Energy Informatics最新文献

In the context of rapid developments in artificial intelligence and the clean energy industry, the optimal scheduling of clean energy storage and charging systems has become increasingly prominent. This study proposes an optimal scheduling method that integrates Differential Evolution (DE) and Kernel Search Optimization (KSO) algorithms. By incorporating DE’s mutation, crossover, and selection operations into the KSO framework, the method effectively avoids local optima while retaining KSO’s advantages in handling complex structures and large-scale data. Experimental results demonstrate that the convergence speed of the fusion algorithm is improved by 34.2%, 30.8%, 28.6%, and 23.4% over four other algorithms for hybrid functions, and by 56.7%, 52.9%, 25.3%, and 21.4% for combined functions. Additionally, the utilization of renewable energy increased from 40% to nearly 70% within 24 h. It can be seen that the convergence speed and renewable energy utilization of the fusion algorithm are significantly improved compared with the four baseline methods, highlighting its effectiveness in large-scale clean energy systems. This research provides an effective scheduling strategy for optimizing clean energy storage and charging systems. This study provides an effective scheduling strategy for optimizing clean energy storage and charging systems, and supports scalable and efficient energy management of urban and rural energy grids. The results show that the optimization of the integrated charging system can not only achieve optimal scheduling in a shorter time, but also reduce operating costs and resource waste, and effectively improve the overall operating efficiency of the energy system. Research to promote the efficient use of renewable energy will help reduce dependence on fossil fuels, thereby reducing greenhouse gas emissions and environmental pollution, which will have a positive impact on achieving the Sustainable Development goals and addressing climate change, and promote a win-win situation for the economy and the environment.

With a focus on larger rooftop or utility-scale solar systems, there is a lack of research on the potential impact of mini photovoltaic (MPV) systems, often referred to as balcony power plants. This work analyzes the impact of varying concentrations of MPV systems, on the stability and control of low-voltage (LV) grids. We offer a comprehensive technical assessment of MPV within a distribution grid and quantify their effects on power quality, losses, transformer loading, and the performance of other inverter-based voltage-regulation devices. For this purpose, this paper introduces the open-source Python-based framework PIDE (Photovoltaic Integration Dynamics and Efficiency), a tool for simulating the integration of distributed energy resources (DER)s and evaluating their impact on autonomous reactive power control in the distribution grid. Our case studies include a one-year sensitivity analysis based on Monte Carlo simulations, compare distributed and decentralized DER control strategies, and demonstrate the role of autonomous inverters in providing ancillary services. With the growing use of battery energy storage (BES) systems in LV grids for these services, the need for adaptable DER control strategies becomes increasingly evident. Our results show that high MPV penetration increases mean transformer load by up to 3%, line load by 2.5% and total power losses by around 17%.

As the rate of energy consumption in intelligent buildings increases, the uneven distribution of energy among different devices in intelligent buildings leads to further acceleration of energy consumption. The study suggested designing an energy dispatch system for intelligent buildings based on the enhanced multi-objective ant-lion optimizer algorithm in an attempt to address the issue that the conventional energy dispatch system for intelligent buildings is unable to carry out energy dispatch in accordance with the electricity price and incentives. The initialization of different energy data parameters was carried out by the multi-objective ant-lion optimizer algorithm, and the variance crossover operation of the data parameters was carried out by the differential evolution algorithm. Based on the improved multi-objective ant-lion optimizer algorithm, a demand response model was constructed, and the energy dispatch system of intelligent buildings was constructed accordingly. The results revealed that the area under the PR curve of the improved multi-objective ant-lion optimizer algorithm was 0.9653, which was significantly higher than the other three algorithms. The root mean square error and the mean absolute error of the algorithm were 0.839 and 0.648, respectively. In the experiments on the practical application of the dispatch system, it was found that the average power of the dispatched energy sources was significantly lower than that of the non-dispatched energy power distribution. The aforementioned findings indicate the suggested approach can more effectively schedule various energy sources in intelligent buildings, offering technical assistance in the area of energy dispatch.

