Power Grid Management Research Articles

Multi-agent Deep Reinforcement Learning for cloud-based digital twins in power grid management

As the power industry becomes increasingly complex, Digital Twin (DT) technology has emerged as a crucial tool for enhancing grid resilience and operational efficiency by creating dynamic digital replicas of physical systems. These replicas enable accurate simulations and proactive management, but the vast amount of data generated by DT systems poses significant challenges for processing and analysis. Cloud computing offers a flexible solution by offloading these computational tasks to distributed resources, allowing for real-time analysis and scalable operations. However, this approach introduces complexities in task distribution and maintaining quality of service. Recent efforts have applied Deep Reinforcement Learning (DRL) to address these challenges, primarily using single-agent methods. However, these methods struggle with scalability and performance in increasingly complex cloud environments. To overcome these limitations, we propose an efficient task scheduling framework based on Multi-Agent Deep Q-Network (MADQN) principles, specifically designed to optimize both response times and operational costs. We provide a comprehensive design overview of our approach and conduct a thorough evaluation of its performance. The experimental results clearly indicate that our approach can considerably reduce response times amd lower operational costs compared to current methods, including state-of-the-art single-agent approaches.

Multimodal Operation Data Mining for Grid Operation Violation Risk Prediction

With the continuous expansion of the power grid, the issue of operational safety has attracted increasing attention. In power grid operation control, unauthorized operations are one of the primary causes of personal accidents. Therefore, preventing and monitoring unauthorized actions by power grid operators is of critical importance. First, multimodal violation data are integrated through information systems, such as the power grid management platform, to construct a historical case database. Next, word vectors for three types of operation-related factors are generated using natural language processing techniques, and key vectors are selected based on generalized correlation coefficients using mutual information, enabling effective dimensionality reduction. Independent component analysis is then employed for feature extraction and further dimensionality reduction, allowing for the effective characterization of operational scenarios. For each historical case, a risk score is derived from a violation risk prediction model constructed using the Random Forests (RF) algorithm. When a high-risk score is identified, the K-Nearest Neighbor (KNN) algorithm is applied to locate similar scenarios in the historical case database where violations may have occurred. Real-time violation risk assessment is performed for each operation, providing early warnings to operators, thereby reducing the likelihood of violations, and enhancing the safety of power grid operations.

Novel hybrid data-driven models for enhanced renewable energy prediction

Global power grid management depends on accurate solar energy estimation, yet present prediction techniques frequently suffer from unreliability as a result of abnormalities in solar energy data. Solar radiation projections are affected by variables such as anticipated horizon length, meteorological classification, and power measuring techniques. Therefore, a Solar Wind Energy Prediction System (SWEPS) is proposed as a solution to these problems. It improves renewable energy projections by taking sun trajectories and atmospheric characteristics into account. In addition to using a variety of optimization methods and pre-processing techniques, such as Principal Component Analysis (PCA), Recursive Feature Elimination (RFE), Least Absolute Shrinkage Selection Operator (LASSO), and recursive feature addition processes (RFA), complemented by a genetic algorithm for feature selection (GAFS). The SWEPS also makes use of sophisticated machine learning algorithms and Statistical Correlation Analysis (SCA) to find important connections. Neural Network Algorithms (NNA) and other metaheuristic techniques like Cuckoo Search Optimization (CSO), Social Spider Optimization (SSO), and Particle Swarm Optimization (PSO) are adopted in this work to increase the predictability and accuracy of models. Utilizing the strengths of machine learning and deep learning techniques (Artificial Neural Networks (ANN), Decision Trees, Support Vector Machine (SVM), Recurrent Neural Networks (RNN), and Long Short Term Memory (LSTM)) for robust forecasting, as well as meta-heuristic optimization techniques to fine-tune hyper-parameters and achieve near-optimal values and significantly improve model performance, are some of this work contributions to the development of a comprehensive prediction system.

