High renewable penetration and reduced system inertia introduce significant challenges for transient stability assessment and control. This article proposes a causality-aware, large language model-enhanced distribution-preserving graph representation learning framework (LLM-DP-GRL) for fast and accurate stability prediction and decision-making. The DP-GRL model captures both structural and distributional properties of network states, whereas large language models provide physics-informed priors that improve data efficiency and generalization under multicontingency and out-of-distribution scenarios. A causal intervention module further quantifies bus-level influence on stability margins, offering interpretable insights consistent with system dynamics. The learned surrogate model is integrated into a cooperative preventive-emergency control strategy, enabling real-time stability margin evaluation and optimization. Tests on the IEEE 39-bus and 118-bus systems show that LLM-DP-GRL achieves higher accuracy, faster convergence, and improved robustness compared with conventional machine learning, LSTM, and GNN-based methods. The proposed approach reduces online control computation from over 35 min (TDS-based) to 39 s while maintaining inference latency below 30 ms. These results demonstrate that combining graph learning, LLM-guided priors, and causal analysis provides an effective and scalable solution for stability assessment and emergency control in low-inertia, high-renewable power systems.

Federated learning (FL) on graphs shows promise for distributed time-series forecasting. Yet, existing methods rely on static topologies and struggle with client heterogeneity. We propose Fed-GAME, a framework that models personalized aggregation as message passing over a learnable dynamic implicit graph. The core is a decoupled parameter difference-based update protocol, where clients transmit parameter differences between their fine-tuned private model and a shared global model. On the server, these differences are decomposed into two streams: (1) averaged difference used to updating the global model for consensus (2) the selective difference fed into a novel Graph Attention Mixture-of-Experts (GAME) aggregator for fine-grained personalization. In this aggregator, shared experts provide scoring signals while personalized gates adaptively weight selective updates to support personalized aggregation. Experiments on two real-world electric vehicle charging datasets demonstrate that Fed-GAME outperforms state-of-the-art personalized FL baselines.

A guaranteed region of attraction (ROA) is essential for ensuring transient stability in DC microgrids. Numerical methods and control strategies for ROA design often rely on parameter tuning tailored to a specific operating scenario. When the steady-state load varies, both the operating point and the associated ROA change, causing fixed-scenario guarantees to break down and offering limited robustness. This letter develops a robust control synthesis method that ensures transient stability with guaranteed ROA. First, a robust stability condition is proposed to ensure a guaranteed ROA that adaptively moves with the operating point; then, we incorporate the condition into an optimal power flow (OPF)-like control synthesis problem that jointly ensures operational feasibility and robust stability guarantees while preserving computational efficiency. Simulation case studies validate the effectiveness of the proposed method.

As the transition to a renewable-dominated grid progresses, traditional offline system strength assessment methods, such as the offline short-circuit ratio (SCR), become inadequate for accurately assessing system strength in near-real time. This study proposes a novel online data-driven framework utilizing phasor measurement unit (PMU) data to estimate SCR in near-real time without requiring detailed grid models or parameters. A deep neural network (DNN) is employed to learn the correlation between the collected PMU data and SCR values across different grid nodes, allowing for efficient and accurate online system strength assessment. The proposed algorithm demonstrates versatility and applicability across various power grids and operating conditions, providing a significant advancement in the near-real time online evaluation of system strength. Across two systems, the estimator achieves testing mean square error (MSE) of 0.142 (IEEE–39) and 0.0191 (simplified Australian 14-generator grid), executes in 1.2–1.4  ms per window on a CPU, and remains robust to 5 % noise and 20 % data gaps (MSE ≤ 0.171/0.184 and 0.0289/0.0315, respectively. This work not only fills existing significant research gaps but also paves the way for improved system monitoring in modern power grids amidst the growing integration of renewable energy technologies.

