Volume 27, Issue 1

Cybersecurity is a critical component influencing the safe and stable operation of the smart grid, as the smart grid (SG) is an information-physical fusion system (also known as a Cyber-Physical System [CPS]) integrating sensing, communication, computation, decision-making, and control built on the foundation of the traditional grid. The most common cyberattack that compromises data integrity in the smart grid is the False Date Injection Attack (FDIA). If this type of attack is not discovered in time, it can take control of physical equipment or target a network transmission line to obstruct control decisions, which can result in power network failure or even a cascade failure in the grid. An integrated learning-based detection method is currently proposed to biclassify power grid data in order to address the issues of low accuracy, high false detection rate, and poor model differentiation ability when applying a single classifier in machine learning to detect false data. The integrated learning detection method is based on GBDT (Gradient Boosting Decision Tree), XGBoost, and Light GBM, with RF- Light GBM and Bagging classifier as the base classifiers, which are integrated by voting strategy after Bayesian tuning. Following simulation experiments, the algorithm is able to significantly outperform the checking rate and accuracy of traditional detection algorithms in detecting false data on the power grid by effectively addressing the issues of low checking rate and accuracy of single classifier detection as well as instability of single classifier detection.

With the continuous progress of social civilization, electricity plays an indispensable role in human life. China is a vast country with wide grid coverage and frequent natural disasters, which cause certain losses to the national economy. Therefore, this paper fully considers the role of meteorological information in the emergency decision-making of smart grid disaster prevention, uses the Dijkstra algorithm to clarify the best path, and integrates many factors such as traffic coefficient, road level, road congestion and road quality to establish the optimal path model, which can effectively improve the management level, reduce the smart grid loss by more than 48 %, protect the residents and industrial electricity consumption, and provide safe operation of the smart grid. Powerful guarantee.

With the advancement of globalization, the rapid development of circulating energy has attracted the attention of all walks of life. Timely monitoring of the accuracy assessment in the Chinese and English transfer environment has an extremely important impact on understanding and promoting the development of circulating energy storage equipment. In the process of Chinese and English transfer, the circulation of energy storage equipment has become more convenient and quickly accepted by users because of its wide application. Based on this background, this study proposed an ICE-LDA model based on the LDA model for developing accuracy assessment in Chinese and English transfer. Through the method of transfer assessment in Chinese and English, the detection of transfer information in Chinese and English and the similarity of accuracy distribution are measured. This study will model the transfer data collected in Chinese and English. In this process, this study will use the TF-IDF algorithm to extract the transfer information in circulating energy storage devices to improve the accuracy of Chinese and English.

New data centers today demand more scalable and efficient solutions with the rapid advances in deep learning, distributed systems, and high-concurrency data processing. Despite the significant improvements in these areas, existing systems continue to suffer from latency, congestion control, and resource allocation, especially during high workloads and network failures. Two new frameworks for optimizing RDMA-based data flow in remote settings – ConVer and DynoFlow – are introduced in this research. ConVer employs user-space path management and multipath transfer to achieve low latency and high throughput, whereas DynoFlow provides a software-defined, modular approach that efficiently handles network failures and varying traffic loads. Per experimental results, ConVer significantly improves throughput and achieves a latency of 1.3 m s (99th percentile). DynoFlow fares better in coping with network failures and optimizing route utilization. We solve significant problems in modern data centers by demonstrating through extensive experimental evaluation that the frameworks significantly enhance throughput and resilience in distributed systems.

The objective of this study is to meticulously forecast wind power generation utilizing Wavelet Transform and neural network techniques. Leveraging extensive wind power generation data, we delve into the intricate characteristics of the dataset to formulate a prediction model that seamlessly integrates WRT and neural network principles. The aim is to attain a precise prediction of wind power generation, thereby furnishing invaluable decision-making insights for optimal dispatching and energy management in the realm of wind power generation. This endeavour holds significant implications for enhancing the efficiency and sustainability of wind energy utilization. Firstly, WRT is used to preprocess the wind power generation data. Utilizing wavelet decomposition and reconstruction techniques, we meticulously extract the pertinent information embedded within the data, thereby mitigating the detrimental effects of noise and outliers on prediction outcomes. Subsequently, a neural network is employed as the core prediction model, facilitating the realization of wind power generation prediction through the meticulous training of network parameters. To rigorously assess the predictive prowess of our model, we leverage actual wind power generation data for rigorous testing and benchmark it against other prediction methodologies. This comprehensive evaluation ensures the reliability and accuracy of our approach in forecasting wind power generation. The results show that the wind power generation prediction model based on WRT and neural network has high prediction accuracy and stability. Specifically, the prediction accuracy of the model on the test set is over 90 %, and the prediction error is reduced by over 20 % compared with the traditional prediction methods. This study also further analysed the source of forecasting error, found that the main reason is due to the uncertainty and volatility of wind power data. Therefore, the research proposed targeted improvement scheme, including optimizing WRT parameters, improving neural network structure and so on. The wind power generation forecasting model based on WRT and neural network proposed in this paper has high forecasting performance and practical value, and can provide strong support for optimal scheduling and energy management of wind power generation.