Research on Power Grid Load Forecasting Method Based on Support Vector Machine

Abstract In order to improve the accuracy of power load forecasting, a power load forecasting model with the support vector machine (SVM) was proposed in the existing load forecasting methods. Firstly, establish the basic analysis on the SVM theoretical foundation. Through the establishment of SVM prediction model, the real power data is sent into the support vector machine prediction model for learning and forecasting once it has been processed. Think about the electricity grid load change rule states, our enterprises accurately understand the local power grid operation and scheduling measures by estimating the electricity uses. Meanwhile, it optimizes the power grid management through careful scheduling means, prevent excessive consumption of resources and large energy consumption. To achieve the prediction effect, this paper builds the prediction model and analyses the data with the MATLAB. In the experiment, the real value is compared with the anticipated value. Through the analysis of more than 8000 load data, the accuracy of the training set is 96.184%, and the accuracy of the test set is 96.6134%. The outcomes demonstrate the great accuracy of the SVM prediction model in predicting electricity load.

Research on the role of LLM in multi-agent systems: A survey

In recent years, the rapid development of large language model (LLM) has demonstrated superior performance in language understanding, text generation, planning, reasoning, and knowledge integration. This has led to the emergence of intelligent agents based on LLM. By leveraging the capabilities of LLM, these agents can effectively make decisions based on given objectives and possess certain learning and adaptation abilities. However, single-agent systems are generally suited to solving relatively simple problems and are limited in handling complex tasks that require coordination. For instance, in fields such as power grid management or traffic control systems, relying solely on a single agent is often insufficient for effective decision-making. In this context, adopting multi-agent systems proves to be more effective: through collaboration among multiple agents, each undertaking specific tasks, complex problems can be efficiently managed through interaction and coordination. This survey will analyze the role of LLM in multi-agent collaboration, discuss and analyze the current research challenges and key issues, and explore potential directions for future development.

Machine Learning Approaches for Short-Term Photovoltaic Power Forecasting

A photovoltaic (PV) power forecasting prediction is a crucial stage to utilize the stability, quality, and management of a hybrid power grid due to its dependency on weather conditions. In this paper, a short-term PV forecasting prediction model based on actual operational data collected from the PV experimental prototype installed at the engineering college of Misan University in Iraq is designed using various machine learning techniques. The collected data are initially classified into three diverse groups of atmosphere conditions—sunny, cloudy, and rainy meteorological cases—for various seasons. The data are taken for 3 min intervals to monitor the swift variations in PV power generation caused by atmospheric changes such as cloud movement or sudden changes in sunlight intensity. Then, an artificial neural network (ANN) technique is used based on the gray wolf optimization (GWO) and genetic algorithm (GA) as learning methods to enhance the prediction of PV energy by optimizing the number of hidden layers and neurons of the ANN model. The Python approach is used to design the forecasting prediction models based on four fitness functions: R2, MAE, RMSE, and MSE. The results suggest that the ANN model based on the GA algorithm accommodates the most accurate PV generation pattern in three different climatic condition tests, outperforming the conventional ANN and GWO-ANN forecasting models, as evidenced by the highest Pearson correlation coefficient values of 0.9574, 0.9347, and 0.8965 under sunny, cloudy, and rainy conditions, respectively.

Construction of integrated network order system of main distribution network based on power grid operation control platform

This study presents a major advance in grid management: the development and deployment of an integrated network command system for the main distribution network. The system integrates cutting-edge information technology, including modules such as command issuance, intelligent routing, security assurance and in-depth data analysis, opening a new era of refined and intelligent power grid management. The research focuses on the application of core technologies such as information communication technology, distributed control system, artificial intelligence and big data analysis, and strengthens the system operation foundation. The chapter on system architecture details the innovative integration of DDQN algorithm and attention mechanism, and carefully constructs intelligent scheduling engine and status monitoring and early warning system, which significantly improves real-time response, decision optimization and active security defense capabilities. Simulation experiments and actual case analysis verify the effectiveness of the system, specifically, the response time is reduced by 75.7%(from 2.1 s to 0.51 s in the traditional system), the data processing speed is still maintained at a high level under high load (100,000 data processing rate is 300/s), and the system stability is as high as 99.97%. The new system also achieved a high degree of automation, reducing annual operation and maintenance costs by 20%, and increasing user satisfaction to 90%, an increase of 28.6% over the previous period. These improvements not only optimize power quality and grid efficiency, but also further confirm that the fault response time is reduced by 30% and the user outage time is reduced by 25%. Therefore, this study not only highlights the innovation of the proposed system, but also demonstrates its significant contribution to accelerating the modernization of power grid management and ensuring safe operation with empirical data.