With the increasing integration of renewable energy into regional power grids, significant spatial differences in carbon intensity have emerged. These differences highlight the need for carbon-aware workload allocation in geographically distributed Internet Data Centers, where aligning computational loads with low-carbon regions can enhance both environmental and economic outcomes. In this paper, we propose a two-stage optimization framework that integrates renewable-aware workload allocation and strategic carbon allowance procurement. In the first stage, a robust optimization model based on column-and-constraint generation is developed to manage uncertainties in workload demand and carbon prices, enabling stable and cost-effective workload distribution across regions with varying renewable energy penetration. In the second stage, a multi-class mean field game model is constructed to capture strategic interactions and behavioral heterogeneity among Internet Data Centers in carbon markets. We apply a Deep Galerkin Method to solve the resulting high-dimensional partial differential equations, yielding a robust and convergent procurement strategy. Simulation results demonstrate that the proposed framework achieves over 28% cost savings while ensuring carbon compliance and workload satisfaction. This study offers theoretical and practical insights for carbon-regulated Internet Data Center operations, and supports the broader integration of renewable energy in large-scale digital infrastructure.

This work presents a novel digital twin (DT) framework integrated with blockchain to ensure real-time optimized energy management of virtual power plants (VPPs). The framework design leveragesenergy trading forecasts for both day-ahead scheduling and real-time operation in a VPP consisting of solar generation, battery storage, and grid supply. Key inputs include forecasted solar irradiance, load patterns, real time sensor readings, and historical data, while the twin outputs span optimized energy dispatch, battery State of Charge (SoC), grid voltage stability, energy transactions, and load profiles for all participants. All critical sensor data and energy trades are securely recorded on the blockchain ledger, enabling transparent verification of operations rather than directly implementing the model on the DT making the overall design lightweight and fast responding. The proposed framework has been implemented and tested in MATLAB-Simulink, with validation against physical system performance to evaluate its effectiveness in efficient energy management. Experimental results show high fidelity between the twin and the physical VPP: real-time measurement discrepancies are essentially zero, with variances in RMS voltage close to 1 V and energy transactions remain below 2 W.

This paper investigates the planning of fast-charging stations when charging demands are highly uncertain. To address this issue, a stochastic programming (SP) model is formulated. Since handling continuous probability distributions is computationally difficult, the sample average approximation (SAA) method is applied. By using SAA, the original stochastic model is converted into a deterministic mixed-integer linear programming (MILP) format. However, directly solving this MILP can be very time-consuming for large-scale networks. Therefore, we design a Benders dual decomposition (BDD) approach. This algorithm improves the traditional Benders decomposition by using Lagrangian relaxation to generate tighter bounds. In our method, the master variables are transferred to the subproblem and subsequently priced in the objective function. We test our model on a 25-node network and the California state road network. The results show that, in comparison with direct exact solvers, the proposed BDD method markedly reduces computational time and iteration counts.

The rapid development of cloud computing, Big Data, and artificial intelligence has driven a surge in data center energy demand, while traditional power grids struggle to meet their high reliability and low-carbon requirements. To address this challenge, this paper proposes a co-optimization framework for islanded data center microgrids and develops a two-stage robust optimization model. The model leverages a coordinated control layer to centralize decision-making for the energy supply and computing resource layers, enabling joint optimization of energy management and workload allocation while minimizing total costs. For computing resources optimization, a cross-regional communication network is established to facilitate coordinated workload spatial allocation among geographically distributed data centers, enhancing server resource utilization and optimizing energy consumption. A quality of service constraint mechanism using dynamic voltage and frequency scaling technology is developed, alongside an average response time analysis based on queuing theory. To address uncertainties in wind power generation, load demand, and outdoor temperature, a data-driven risk-tunable modeling approach is proposed. The method constructs distributionally robust chance-constrained programming using Wasserstein distance, transforms it into a risk-bound optimization problem, and designs a polyhedral uncertainty set that adaptively adjusts to risk levels and sample sizes. Finally, case studies demonstrate the effectiveness of the proposed model in improving data center resource utilization, reducing energy consumption, and optimizing robustness adjustment capabilities.

The global transition to renewable-dominated power systems is reshaping grid operation while introducing new cybersecurity risks. As renewable technologies scale and rely on digital control and communication platforms, the cyberattack surface expands and exposes critical vulnerabilities. Here, we report an assessment of cybersecurity threats in renewable-dominated grids by examining representative attack scenarios, including false data injection into power control, denial of service on distributed energy resources and cloud platforms, inverter parameter manipulation, and GPS time synchronization spoofing. These threats are shown to compromise system stability, reliability, and resilience. We further evaluate current industrial practices, regulatory frameworks, and emerging standards in addressing these risks. We find that existing approaches remain insufficient for the complexity of renewable-dominated systems. We conclude by identifying the root causes of past failures and outlining research and policy directions to strengthen cyber resilience in future power systems.

