Volume 26, Issue 6 - Special Issue: Control Strategies for Distributed Renewable Energy Resources using AI

This paper aims to solve the problem of insufficient accuracy of abnormal behaviour recognition in smart grid monitoring. By applying the CNN (Convolutional Neural Network)- BiLSTM (Bidirectional Long Short-Term Memory) model, combined with spatial feature extraction and time series analysis, the detection ability of complex abnormal patterns is improved. The safety and stability of power grid operation are guaranteed, and efficient management and accurate fault diagnosis of smart grid are realized. Multidimensional time series data, including voltage, current, power factor, and other information, are obtained from sensors and equipment of smart grids, and data cleaning, abnormal annotation, and standardization are performed. The sliding window method is used to divide the time series data to adapt to the input of the deep learning model. The model structure includes a CNN layer for extracting local features, a BiLSTM layer for capturing time series dependencies, and finally a fully connected layer for abnormal behaviour classification. In the test set, the average accuracy of epochs within 30–50 times reaches 97.6 %. Experimental findings demonstrate that the accuracy, precision, and recall of the CNN-BiLSTM model on the test set are better than those of the traditional CNN-LSTM, BiLSTM, and Transformer models. The CNN-BiLSTM model can effectively enhance the accuracy of abnormal behaviour detection in smart grids and provide a reliable solution for the safety monitoring of power grids.

Electricity theft increases the line loss rate of the power grid, causing a greater burden to the power grid company. At the same time, in terms of social impact, electricity theft disrupts the normal order of electricity consumption and has a bad effect on society and electricity-using enterprises. Most of the methods of stealing electricity are by affecting the metering method of the electric energy meter or privately connecting the wires to make the electric energy less metered or not metered, which not only destroys the power metering device but also seriously affects the stable operation of the power grid. The main purpose of this paper is to research the design of an intelligent anti-stealing monitoring device based on ARM. In this paper, the overall scheme of the anti-stealing device involved is designed through the analysis of the characteristics of electricity stealing. At the same time, the main components of the system and the work to be completed by each part of the module are pointed out. Finally, the software flow of the system is designed, and the whole system is tested and analyzed. Experiments show that the on-site verification of each abnormal power meter is consistent with the software display results, indicating that the anti-power stealing monitoring device has a high accuracy rate in power stealing monitoring, has a good on-site power stealing monitoring effect, and accurately monitors abnormal tables and abnormal conditions at the same time, which can reduce the line loss rate, which proves the feasibility of the device.

Translation is the process of converting a text sequence from one language to other. Translation systems are often used to translate between various language texts, but they may also be used for speech or a mix of the two, such as text-to-speech or speech-to-text. Statistical machine translation (SMT) employs large target language models (LMs) to increase the fluency of produced texts, and it is often considered that “more data is better data” when developing language models. This work focuses on translating the Renewable Energy (RE) terms in English and Chinese languages. The developed work mainly focuses on two models namely; Translation language model and Context adaptability model. Initially, a model is developed that is capable of translating renewable energy terms from English to Chinese. Subsequently, context adaptability model is built, where, two processes are carried out. The two processes include, (a) building a word corpus related to RE and (b) building dictionaries related to RE. The dictionaries are built by generating sentiment lexicons using Pointwise Mutual Information (PMI) model. Finally, the language is trained using improved BERT model.

A vehicle that runs on electricity rather than conventional fossil fuels is known as an Electric Vehicle (EV). They have advantages for the environment making them more pollution-free and helping to reduce the cost. However, there are still certain obstacles to overcome, such as cost and a poor network for charging. Utilizing Hadoop helps maximize the effectiveness of charging, cut expenses, and enhance the EV experience in general. So, this paper presents the performance scoring model for new energy vehicles based on Hadoop for improving the charging efficiency of PHEVs and reducing fuel costs. By leveraging the power of Hadoop and implementing optimization algorithms, this model can determine the best charging hours for PHEVs, leading to more efficient and cost-effective charging. By utilizing Hadoop’s distributed computing framework, the model can process and analyze the data in parallel, enabling faster and more accurate. The developed Hadoop model can consider various parameters to determine the best charging hours for PHEVs. Moreover, the Enhanced Bird Swarm Optimization Algorithm (EBSO) integrated with the Hadoop model to minimize the operation cost of the Battery Swapping Stations (BSS) while EV charging. This can provide insights into the effect of PHEV fuel costs. By analyzing data on fuel consumption, electricity prices, and charging patterns; it can calculate the cost savings achieved by optimizing charging schedules. This information could be valuable for PHEV owners, helping them make informed decisions about when and how to charge their vehicles to minimize the operation cost of the BSS. With further advancements in technology, the developed model has the potential to significantly contribute to the widespread adoption of PHEVs and the transition to a more sustainable transportation system.