Ensuring personnel safety and operational reliability in 0.4 kV (kilovolt) distribution networks during maintenance remains a critical challenge, especially under diverse environmental and load conditions. Traditional fixed-threshold protection schemes often fail to adapt to dynamic fault signals caused by varying earth impedance, moisture, and nonlinear loads, resulting in false alarms or missed detections. To address these limitations, this paper proposes an adaptive protection strategy for 0.4 kV maintenance environments, combining harmonic distortion rate and transient waveform slope as detection criteria. An experimental platform was constructed, featuring a 10/0.4 kV transformer, dual-feeder topology, configurable loads, adjustable fault resistances, and arc generators to emulate realistic fault scenarios. Multi-point measurements of voltages, currents, and residual currents were conducted under varying conductor cross-sectional areas, fault locations, contact areas, and humidity levels. Results show that the proposed method accurately detects earth faults and electric shock events with higher sensitivity and specificity than conventional fixed-threshold methods, even under adverse conditions. The strategy effectively mitigates the impact of load switching and environmental noise, improving protection reliability.

Image monitoring and transmission systems widely used in smart grids face problems such as complex electromagnetic interference, privacy leakage and inefficient retrieval. To address these challenges, this paper proposes an anti-interference transmission algorithm for sensitive images of smart grid based on the combination of Tent chaotic mapping and deep learning. The method improves NSGA-II optimization algorithm through Tent chaotic mapping to achieve global convergence enhancement and local optimal jumping out ability; meanwhile, it integrates DenseNet deep features with traditional HOG, BOW, ColorSpace features, and adopts PCA dimensionality reduction to improve the anti-interference expression ability of the features. In order to safeguard privacy security and retrieval efficiency, this paper designs a fine-grained access control based on CP-ABE and ρ -stable LSH secure indexing mechanism to support efficient similarity retrieval in ciphertext state. The experimental results show that the method can significantly reduce the BER, stabilize the DC bus voltage, maintain the fast recovery of three-phase current and voltage waveforms, and outperform the existing schemes in terms of image retrieval accuracy and computational efficiency under the extreme interference conditions, such as high-voltage overlay discharge and power switching. This study provides a new theoretical support and engineering realization path for the safe and efficient transmission of sensitive images in smart grids.

The increasing deployment of photovoltaic (PV) systems necessitates efficient fault detection methods to ensure optimal performance and reliability. The performance of the PV panels can be significantly affected by faults and anomalies. Various machine learning techniques can be used for this purpose, each with its strengths and weaknesses. Developing a proper fault diagnosis strategy to identify PV faults for the smooth and uninterrupted operation of PV panels plays a vital role. This paper presents a comparative assessment of different machine learning techniques for photovoltaic (PV) panel fault diagnosis. This study aims to evaluate and compare the performance of various machine learning algorithms including support vector machine (SVM), k-nearest neighbour (KNN), decision tree (DT), discriminant analysis (DA) and ensemble classifier (EC) in identifying and diagnosing faults in PV panels. Through extensive experimentation and analysis, the effectiveness of these machine learning techniques in fault detection and classification is assessed in terms of accuracy, precision and efficiency. Additionally, the robustness and scalability of the models are examined under different operating conditions and fault scenarios. This study used simulated data to train and test machine learning models and evaluated their performance based on accuracy and precision. The findings of this study provide valuable insights into the applicability and performance of data-driven machine learning techniques for fault identification in PV panels, thereby contributing to the development of more efficient and reliable PV systems for the renewable energy industry.