Peer-to-peer (P2P) energy management facilitates decentralized resource allocation among prosumers, improving local hosting capacity for renewables and minimizing energy expenditures while ensuring data privacy through distributed coordination. However, conventional P2P energy management methods are confined to synchronous scheduling paradigms, creating synchronization bottlenecks that fundamentally conflict with the dynamic and decentralized nature of P2P energy management tasks. To bridge this gap, this paper focuses on resolving a class of dynamic energy management problems over asynchronous P2P (Asyn-P2P) transactive networks. We first recast the dynamic energy management problems into a saddle-point problem, and then propose a synchronous decentralized dynamic energy management algorithm, dubbed Syn-DYNA,based on operator splitting theory. To eliminate the global synchronization clock in Syn-DYNA, we introduce a random activation scheme, together with local buffers for latest state tracking, to develop an asynchronous variant of Syn-DYNA, namely Asyn-DYNA. Based on monotone operator theory, theoretical analysis proves a non-asymptotic linear convergence rate for Syn-DYNA and establishes the almost sure convergence ofAsyn-DYNA. Numerical experiments validate effectiveness of Syn-DYNA and Asyn-DYNA algorithms by tackling a dynamic energy management task over P2P transactive networks.

Integrated Energy Systems (IES) combined with District Heating Systems (DHS) are vital for reducing carbon emissions through multi-energy conversion and flexible heat utilization. However, coordinated heat and power dispatch (CHPD) in such coupled systems is large-scale and complex. Centralized optimization requires heavy data exchange and poses privacy risks, while distributed or federated learning methods still face limits in scalability, heterogeneity, and data security. This paper proposes a fog-based multi-agent framework for privacy-preserving CHPD in IES-DHS. The framework integrates decentralized federated learning with homomorphic encryption for secure and scalable optimization. A decentralized Federated Learning algorithm is developed to coordinate agents without sharing raw data. A fog-based structure is designed to match the IES-DHS topology, and homomorphic encryption is embedded to protect communication and aggregation. Case studies in Jilin Province verify the effectiveness of the proposed framework. Results show improved optimization performance, communication efficiency, and privacy protection, providing a practical and secure solution for low-carbon energy management.

Distributed power systems complement centralized grids by coordinating distributed energy resources (DERs) to achieve regional energy self-sufficiency. Scaling such systems raises four persistent challenges: decentralized coordination, fair economic settlement, trustworthy operation, and system optimization, all without a central authority. This paper proposes Proof of Energy (PoE), a blockchain consensus mechanism that addresses these challenges through cryptographically secured, contribution-proportional node selection. In PoE, block generation rights are tied directly to real-world energy contributions, enabling distributed consensus without centralized dispatch. An Energy Contribution Unit (ECU) model is introduced to map heterogeneous energy services onto a unified value metric via scarcity-weighted normalization. A Verifiable Random Function (VRF)-based proposal mechanism then ensures selection probability is strictly proportional to node contribution, preserving fairness and resisting manipulation. Case studies validate PoE across three dimensions: grid coordination, incentive fairness, and optimization efficiency. The result is a cryptographically secured, incentive-compatible framework for decentralized value distribution in energy systems.

Modern energy systems, including those supporting transportation electrification, are increasingly exposed to volatile market conditions and external events. Effective decision-making therefore requires the integration of structured operational data with unstructured contextual information. Existing studies on Large Language Model (LLM)-assisted energy systems have mainly applied LLMs to individual tasks such as forecasting, scheduling, or decision support, while forecasting and control are typically treated separately. As a result, semantic information extracted from external events is not consistently propagated from market prediction to operational decision-making. This paper proposes a unified dual-stage framework in which the LLM functions as a shared semantic information processor, converting raw event data into structured representations used by both forecasting and control modules. In the forecasting stage, these representations improve price prediction under non-stationary conditions. In the control stage, the same information provides an event-aware contextual action prior for reinforcement learning-based energy management. This design allows external event information to inform both future-state estimation and subsequent control decisions, establishing a consistent connection between prediction and decision-making. The framework is evaluated using real-world electricity market data and a battery energy management environment. The results show that the proposed framework achieves the highest average cumulative reward among the evaluated methods while maintaining greater robustness than the forecasting-only LLM configuration. Overall, this work demonstrates the benefit of consistently propagating structured semantic information across forecasting and control and provides a viable approach to event-aware intelligent energy management, with potential extensions to multi-energy transportation systems and electrified mobility applications.