Data-driven electric vehicle usage and charging analysis of logistics vehicle in Shenzhen, China

The electrification of transportation is profoundly reshaping human society and presenting new challenges in terms of travel modes, infrastructure development, and energy supply. Given the potential for large-scale scheduling of electric logistics vehicles (ELVs), it is crucial to thoroughly analyze the usage characteristics and establish reliable models. This study examines the usage patterns and charging behaviors of 29 ELVs in Shenzhen, China, encompassing 34,856 trips and 14,464 charging events. Furthermore, behavior-time probability density models were constructed based on an improved Gaussian mixture model (GMM), which avoids the fitting error caused by misclassification of time series data across time nodes. The article also provides a comprehensive analysis of other statistical findings related to the travel and charging activities of ELVs. The conclusions drawn from this research can serve as valuable references for industries involved in infrastructure construction, power grid management, battery virtual aggregation, and similar sectors.

AI-Driven Solar Energy Generation and Smart Grid Integration A Holistic Approach to Enhancing Renewable Energy Efficiency

This paper comprehensively analyzes AI-driven solar energy generation and smart grid integration, focusing on enhancing renewable energy efficiency. The study examines applying advanced artificial intelligence techniques in optimizing solar power production, forecasting, and grid management. Machine learning algorithms, including Support Vector Regression (SVR) and Artificial Neural Networks (ANN), are evaluated for effectiveness in solar irradiance prediction and PV system performance estimation. The integration of AI in smart grids is explored, highlighting its role in demand-side management, energy storage optimization, and grid stability control. A holistic approach to improving renewable energy efficiency is proposed, encompassing integrated AI frameworks for solar-plus-storage systems, multi-objective optimization techniques for energy management, and AI-enabled microgrids and virtual power plants. The paper also addresses the challenges and future trends in AI application to renewable energy systems, including scalability issues, regulatory considerations, and ethical implications. By leveraging big data analytics and advanced AI algorithms, this research demonstrates the potential for significant improvements in overall system efficiency, reliability, and sustainability of solar energy systems integrated with smart grids.

Short-Term Photovoltaic System Output Power Prediction Based on Integrated Deep Learning Algorithms in the Clean Energy Sector

Photovoltaic power generation system plays an important role in renewable energy. Therefore, accurately predicting the short-term output power of photovoltaic system has become a key challenge for real-time power grid management. This study focuses on Yingli's green energy photovoltaic system, and uses the convolution neural network and long-term and short-term memory network fusion model (CNN-LSTM) to predict the short-term power. The model integrates CNN's data feature extraction and LSTM's time series prediction ability, showing high accuracy and stability. The experimental results show that CNN-LSTM model has a low mean and variance of prediction error, and the prediction is stable and reliable, and it is consistent in different scenarios. This provides theoretical support for the output power prediction of photovoltaic system based on deep learning.

Intelligent Video Monitoring and Analysis System for Power Grid Construction Site Safety Using Wireless Power Transfer

Power grid construction significantly enhances power grid management and risk control. Inconsistent operations at construction sites can jeopardize grid stability and crew safety. Traditional power lines are less favored due to mobility limitations, while batteries add burden and impracticality. To address this, a wireless power transfer and video monitoring system is developed using RF technology and Yolo V3 model. This enables continuous monitoring and employee safety analysis. The system's detection performance is optimized using HPSO, surpassing existing methods in accuracy and speed. It ensures real-time monitoring and improves the detection of potential risk sources, crucial for construction site safety.

Seasonal Electrical Load Forecasting Using Machine Learning Techniques and Meteorological Variables

Accurate forecasting of seasonal power consumption is crucial for effective grid management, especially with increasing energy demand and renewable energy integration. Weather patterns significantly influence energy usage, making load prediction a challenging task. This study employs machine learning algorithms, including Random Forest (RF), Artificial Neural Networks (ANN), and Decision Tree (DT) models, to forecast electricity consumption using meteorological variables such as solar irradiance, humidity, and ambient temperature. The impact of weather elements on load prediction accuracy across different seasons is explored using seasonal forecasting techniques. The results demonstrate the superior performance of ANN and RF models in forecasting summer and winter loads compared to the rainy season. This discrepancy is attributed to the abundance of data for the summer and winter seasons, and the ability of the models to capture complex patterns within the data for these particular seasons. The study highlights the potential of machine learning techniques, particularly ANN and RF, in conjunction with meteorological data analysis, for enhancing the accuracy of seasonal electrical load forecasting. This can contribute to more effective power grid management and support the transition towards a more sustainable energy landscape. The findings underscore the importance of data quality, quantity, and appropriate model selection for different seasonal conditions.