Existing energy storage system is difficult to balance the energy distribution and dynamic response efficiency issues of lithium-ion batteries and supercapacitor, resulting in low energy utilization. Therefore, the study proposes a hybrid energy storage system for intelligent electric vehicles incorporating improved particle swarm optimization. The study analyzes the relationship between vehicle driving speed and power demand through equivalent model, constructs an objective function containing power demand and state of charge, and uses an improved algorithm for optimization and solution. The performance test results indicated that the proposed improved algorithm exhibited the fastest convergence speed by rapidly decreasing the objective function value and approximating the optimal solution within the first 20 iterations in both single-peak and multi-peak functions. The simulation experiments were validated under urban working conditions and highway working conditions, respectively. The results indicated that the energy efficiency in both working conditions was improved to 92.5% and 94.9%, respectively. In addition, good results were achieved in the contribution of supercapacitor, which were 27.2% and 29.6%, respectively. In the test results based on HIL environment, the system proposed by the research institute can also maintain energy efficiency of over 80% under extreme conditions. The findings support the optimal design of intelligent electric vehicle energy storage systems both theoretically and practically, showing that the study’s revised algorithm performs well in both energy allocation efficiency and dynamic response performance.

The current power plant production systems face issues such as insufficient monitoring accuracy, data transmission delays, and low energy utilization efficiency. In response, this study proposes a real-time monitoring and intelligent data analysis system based on Fifth-Generation Mobile Communication Network (5G) technology. Building upon an analysis of the limitations inherent in traditional systems, the experiment capitalizes on the extensive connectivity capabilities of 5G to design an intelligent monitoring architecture tailored for power plant production environments. To enhance system performance, the study introduces an innovative resource scheduling and data analysis model that combines an improved Hybrid Advantage Actor-Critic (A3C) algorithm with a Dueling Deep Q-Network (DQN) algorithm. This model integrates the global optimization capabilities of the A3C algorithm with the efficient learning mechanism of the Dueling DQN algorithm to optimize communication resource scheduling and energy storage management within a 5G Cloud Radio Access Network (C-RAN) environment. Simulation experiments demonstrate that this approach significantly improves system energy efficiency, optimizes resource utilization, and reduces energy waste. The results show that data transmission delays decreased by 25%, energy utilization increased by 18.25%, and renewable energy consumption rose by 12.55%. This study offers a new technical approach for the intelligent upgrade and green, efficient operation of power plant production systems, providing both theoretical and practical support for the optimization of power systems in the 5G era.

The growing complexity and need for electricity in contemporary grids have resulted in an increased dependence on Distribution Automation Technology (DAT) to improve the effectiveness and reliability of distribution networks. Automation technologies, like smart sensors and fault detection systems, are critical for enhancing operational efficiency and lowering power outages in distribution networks. This study investigates the influence of distribution automation on the dependability of electricity networks, concentrating on important functional metrics and their relationship with network efficiency. Objectives: The main objective of this research is to examine the factors that influence the reliability of distribution networks, with a focus on distribution automation technology. This study uses a variety of efficiency indicators, like automation coverage, fault detection time, and consumer complaints, to discover the primary factors of network reliability. This paper introduced the Reliability-Optimized Meta-Learning Ensemble (ROME) algorithm, which seeks to predict the reliability category of various areas using these indicators. Methodology: This study utilizes the Distribution Network Reliability Dataset, which includes several areas with a variety of characteristics such as network age, automation coverage, smart sensor installation, power outages, fault detection time, and other operational metrics. The ROME algorithm is used, which integrates numerous base models (SVM, Random Forest, MLP) and a meta-learner (Gradient Boosting) to predict each region’s Reliability Category (High, Medium, Low). The dataset is thoroughly preprocessed, which includes mean and mode imputation, label encoding, standardization, and SMOTE balancing. Recursive Feature Elimination (RFE) is used for feature selection. Results: The findings show a strong correlation between automation coverage, fault detection time, and reliability category. When compared to traditional classification techniques, the ROME algorithm surpassed SVM, RF, MLP, and GB models with 94.7% accuracy, 0.18 Log-Loss, 91.2% Jaccard Index, 0.08% fall-out, and 95.3% specificity. Conclusion: This research emphasizes the value of distribution automation in improving network reliability. Utilities and grid operators can use the ROME algorithm to better predict and enhance network reliability. The results highlight the requirement for targeted investments in automation technologies, particularly in regions with lower reliability scores, to guarantee sustainable and effective electricity distribution.