The increasing penetration of uncertain wind power in microgrids challenges scheduling, security, and reliability. To enhance forecast performance, this paper proposes a new method utilizing a large language model (LLM) with parameter-efficient fine-tuning (PEFT) technique, specifically Low-Rank Adaptation (LoRA) for wind power forecasting. Leveraging LLMs’ superior pattern recognition and contextual understanding capabilities, the historical wind power data, meteorological inputs, and relevant context are used in text-based prompts via a supervised instruction fine-tuning strategy. This enables the model to capture complex dependencies and predict future output. Building on the LLM forecasts, a data-driven, risk-tunable uncertainty set based on the Wasserstein distance is developed to characterize forecast errors. This set incorporates the operator’s risk preference. It feeds a robust optimization framework for microgrid scheduling, enabling optimization at varying risk levels. The prompt-optimized LLM forecasting model achieves a 37.99% reduction in mean squared error versus traditional LSTM on the target dataset. Furthermore, the derived risk-tunable uncertainty set enables a robust optimization framework, validated on a modified IEEE 39-bus system, that reduces operational cost by about 5.53% compared to the traditional robust method and improves computational efficiency by about 94.96% versus stochastic optimization, maintaining an effective conservatism-efficiency balance.

Green hydrogen produced from renewable energy is a promising pathway to decarbonize hydrogen fuel cell vehicle transportation. However, existing systems face three interrelated challenges: reliance on single-site renewable generation limits supply stability, hydrogen pricing mechanisms often neglect the carbon intensity of production, and refueling strategies rarely integrate dynamic pricing with congestion-aware routing. To address these gaps, this paper proposes a multi-park hydrogen energy system that coordinates hydrogen production across multiple geographically distributed renewable energy parks via a combined hydrogen production station, exploiting spatial complementarity and economies of scale. A Stackelberg game-based carbon intensity pricing mechanism is developed for inter-regional hydrogen trading between the combined hydrogen production station and decentralized hydrogen refueling stations, dynamically linking hydrogen prices to the carbon emission intensity of electricity used in production. Furthermore, a load-responsive dual-path navigation strategy is designed to jointly optimize HFCV refueling prices and routing decisions, alleviating congestion and reducing user costs. Simulation results demonstrate that the proposed framework reduces peak refueling demand by 5.10 %, lowers average refueling costs by 0.37 %, and significantly improves spatiotemporal load balance, thereby enhancing the economic efficiency, environmental sustainability, and operational reliability of hydrogen mobility systems.

Variance reduction is indispensable in Byzantine-resilient decentralized stochastic optimization over multi-agent systems (MASs) for its ability to mitigate gradient noise and thereby enhance the resilient aggregation process. However, most existing Byzantine-resilient decentralized variance-reduced (VR) stochastic gradient algorithms rely on random data sampling, which proves inefficient in data-scarce yet high-dimensional tasks, for instance, image deblurring. This paper pursues an alternative technical line that achieves variance reduction via gradient sketching. To this end, we first formulate a class of structural risk minimization (SRM) problems, where the local objectives are not necessarily decomposable and their gradients may be unavailable. To solve the SRM problems in a decentralized manner, we integrate a gradient-sketching technique into decentralized stochastic proximal gradient descent with gossip communication to propose a decentralized VR stochastic gradient algorithm, dubbed Gossip-SEGA.Since Gossip-SEGA does not provide any resilience against Byzantine attacks, a resilient extension of Gossip-SEGA,namely RED-SEGA,is developed via replacing the weighted average in Gossip-SEGA by a norm-penalized approximation. Theoretically, we derive sufficient conditions for both consensus (among reliable agents) and linear convergence rate of RED-SEGA over time-varying networks. The effectiveness and resilience of the proposed algorithms are validated through numerical experiments.