The role of grid management in community risk governance: a case study in Yuelu, China.

In this study, we aim to provide a comprehensive analysis of the effectiveness of the risk prevention and control mechanism within the grid management model for community risk prevention. We emphasize the importance of thoroughly examining the risk prevention and control mechanism to enhance risk management efforts in urban communities, particularly in response to unforeseen outbreaks such as COVID-19. Case studies are widely acknowledged as one of the most effective approaches to examine governance in China. In this study, the "Yuelu Model" serves as an illustrative example to demonstrate the application and effectiveness of grid management in community risk governance. To ensure the validity of the case study, it is imperative to adhere to the principle of representativeness. The collection of case data involves a combination of primary and secondary sources, and supplementary information is obtained through follow-up investigations conducted via WeChat, telephone, and other means, thereby enhancing the comprehensiveness and accuracy of the data. Our analysis reveals significant findings regarding the impact of the grid management model, fulfilling a triple role as a "Social Safety Valve" in the management process: (1) Community stress reduction function, (2) Community alarm function, and (3) Community integration function. Furthermore, we explore the adaptability of the grid management mechanism in addressing community risks, highlighting its effectiveness and potential for broader application. The findings of this study suggest that: Firstly, it is crucial to establish a shared information repository among different departments on a big data platform. Secondly, a dynamic government public information internal network should be established through collaborative efforts among multiple departments. Thirdly, implementing a regular (or periodic) early warning mechanism is essential. Lastly, the establishment of a high-quality talent team for power grid management is highly recommended. Our research provides valuable insights to enhance community risk governance.

A BPSO based collaborative management platform for multi-source data security in power grids

By designing a digital power grid multi-source data security collaborative management platform, the system configuration problem of the OMS system and the power grid management platform for the main distribution network of the power grid is solved. A design method for the digital power grid multi-source data security collaborative management platform based on discrete particle swarm optimization algorithm is proposed. Based on the design concept of SOA, realize the overall design framework of the platform according to the design method of multi-layer technical system in the business presentation layer, business process and composition layer, service layer, component layer and resource layer, realize the basic layer design of the system management platform through the basic application platform design scheme of XML configuration, implement the query, processing and output representation of the grid’s multivariate data using B/S architecture protocol, and use the Spring Framework The platform software architecture is implemented using J2EE technology and multi module component design scheme. The discrete particle swarm optimization algorithm is used for the fusion and scheduling of multi-source data in the digital power grid. The interface design and functional construction of the power grid management platform are implemented in the OMS system of the power grid main distribution network, and the logical model of the transformation project is constructed to achieve platform optimization and construction. Tests have shown that the designed digital power grid multi-source data security collaborative management platform has good human-machine interaction, strong data fusion scheduling ability, reduced resource and subsystem coupling, and supports the flexibility of physical deployment and maintenance.

The Construction and application of analytical framework of power grid management system

The power grid management system is an important institutional foundation supporting the development of the electric power industry. The current power grid management system analysis framework with the relationship between transmission and distribution as the core cannot explain the facts well. This paper builds a power grid management system analysis framework composed of "structure layer + factor layer", which can analyze the basic outline of a country's power grid management system and the driving factors behind it in a more comprehensive and systematic way.

Smart Grids Data Aggregation Method on Paillier Homomorphic Encryption

ABSTRACT Smart grid (SGs) is a highly integrated power system, it is gradually replacing the traditional power grid, but at present there are problems in the SGs data aggregation, such as user privacy leakage, grid data query inflexibly and data leakage. In order to solve these problems, this paper studied the SGs data aggregation method based on Paillier homomorphic encryption (HE). In this paper, the Paillier HE algorithm was used to study the power grid data, and the encrypted data was numerically calculated under different encryption states. It used the cloud computing center in the blockchain to directly aggregate the ciphertext data generated by users, and used the hash function to set the secret key to prevent the ciphertext data from being tampered with, and obtained the aggregate result of the original data after decryption by the power grid management center. When the number of security parameters was 30,000, the encryption time required by this method was 21.71 seconds. When the number of smart meters (SMs) was 80, the signature verification time required by this method was 2.31 seconds.