This paper explores electronic product design with a focus on embedded systems, highlighting the strategic approaches used by design engineers to balance innovation, efficiency, and competitiveness. It examines design practices across firms and introduces a framework that portrays design engineers as “administrative men,” adept at navigating project complexities by choosing among four distinct design paths: “Repeat Order, Variant Design, Innovative Design, and Strategic Design.” The “Repeat Order” path focuses on reproducing existing designs with minimal changes, while “Variant Design” adapts designs to specific requirements. “Innovative Design” aims for novel solutions, and “Strategic Design” integrates long-term innovation with business strategies. Each path reflects unique goals, resource demands, and methodologies. To support the selection of the most suitable design path, this study introduces a novel metric and a structured methodology for evaluating design yield and cost. This approach includes analyzing complexity, estimating yield, and calculating total costs, culminating in an objective function designed to enhance design efficiency and accuracy. By offering actionable insights and a comprehensive decision-making framework, the paper aims to optimize the embedded design process, addressing the evolving challenges and opportunities in electronic product development.

Fault detection becomes especially crucial for maintaining the dependability and safety of power equipment as power systems become more complicated. This work aims to increase the accuracy and efficiency of fault signal identification in power equipment by proposing a method based on nonlinear stochastic resonance and a Duffing chaotic oscillator to detect weak signals in power systems. First, this study develops a weak signal detection model based on the Duffing chaotic oscillator approach. By modifying the system parameters to put the signal in a critical condition, the model greatly increases the sensitivity of the signal to the tiny periodic sinusoidal waveforms. In the meantime, the stochastic resonance method’s high resistance to noise interference improves the dependability of signal identification. The experimental results demonstrate that the approach is more accurate and stable than the conventional method and can extract fault signal features when working with weak signals in high noise conditions. Ultimately, this paper’s research offers a fresh, efficient method for identifying power system faults.

In recent days, solar energy has achieved tremendous growth and technological developments that result in reduced cost requirements and increased power quality. Marketing such renewable products, branding and promotion is important because the customers are attracted through these branding names. Vocabulary selection of energy products for branding and promotion empowers the attention of customers and hence the selection of brand names is considered as the initial step in the marketing field. The visual attention of the data is extracted by the eye tracker. So, in this research work, vocabulary selection methods for solar panel branding and promotion in the Chinese language are analyzed. The awareness programs conducted by the Chinese government for the branding and promotion of solar panels are discussed. In addition, vocabulary selection methods with their requirements for branding are analyzed. Eye tracking-based vocabulary selection method is briefly discussed and the measures used for branding and promoting solar panels are also elucidated. After, vocabulary building is accomplished in the second stage through eye tracking-based vocabulary selection. The performance analysis is carried out in terms of correlation, Jaccard, cosine similarity and word error rate. The correlation between the eye tracking-based vocabulary selectionsis the same as that of the user feedback-based vocabulary selection. The selected vocabulary learning method could accomplish user satisfaction and perception in terms of various performance measures.

In the construction, operation and maintenance of modern power grids, the monitoring, acceptance testing, and operation and maintenance of distribution networks have always been key links. With the rapid development of technologies such as smart grids, the Internet of Things and artificial intelligence (AI), how to enhance the efficiency of distribution network acceptance testing and operation and maintenance management has become an urgent issue to be addressed. However, traditional distribution network acceptance testing and operation and maintenance management rely on manual operations, which are inefficient, error-prone, and lack data integration and intelligent support. In response to the above problems, this paper studies and designs an automatic verification platform for distribution network monitoring information based on AI. The platform first collects and integrates distribution network information, performs preliminary data processing, and then realizes automatic verification and anomaly detection of the monitored platform data through the long short-term memory (LSTM) network model. In addition, the particle swarm optimization (PSO) algorithm is utilized to optimize the task scheduling of the platform, and the greedy algorithm is used to optimize the platform’s load balancing, aiming to enhance the platform’s processing efficiency, response speed, and resource allocation capability. After acceptance testing, the outcomes demonstrate that the platform designed in this paper has good verification consistency for the distribution network, and the verification consistency of various major equipment remains above 90 %. It can detect abnormalities and give feedback for different monitoring points in different distribution network areas. The traditional operation and maintenance solution is compared with the platform designed in this paper regarding operation and maintenance management efficiency. The results show that in the seven key indicators of operation and maintenance management efficiency, the platform designed in this paper has shown obvious advantages, providing a practical solution for the intelligent management of the distribution network and contributing to the development of automation, intelligence, and efficiency of power grid operation and maintenance.

Accurate and timely wind power forecast is difficult because of the volatility and intermittency of wind energy, which are strongly influenced by meteorological conditions. In order to solve this problem, this research suggests a data-driven wind power prediction technique based on Elman neural networks and an enhanced generative adversarial network (GAN). First, outliers in the supervisory control and data acquisition (SCADA) power and wind speed data are found by combining the methods of density clustering and regression threshold truncation. To guarantee the temporal continuity of wind power data with the original characteristics, minority class samples are then generated using an enhanced GAN and added to the dataset. In order to obtain an accurate wind power prediction, the balanced data is then fed into a prediction model that combines the ABC algorithm and Elman neural networks. The fitness value is the mean squared error (MSE) of the training data, and the optimization objective is the connection weights of the Elman neural network. Results from experiments show that the suggested model outperforms earlier state-of-the-art techniques and significantly increases wind power prediction accuracy. It also has the advantages of high stability and quick convergence speed, and it can capture the long-term dependencies of wind power data.