Islands, due to their geographical characteristics, face unique challenges in energy production, distribution, and storage, particularly with respect to integrating renewable energy and reducing reliance on fossil fuels. Hydrogen, in this context, is gaining attention as a potential energy carrier capable of facilitating renewable integration, offering energy storage solutions, and aiding in the reduction of emissions in maritime transport. This paper proposes a two-stage sequential optimization model for the scheduling of Hydrogen Carrier Vessel (HCV) and Island Hybrid Energy System Network. The objective is to efficiently coordinate energy production, storage, and hydrogen distribution while minimizing operational costs under uncertainty in renewable energy generation and demand. The first stage focuses on scheduling energy systems. A scenario-based stochastic optimization approach is applied to determine optimal schedules of shore-side and island hybrid energy systems. In the second stage, based on the first stage's results, the optimal dispatch and routing of HCV are determined. The HCV scheduling ensures all islands receive hydrogen within their specified time windows while minimizing transportation costs. A 24-hour case study demonstrates the effectiveness of the proposed model, the proposed optimization reduced the required peak hydrogen reserves by 11.58% compared with a conservative uncoordinated baseline, while avoiding the reliability violations observed under deterministic scheduling. This research provides a viable framework for incorporating green hydrogen production, storage, and distribution into island energy systems, supporting decarbonization initiatives in maritime transport and island communities.

This paper investigates distributed low-carbon scheduling for interconnected multi-microgrid integrated energy systems. An electricity-heat coupled carbon-aware model is established to characterize renewable utilization, multi-energy conversion, storage dynamics, inter-microgrid energy exchanges, and carbon trading. The coordinated scheduling problem is reformulated into a distributed optimization framework with local operational constraints and exchange consistency conditions. A distributed electricity-heat coupled coordination algorithm is developed to enable scalable and privacy-preserving decisionmaking via neighbor-to-neighbor information exchange. Convergence is guaranteed under standard assumptions. Numerical results indicate that the proposed approach enhances multi-energy coordination, achieves effective carbon-emission control under emission constraints, and attains stable distributed convergence.

Accurate electricity price short-term forecasting plays an essential role in the digitization of the electricity market. However, due to the expansion of renewable energy resources and the development of electricity demands, electricity prices are increasingly volatile and difficult to predict, posing a significant threat to the security of daily electricity market operations. The uncertainty of the supply-demand balance, the spatiotemporal correlation of the electricity market are two major obstacles to making the forecasting precisely. In this paper, a multi-task learning model (MGAAL) utilizes a graph attention mechanism and incorporates an auxiliary task focused on predicting abnormal price spikes, enhancing generalization and reducing overfitting risk. Specifically, MGAAL employs attention-based Graph Neural Networks to enhance price forecasting by capturing temporal and spatial power flow dynamics. In addition, MGAAL can also adaptively assign task weights based on homoscedasticity uncertainty and gradient normalization of the tasks. Finally, our experiments, conducted using data from Australia's National Electricity Market (NEM), demonstrate the effectiveness of MGAAL, surpassing current state-of-the-art methods.

To address the strong coupling between schema linking and structural parsing in text-to-SQL tasks for small-scale language models, as well as the neglect of SQL skeleton guidance in existing decoupling methods, this paper proposes SKT-SQL, a multi-stage decoupling framework. The framework redesigns the generation process through a three-stage decoupling mechanism: (1) leveraging a schema decoupler to eliminate irrelevant schema items and reduce semantic noise; (2) predicting query hardness to generate an abstract SQL skeleton, forming a structured template with operator logic; (3) utilizing the skeleton as dynamic prompts to guide a transformer-based seq2seq T5 model in precisely filling specific schema items, followed by execution-guided beam search to derive the final SQL query. This "structure-first, entity-later" paradigm eliminates the need to simultaneously resolve syntactic complexity and schema correlations, significantly reducing cognitive load. Experiments on the Spider 1.0 benchmark show that SKT-SQL-base achieves 80.8 × larger T5-3B model by 6.4 https://github.com/JarvenYi/SKT-SQL.