Energy Informatics最新文献

In the context of intensifying competition within the power market, power companies face the dual challenges of enhancing customer loyalty and optimizing marketing strategies. This study addresses these challenges by employing the long-term and short-term memory (LSTM) network model to analyze data-driven power marketing strategies and their impact on customer loyalty. The LSTM model is trained on a dataset combining time-series power consumption data with customer interaction scores and market response rates. This enables the model to predict and explain customer responses to marketing efforts with greater accuracy. Unlike traditional marketing models, which lack the capacity to capture dynamic customer behavior over time, the LSTM model accounts for both the temporal nature of consumption patterns and static customer feedback, offering a more holistic view. Key findings indicate that improving the quality of customer service interaction and accurately targeting marketing activities significantly boosts customer loyalty. In particular, customer interaction scores and market response rates are the most influential factors driving customer loyalty, providing critical insights for companies to adjust their strategies effectively. This study’s novelty lies in its application of advanced machine learning methods, such as LSTM, to the power industry—a sector traditionally slower to adopt such innovations. By bridging this gap, the research provides actionable recommendations on how power companies can implement data-driven marketing strategies to improve service quality, increase customer retention, and enhance their competitive position in the market. Additionally, the results underscore the model’s effectiveness in forecasting and optimizing marketing outcomes, offering a scalable solution for the evolving power sector. In the power market, companies face challenges in enhancing customer loyalty and optimizing strategies. This study employs the LSTM network model, trained on combined time-series power consumption, customer interaction scores and market response rates data. Unlike traditional models that struggle with dynamic customer behavior capture, LSTM accounts for consumption pattern temporality and static feedback. It outperforms other techniques like Random Forest and XGBoost in handling time-dependent consumption data. The key findings highlight the importance of customer interaction and targeted marketing. By applying LSTM, power companies can better predict customer responses, optimize marketing, improve service quality and enhance competitive position, providing a scalable solution for the evolving power sector.

This paper proposes a novel approach for transformer fault diagnosis. Initially, a high-dimensional feature set comprising 19 features related to five gas concentrations is constructed to reflect the gas-fault relationship. Subsequently, the Shapley Additive Explanations (SHAP) method is employed to evaluate feature importance and select a subset that significantly influences model predictions, thereby simplifying the model and enhancing its interpretability. Following this, the bald eagle search (BES) intelligent optimization algorithm is utilized to optimize the hyperparameters of the light gradient boosting machine (LGBM) model, further improving its predictive capability. Comparative experiments with various traditional machine learning models validate the effectiveness of the proposed method. The SHAP-BES-LGBM model achieves the highest accuracy of 0.9509 and an f1 score of 0.9606 on the test set, with only 11 samples misclassified, demonstrating superior classification performance and underscoring the advantages of this integrated approach in transformer fault diagnosis.

The evolution of the electrical grid from its early centralized structure to today’s advanced “smart grid” reflects significant technological progress. Early grids, designed for simple power delivery from large plants to consumers, faced challenges in efficiency, reliability, and scalability. Over time, the grid has transformed into a decentralized network driven by innovative technologies, particularly artificial intelligence (AI). AI has become instrumental in enhancing efficiency, security, and resilience by enabling real-time data analysis, predictive maintenance, demand-response optimization, and automated fault detection, thereby improving overall operational efficiency. This paper examines the evolution of the electrical grid, tracing its transition from early limitations to the methodologies adopted in present smart grids for addressing those challenges. Current smart grids leverage AI to optimize energy management, predict faults, and seamlessly integrate electric vehicles (EVs), reducing transmission losses and improving performance. However, these advancements are not without limitations. Present grids remain vulnerable to cyberattacks, necessitating the adoption of more robust methodologies and advanced technologies for future grids. Looking forward, emerging technologies such as Digital Twin (DT) models, the Internet of Energy (IoE), and decentralized grid management are set to redefine grid architectures. These advanced technologies enable real-time simulations, adaptive control, and enhanced human–machine collaboration, supporting dynamic energy distribution and proactive risk management. Integrating AI with advanced energy storage, renewable resources, and adaptive access control mechanisms will ensure future grids are resilient, sustainable, and responsive to growing energy demands. This study emphasizes AI’s transformative role in addressing the challenges of the early grid, enhancing the capabilities of the present smart grid, and shaping a secure, efficient, and adaptive next-generation grid aligned with future needs.