A resilient distributed secondary control strategy is proposed for DC microgrids (MGs) with denial-of-service (DoS) attacks on communication channels to achieve accurate current sharing and voltage regulation. According to the designed resilient control strategy, when there is no DoS attacks, the quantization information is sent to neighboring distributed generators (DGs) only when the triggering condition is satisfied. When there is a DoS attack, the communication can not be carried out normally. Then, each DG would attempt to send messages to its neighbors at a preset period until the attack disappears. The maximum allowable communication attempt period is determined by theoretical analysis, which is one of sufficient conditions to ensure current sharing and voltage regulation. As a result, the control signals are updated by discrete-time quantized information, which greatly reduces the communication burden, and effectively defends against DoS attacks. The asymptotic stability of the system is analyzed to show the results that voltage regulation and accurate current sharing are achieved. The proposed control strategy is verified through simulations.

A transformer is one of the most important high voltage (HV) power apparatus in any electrical system. In transformers, liquid insulation, primarily mineral oil, is used in bulk. It acts as an insulation and cooling medium in the transformer. Therefore, it plays an important role in the safe and reliable operation of the transformer. It is understood from several studies that because of the shortage of petroleum oil and its non-environment-friendly nature, scientists, engineers, and researchers are concentrating their studies on synthetic oil, natural ester oil, so that they can replace mineral oil successfully. Studies related to the characterization of different types of natural ester oil are made for their superiority over mineral oil. However, very few studies have been conducted on different combinations of mineral oil (MO) with natural ester oil, such as coconut oil (CO) and rice bran oil (RO), as an alternative insulation. In this paper, a study has been made on different combinations of MO with CO and RO for their performance characteristic studies like breakdown voltage (BDV), flash point, DC resistivity, and tan delta on the lines of International Electrotechnical Commission (IEC) and American Society for Testing Materials (ASTM) standards. It is observed from the studies that different combinations of MO with CO and RO enhanced its BDV, which is the prime electrical property for studying the dielectric strength of any insulation. However, blended RO with MO has balanced properties in terms of dielectric strength, flash point, and DC resistivity compared to all other virgin oil and blended oils. It is also less toxic compared to the mineral oil. Therefore, blended rice bran oil with mineral oil may be considered an alternative to traditional mineral oil as it is cost-effective and less toxic. Therefore, it may be an efficient alternative to insulating oil for the transformer industries.

The increasing penetration of distributed energy resources (DERs) into Active Distribution Networks (ADNs) demands innovative control solutions to address critical challenges such as voltage stability and power quality. Existing decentralized control strategies often fail to handle the intricate interactions between control loops in highly coupled systems, leading to suboptimal performance, excessive reactive power requirements, and operational limitations. This paper introduces a novel multivariable control framework designed for voltage regulation in ADNs. The proposed framework incorporates advanced Control Allocation and Measurement Combination modules, developed through a systematic methodology that merges Process Systems Engineering and Power Systems principles. By formulating the control design as a bilevel mixed-integer nonlinear programming (BMINLP) optimization problem, the framework minimizes voltage deviations, reactive power usage, and the reliance on actuators and sensors, ensuring system stability even under extreme conditions such as significant active power generation drops. Dynamic simulations performed on the IEEE 33-bus system demonstrate that the proposed approach outperforms traditional decentralized controllers by effectively preventing reactive power saturation and delivering superior voltage regulation. The results indicate that the voltage at all nodes stays within the permissible limits, with a maximum deviation of 0.7 %. Moreover, it achieves these results with reduced hardware and fewer control loops, highlighting its resource efficiency. This work represents a significant step forward in the control of ADNs with high DER penetration. By addressing the limitations of conventional strategies, it offers a scalable and robust solution that optimizes control resources, enhances system stability, and supports the transition to decentralized, renewable-powered distribution networks.

Substations are one of the key elements in electrical power system network. Efficient grounding in substation design forms the base for ensuring the safety of the person and equipments. Depending on the geometrical aspect of the substation, five different grid shapes are applicable. In this paper all 5-grid shape analysis is carried out to have comparison on performance analysis of the grid designing process. Also, in grid design calculations a concept of three different lengths is introduced which makes the grid designing simple and error free. Also, all the calculations are carried out owing to five different grid designing methods. Various softwares are used to make grid designing effective and lesser time consuming. Mathematically calculated results incorporating the length updation with grid shape applications are verified by ESGSD software. The methodology developed in this paper is applicable to any shape of the grid and ensures the safety and reliability of the grounding grid.