This paper aimed to assess new connotations and characteristics of power distribution networks in new situations like integrating photovoltaic (PV) systems. Power system emission reduction is an ongoing subject of discourse, and solar energy production using PV is gaining popularity. Centralized and unidirectional systems, nevertheless, provide difficulties. An investigation is expected to comprehend the network’s design and PV integration capacity’s (PV-IC’s) responsiveness to subsequent generations.With an emphasis on low and medium-voltage networks, the paper presents a unique dwarf mongoose optimization (DMO)approachfor developing efficient network configurations. It analyzes the effect of network configuration on PV-IC.This study is experimented with MATLAB/Simulink platform based on the DMO technique. Different PV system numbers and peak powers, together with alternate providing substations, have been modeled for a certain set of load locations. The load time series computed shows rural-urban zones, and the proposed DMO is implemented on several topological generations. The computed results indicate that network topologies must be changed to accommodate raised solar energy production and PV-IC, with rural regions attaining up to 8.2 kW using TF (+). Our proposed DMO approach surpassed alternatives, with rural regions having a higher PV-based load of 190 kW compared to 120 kW in urban areas. Voltage control tactics, like RPC and Curt, benefit up to 55% of rural customers versus 15% in urban areas. Policy changes for household PV incorporation may be needed to maximize solar energy use.

Advancements in renewable energy technologies have positioned microgrids as essential applications of the Internet of Things (IoT), necessitating innovative energy management systems. This study introduces a dual-layer energy trading framework designed to optimize interactions among interconnected microgrids and users. The upper layer focuses on energy exchanges between microgrids, while the lower layer manages transactions among local users. A novel Energy-Trading Management Algorithm (ETMA), based on the Meerkat Optimization Algorithm (MOA), is proposed to tackle the complexities of this system. By integrating a multiblockchain structure with a Delegated Proof of Reputation (DPoR) consensus protocol, the framework ensures secure and private transactions while incentivizing compliance among participants. Experimental validation with real-world data from Guizhou demonstrates significant improvements in efficiency and utility for both users and microgrid operators (MGOs) compared to traditional methods. This approach sets a new benchmark for scalable, secure, and efficient energy management in microgrid environments.

Forecasting energy usage in buildings is essential for implementing energy saving measures. Precisely forecasting building energy use is difficult due to uncertainty and noise disruption.To achieve enhanced accuracy in predicting energy use in buildings, a deep learning approach is proposed. This paper proposes a customized convolutional neural network with Q-Learning (CCNN-QL) based reinforcement learning algorithm for predicting energy consumption in building.The suggested CCNN-QL model offers an auto-learning feature that predicts building energy consumption through an automated method, continually improving its predictive accuracy.To assess its performance, various building types were selected to study the factors influencing excessive energy consumption, and data were collected from multiple Chinese cities. The suggested model’s performance has been assessed using evaluation metrics, resulting in a low Mean Absolute Error (MAE) and Root Mean Square Error (RMSE), indicating superior accuracy relative to comparable studies. Experimental results indicate that the suggested technique has superior predictive performance across several scenarios of building energy usage.

The need to harness the flexibility of small-scale assets for system stabilization, including redispatch, is growing rapidly with the increasing prevalence of distributed generation, such as photovoltaic systems and heavy loads, in particular heat pumps and electric vehicles. Integrating these resources into the redispatch process presents special requirements: On the one hand, building trust with the owners of such assets requires privacy and a reasonable degree of autonomy and engagement. On the other hand, besides the system’s scalability and robustness, the verifiability and traceability of provided data are essential for grid operators who depend on the reliable provision of redispatch services. To date, research and practice have encountered significant challenges in defining a system that enables the inclusion of decentralized flexibilities while satisfying necessary requirements. To that end, we present a novel conceptual system design that addresses these challenges by combining a multi-agent system (MAS) approach with verifiable information flows through digital self-sovereign identities (SSIs) and Zero-Knowledge-Proofs (ZKPs). Single agents, as edge devices, operate locally and autonomously, respecting customer preferences, while MAS provide the ability to design robust, reliable, and scalable systems. SSI enables agents to manage their data autonomously, while ZKPs are used to protect users’ privacy through selective data disclosure which allows the verification of the correctness of information without disclosing the underlying data. To validate the feasibility of this design, a case study is included to demonstrate the functionality of key sub-processes, such as baseline optimization, aggregation, and disaggregation, in a realistic scenario. This case study, supported by a prototype implementation, provides initial evidence of the concept’s soundness and lays the groundwork for future evaluation through extensive simulations and field testing. Together, the technologies included in the conceptual system design balance full transparency for grid operators with autonomy and data economy for asset owners.