With the increasing demand for renewable energy in new rural areas, the integration and optimization of multi-energy systems such as wind and photovoltaic have become a key issue in distribution network planning. Existing methods are difficult to cope with the volatility and uncertainty of energy sources, resulting in uneven load distribution, high energy loss and low system efficiency. In this paper, the fuzzy algorithm is used to optimize the multi-energy distribution network, and the efficiency of the system is improved by real-time scheduling and load balancing. The results show that the fuzzy algorithm can effectively improve the utilization rate of renewable energy, reduce energy loss, and improve the stability and load matching degree of the system, which provides an optimization scheme for the new rural multi-energy system.

This study presents an optimization framework for residential energy management that integrates photovoltaic (PV) systems, battery storage, and demand response strategies. The primary objective is to minimize electricity costs while ensuring efficient use of renewable energy resources. The proposed method utilizes the Meerkat Optimization Algorithm (MOA), which is compared against other optimization algorithms, including Genetic Algorithm (GA), Particle Swarm Optimization (PSO), and Teaching-Learning-Based Optimization (TLBO). The results show that the proposed MOA achieves significant cost reductions. For example, under Time-of-Use (TOU) tariffs, the total electricity cost is reduced by 14% compared to the base case, while under Real-Time Pricing (RTP), the reduction is 16%. The optimized system also yields a 5 kW PV system and a 10 kWh battery, compared to 3 kW PV and 6 kWh battery in the GA and PSO cases. Additionally, the MOA provides a more computationally efficient solution, with a calculation time of 73 s, compared to 91 s for GA and 102 s for PSO. This study demonstrates the effectiveness of the MOA in optimizing residential energy systems, providing a robust solution for reducing electricity costs while integrating renewable energy sources. The approach is generalizable to other energy management applications and can be adapted for various regions and household configurations.

With the transformation of the energy structure, distributed photovoltaic (PV) power generation has become increasingly important. However, due to uncertain factors such as weather, equipment, and load demand, the consumption problem is prominent, which restricts the healthy development of the system. It is important to accurately measure the absorptive capacity of distributed PVs, but there are many shortcomings in existing research methods. This paper proposes a comprehensive measurement method to solve this problem and thus conducts a comprehensive metric study of the distributed PV Consumption Capacity considering multiple uncertainties. Based on the output uncertainty and load uncertainty of the distributed PV power generation, a mathematical model of the distributed PV power generation uncertainty is constructed. Based on the distributed PV operation data under various uncertain factors, considering the PV capacity and active power loss connected to the distribution network as objective functions, and setting constraints such as power balance, node voltage, line power flow, and distributed PV output, a comprehensive measurement model of the distributed PV absorption capacity is constructed. A local chaotic search is introduced to improve the firefly algorithm, and the improved firefly algorithm is used to solve the comprehensive measurement model and output the comprehensive measurement results of the absorption capacity. The experimental results show that this method can effectively evaluate the absorptive capacity. In a typical IEEE 32 - node distribution network, the network loss is 30 kW when PV access reaches 534 kW. This method is better than other methods in terms of maximum absorptive capacity, annual PV absorption, and annual network loss, and provides a scientific basis for the planning, operation, and management of distributed PV systems.

In the context of extensive integration of renewable energy sources into the electrical grid, the grid's fault transient behaviors have undergone significant changes. However, conventional single-unit equivalent models fail to accurately capture the fault transient responses of combined wind-solar-hydro power stations and often require substantial computational resources, leading to reduced simulation efficiency. This study proposes a cluster-based equivalent modeling approach for hybrid wind-solar-hydro power plants using the Sparrow Search Algorithm. Key factors influencing fault characteristics, including the distance to the Point of Common Coupling, DC-side current-limiting measures, irradiance levels, water flow rates, wind speeds, and reactive power at the outlet, are identified and used to construct a transient model. Euclidean distances are computed for these factors, and initial clustering centers for wind turbines are determined using an improved max–min distance technique. These factors and clustering centers serve as the training dataset to establish the clustering equivalent model. Simulation results, conducted on the MATLAB2022 platform, demonstrate that the SSA-based model outperforms the single-unit equivalent model by over 150 times in terms of accuracy. Additionally, the SSA-based model achieves a delay time, defined as the time required to compute the system's transient response after a fault, of less than 5 ms, which is less than one-twentieth of the delay time of the single-unit equivalent model. These improvements highlight the model's ability to accurately capture dynamic power responses under various disturbances, making it highly suitable for real-time applications in hybrid renewable energy systems.