Energy Engineering is a well-liked branch of engineering, wherein the graduates in this discipline are estimated for their English skills under several cases. The capability of listening broadly enhances the effectiveness of English listening teaching; thus, this makes college English listening lessons dynamic, engaging, information-rich, and facilitates an immersive learning experience for students. This is to cultivate students’ attention in learning, form learning inspiration, and finally, attain the objective of enhancing their English skills. This work explores the findings from experimental work associated with establishing an educational setting specifically designed for the foreign language instruction of upcoming energy engineering students. The aim of this work is to a) design a series of short interactive videos or reusable learning objects (RLOs) covering an extensive area of psychosocial and practical problems appropriate to the listening capability of Energy Engineering students; b) establish the ease of access, suitability, take-up, and adherence of the RLOs; and c) evaluate the advantages and cost-effectiveness of the RLOs. Ultimately, the experimentation analysis is performed for the student’s average score for basic level, middle level, and higher level students from five universities. The analysis states that regarding university 5, the basic level females attained the highest score of 92.9, while the middle-level female attained the average score of 86.27. Meanwhile, for higher level, both attained average score of 68.89. From the analysis, it is clear that the basic level student’s average score is better than the middle level and higher level.

Power metering system electricity theft may jeopardize the security and stability of the power system in addition to reducing the financial gains of the power grid. This research suggests a power theft detection model based on the Pearson correlation coefficient and support vector machine (SVM) to increase the detection accuracy. The relationship between the user’s daily electricity consumption and the line loss of electricity of the day in the station region is first examined using the Pearson correlation coefficient to determine whether a user is committing power theft. The SMOTE technique is used to oversample a few classes of samples to address the issue of the limited amount of data on electricity theft users. This prevents sample imbalance from having an impact on the detection model. The SVM algorithm is then used to build the electricity theft detection model, and the Gaussian kernel function is used to translate the linearly indivisible power consumption data to the high-dimensional space to guarantee the dataset’s linear divisibility. With a classification accuracy of 93.8 % and a positive sample classification accuracy of 95.4 % for the detection of electricity theft users, the experimental portion uses the State Grid of China’s actual electricity consumption dataset. Following model training and cross-validation, the results demonstrate that the model can successfully identify electricity theft, demonstrating the model’s superiority and applicability in this area.

Traditional monitoring systems rely on different types of sensors in the field of new energy grid connection monitoring. The collected data have inconsistent problems; the monitoring method is not real-time enough; the fault detection and anomaly recognition accuracy is low. This paper combines intelligent sensing technology and GNNs (Graph Neural Networks) to design a more efficient, smart, and precise new energy grid connection monitoring information verification method. High-precision intelligent sensors are used to collect multidimensional data of the new energy grid connection system in real-time, and data fusion technology is used to solve the data inconsistency problem between different sensors to ensure efficient and accurate data collection. The graph neural network algorithm framework is used to build the relationship diagram of nodes and edges, and the GNN model is used for information verification and fault detection. The advantages of graph structure are used to accurately obtain the information transmission of each node and improve the accuracy and real-time performance of fault detection. The collected data of the intelligent sensor and the graph neural network model are synergistically optimized to form a closed loop of data processing, model training, and fault prediction. The experimental results show that among the 10 different folds, the GNN model has a lower loss value, with an average loss value of 0.312, which can reduce the error in the information transmission process when processing the monitoring information of the new energy grid connection system; the fault recognition rates of the GNN model in abnormal voltage, current, frequency, and temperature scenarios are 0.92, 0.87, 0.9, and 0.85, respectively, which is suitable for complex fault detection tasks.

The incorporation of Hybrid Energy Storage Systems (HESS) into alternating current (AC) electrified railway systems has attracted considerable attention. Nevertheless, only a little focus has been placed on optimizing the dimensions and daily delivery of HESS throughout the entire project duration. This research introduces an innovative bi-level method for Railway Traction Substation Energy Management (RTSEM). At the master level, HESS dimensions are refined, whereas the slave level emphasizes daily allocation. The slave model utilizes Mixed-Integer Linear Programming (MILP) to control energy distribution, synchronizing HESS, renewable sources, and regenerative braking. An extensive cost evaluation takes into account battery deterioration and replacement expenses throughout the project duration. Using an integrated CPLEX solver, Ray optimisation is utilized to address the RTSEM challenge. The model is evaluated on an actual high-speed railway line scenario in China. Findings indicate that combining HESS with renewable energy yields considerable cost savings, as the suggested Ray Optimization-based model achieves a lowest fitness value of 0.016, in contrast to the highest fitness values of 0.033, 0.005, and 0.002 realized by BSO, GA, and GOA models, respectively.