for maintaining faulty insulators in railway catenary.

The oil-immersed power transformer plays a crucial role in ensuring the safe and stable operation of the entire power system, highlighting the significance of enhancing the precision of its fire detection Aiming at the problems of difficulty in status sensing and inaccurate sensing results of traditional oil-immersed transformer status sensing models, a fire status sensing model for oil-immersed power transformers based on TOPSIS (Technique for order preference by similarity to an ideal solution) method and combination weighting of game theory is proposed in this paper. The model combines subjective weighting with objective weighting, while using game theory to determine the subjective and objective weight coefficients, eliminating human influence, and the TOPSIS method determines the final status sensing results. The test data of oil-imm supplied by a power company serves as the verification and analysis sample. This paper focuses on the oil-immersed power transformer, a key equipment in the power system, and carries out research on fire state sensing, which promotes the progress of state sensing technology and is of great significance to ensure the safe and stable operation of the power system and the normal life of residents, and providing early warning of potential fires, which holds significant practical implications in engineering.

Partial shading can reduce the power output of a photovoltaic (PV) array due to mismatch losses. Therefore, various static and dynamic reconfiguration techniques have been proposed to address this problem. Although dynamic reconfiguration methods are fast and flexible, they face challenges such as complex hardware circuits and high costs. In contrast, static reconfiguration methods simplify control complexity, improve system stability, and significantly reduce costs. Among the existing static reconfiguration techniques, the competence square (CS) method, which changes the physical positions of PV panels without modifying the total cross tied (TCT) based electrical connections, has been introduced to enhance power generation. However, this arrangement faces drawbacks due to the lack of effective chromatic dispersion and insufficient power enhancement under shading conditions. This paper proposes an improved competence square (ICS) reconfiguration method to further improve the power output. The ICS method first reconfigures the PV array using the CS method, then applies the lo shu (LS) method for secondary reconfiguration to determine the optimal connections. This method is compared with TCT, CS and improved northern goshawk optimization (INGO) methods. Additionally, the performance of the proposed method is evaluated against various existing photovoltaic array configurations by comparing the global maximum power point (GMPP), fill factor (FF), mismatch loss (ML), and power enhancement (PE). The experiment shows that under shading conditions of square shading, rectangle shading, trapezoid shading, triangle shading, L-shaped shading, cross shading, discontinuous shading, and irregular shading, reconfiguring PV arrays using ICS consistently achieves the highest power improvement. Compared to TCT, the power output increases significantly by 16.6%, 2.1%, 18.0%, 16.0%, 19.1%, 5.0%, 4.0%, and 3.0%, respectively. Comparison results validate the proposed ICS method in enhancing the global maximum power under shaded conditions.

By the end of 2021, the urbanization rate of China had reached 64.72 %. Increasing urbanization rates have led to an 80 percent share of carbon emissions in urban areas. Therefore, a load optimal scheduling model of smart buildings is proposed in this paper, which takes into account the cost of carbon emissions. The model aims to minimize the electric cost for building residents and maximize the utilization of distributed energy. The research focuses on the load optimal scheduling for two types of residents, i.e., the working residents and the retired ones, based on their electricity usage patterns and habits. To address the challenges of uneven load distribution, nonlinear behavior, and uncertainties introduced by electric vehicles and photovoltaic energy, the multi-objective beluga whale optimization algorithm with hybrid reverse learning competitive strategies (RCMBWO) is introduced. It utilizes the energy storage and discharge capabilities of electric vehicles to mitigate the uncertainties of photovoltaic energy generation. The simulation results show that load optimal scheduling for both types of residents can lead to significant cost savings for a smart building with 100 households, approximately 29,554 RMB per year. Furthermore, the optimized results can guide the determination of the appropriate size for the photovoltaic energy storage system, reducing energy waste resulting from insufficient energy consumption capability of smart buildings. The research on targeted load optimal scheduling for classified residents presents a viable solution for enhancing the cost-effectiveness and environmental benefits of smart buildings.

Rail freight is a common mode of transportation for bulk commodities and is widely used around the world, with heavy quality and large inertia. Conventional control methods are influenced by the dynamic parameters of the train model. To solve this problem, this paper designs a mayfly algorithm-predictive active disturbance rejection controller (MA-PADRC) for speed tracking and wheel anti-slip control of electric freight trains. This control method reduces dependence on accurate modeling. MA-PADRC integrates the searching ability of MA, predicting ability of Smith estimator and high efficiency of ADRC, and it can accurately estimate the driving disturbance of the freight train. Meanwhile, to prevent wheel slip, an Elman neural network-Kalman filter (ENN-KF) is designed. Through it, the precise observation of anti-slip parameters and the constraint of anti-slip control are realized. The proposed control is validated by simulation of both real and virtual driving speeds. The results show that the proposed control has a strong tracking performance, and can effectively prevent wheel slip and guarantee the safe driving of the freight train with large inertia.