To optimize the interaction effect between internal units and demand response of virtual power plants and enhance their transaction profit, a study on the the rolling optimization method of demand response for virtual power plants based on Bayesian Stackelberg game is conducted. Following the construction of a virtual power plant model and analysis of its operation strategy and process content, this method employs a power demand forecasting approach based on multidimensional fusion and Bayesian probability update to forecast the demand-side power requirements within the jurisdiction of the virtual power plant. Utilizing the forecast results of dynamic electricity demand, a demand response elastic matrix for virtual power plant is constructed through a rolling optimization model based on Stackelberg game. The two optimization objective functions, maximizing the supply-side income and minimizing the demand-side electricity purchase cost of virtual power plant, are transformed into maximizing the profit of power transaction for the virtual power plant. This is iteratively solved using the whale algorithm to determine the optimal power generation distribution scheme for each unit on both the supply side and demand sides. Upon testing, this method demonstrates not only the capability for peak shaving and valley filling but also improves the operating profit of the virtual power plant and optimizes user satisfaction, resulting in a relatively high comprehensive benefit index.

In the context of today’s global environmental challenges, manufacturing enterprises are gradually taking green technology innovation as a strategy to enhance sustainable development ability. This study discusses the application of hybrid intelligent technology in promoting green technology innovation decision-making in manufacturing enterprises. Through data preprocessing and model construction, it is found that energy consumption, emissions, standard compliance, environmental quality and market response are closely related to the green technology innovation score of enterprises. The results show that efficient energy management and active compliance with environmental standards have a significant impact on improving the environmental performance and technological innovation of enterprises. The market’s positive response to green technology has significantly promoted the rapid development and application of the technology. This study not only provides manufacturing enterprises with strategy and decision support for green technology innovation, but also provides policy makers with insights for promoting sustainable development. Through in-depth analysis, this paper emphasizes the importance and effectiveness of comprehensive application of hybrid intelligent technology in the process of green technology promotion.

The integration of large-scale renewable energy into the grid has significantly advanced research on flexible DC distribution networks. However, the potential of flexible loads—possessing both source and load characteristics—in supporting the low-carbon economic operation of integrated energy systems (IES) remains underexplored. Furthermore, the optimization of IES scheduling is inherently a multi-dimensional nonlinear problem, where traditional intelligent optimization methods struggle to achieve satisfactory solution accuracy. In this paper, an IES model is developed based on the concept of an energy hub, incorporating elements such as wind turbine output, photovoltaics, energy storage systems, gas turbines, and flexible loads, while considering the transferability, interruptible nature, and reverse energy flow characteristics of demand-side flexible loads. To address the current challenges in balancing environmental and economic benefits in IES, a carbon trading strategy and demand response mechanisms are applied to the optimization scheduling process, with the objective of achieving low-carbon and low-cost operations. The proposed model is solved using a novel Hiking Optimization Algorithm (HOA), and comparative analysis across different scenarios is conducted to investigate the impact of the carbon trading strategy on low-carbon operation, alongside an evaluation of the system’s economic and environmental performance under reasonable scheduling of both the carbon trading strategy and flexible loads. The results indicate that the total cost and carbon emissions of the system decreased by 8.98% and 15.13%, respectively, indicating that appropriate scheduling of the carbon trading mechanism and flexible loads effectively improves the system’s economic and environmental performance. In addition, a comparative study with traditional particle swarm and genetic algorithms demonstrates that the HOA, by incorporating adaptive mechanisms for both search space resolution and speed adjustment, enhances both global exploration and local exploitation, effectively avoiding local optima traps. This leads to improved optimization accuracy, further validating its effectiveness in IES optimization.