The new situation and optimization strategy of China's power grid management system

The system mechanism of power grid management is an important institutional basis to support the development of electric power industry, and has always been an important controversial issue in China's electric power system reform. Based on the current situation of China's power grid management system, this paper analyzes the new situation, analyzes the main problems, and finally, based on the background of new power system construction, systematically puts forward the overall thinking and specific strategies of China's power grid management system optimization.

Revolutionizing Power Systems Through Electrical Automation: Efficiency, Reliability, and Sustainability

This paper explores the transformative role of electrical automation technologies in the modernization and optimization of power systems. With a comprehensive analysis spanning power generation, distribution, and grid management, we delineate the multifaceted impacts of automation on enhancing operational efficiency, ensuring reliability, and promoting environmental sustainability. Through the lens of operational case studies and theoretical frameworks, we identify key insights into how electrical automation contributes to the streamlining of processes, reduction of human error, and integration of renewable energy sources. Moreover, we delve into the future trajectory of electrical automation, highlighting the potential integration of Artificial Intelligence (AI) and Machine Learning (ML) algorithms, the expansion of smart grid capabilities, and the paramount importance of cybersecurity in the digital era of power systems. Our discussion synthesizes these perspectives to underscore the pivotal role of electrical automation in driving the evolution of power systems towards greater adaptability, efficiency, and resilience. This paper posits that the future of power systems hinges on the strategic implementation of advanced automation technologies, guided by an overarching commitment to innovation, security, and sustainability.

Distributed Energy Resources Management System (DERMS) and Its Coordination with Transmission System: A Review and Co-Simulation

Ever-increasing penetration of distributed energy resources (DERs) in the power grids, alongside their numerous benefits, brings new challenges that call for enhanced solutions in the field of control and management of power grids. The majority of the available research have considered either distribution or transmission grids in their studies. In this paper, a comprehensive review of the effects of DERs on the distribution and transmission grids is performed. The focus of this paper is on hierarchical management methods in order to categorize different approaches and highlight the gaps. Moreover, a review is conducted in the field of the newly introduced distributed energy resources management system (DERMS) concept. A DERMS can facilitate the hierarchical energy management procedure due to its functionalities and broad capabilities. Hence, its implementation in energy management and its impact on the power grid will be assessed with the aid of a co-simulation platform that considers both transmission and distribution grids.

Self-Healing Grids: AI Techniques for Automatic Restoration after Outages

The reliability and resilience of power grids are paramount for sustaining modern society's energy demands. However, power outages resulting from natural disasters, equipment failures, or human errors remain persistent challenges. Traditional approaches to power grid restoration, relying heavily on manual intervention, often lead to delays and inefficiencies in restoring services. [1],[2] Recent advancements in artificial intelligence (AI) have spurred the development of self-healing grids capable of autonomously detecting, diagnosing, and restoring power after outages. This paper presents a comprehensive overview of AI techniques employed in self-healing grids and their applications in automatic restoration following outages. The traditional methods of power grid restoration, characterized by manual inspection and decision-making processes, are discussed, highlighting their limitations and challenges. Subsequently, the paper delves into various AI techniques employed in self-healing grids. Machine learning algorithms, such as supervised and unsupervised learning, are utilized for outage detection by analyzing historical data to identify patterns indicative of faults or anomalies. Fault diagnosis is facilitated through the application of Bayesian networks, neural networks, and fuzzy logic systems, enabling operators to accurately identify the root cause of outages and prioritize restoration efforts. Optimization algorithms, including evolutionary algorithms and reinforcement learning, play a crucial role in planning and coordinating restoration efforts to minimize downtime and maximize efficiency. The benefits of self-healing grids, including improved reliability, reduced downtime, and enhanced safety, are discussed alongside the challenges posed by data quality, scalability, and cybersecurity concerns. Finally, the paper outlines future directions, emphasizing advancements in AI techniques, integration with emerging technologies, and the importance of standardization and regulatory frameworks for the future of power grid management. DOI: https://doi.org/10.52783/pst.302