To address the insufficient regulation capability and bidirectional stochastic fluctuations in power systems integrated with high-penetration renewable energy generation, this paper proposes a multi-time scale rolling optimization and distributed scheduling strategy for a Virtual Power Plant with Proactive Support (VPP-PS). First, a cascaded scheduling framework of “day-ahead global optimization - intraday rolling correction -real-time feedback restoration” is constructed. Box uncertainty sets with budget parameters are utilized to quantify the prediction deviations of heterogeneous resources across various time scales. Secondly, a hierarchical distributed collaborative algorithm based on the Adaptive Alternating Direction Method of Multipliers (ADMM) is designed to achieve the optimal spatial allocation of support energy and regulation margins while ensuring resource privacy. Specifically, for load-side resource-dominant scenarios with limited regulation potential, a State of Charge (SOC) benchmark anchoring mechanism based on state feedback and a real-time surplus mutual aid strategy are developed. Simulation results demonstrate that the proposed strategy could effectively suppress state drift and significantly reduce real-time imbalance penalty costs, providing a workable solution for VPP participation in electricity market operation.

The increasing penetration of line-commutated converter based high-voltage direct current (LCC-HVDC) systems can diminish power grid strength, potentially precipitating static voltage stability issues. To maintain the required level of grid strength while minimizing the total system cost, this paper proposes a coordinated expansion planning framework for multi-infeed LCC-HVDC (MI-HVDC) systems that explicitly incorporates generalized short-circuit ratio (gSCR) constraints. Specifically, the gSCR requirement is innovatively formulated as a semidefinite constraint, and the proposed framework is reformulated as a mixed-integer semidefinite program (MISDP) model to guarantee the global optimality. Then, to address the computational intractability of the MISDP model, a generalized Benders decomposition (GBD)-based approach is employed for efficient solution. Time-domain simulations conducted in PSCAD/EMTDC demonstrate the effectiveness of the proposed framework in enhancing the voltage stability of MI-HVDC systems.

With the rapid development of renewable energy technologies, numerous distributed energy resources (DERs) have been integrated into power systems. How to fully exploit renewable energy while maintaining the stable operation of power systems remains an urgent challenge. Furthermore, the diversity of DERs’ ownership requires scheduling approaches that account for the distinct interests and characteristics of multiple stakeholders. To address these challenges, this study introduces a two-stage operational optimization framework for the virtual power plant (VPP), which is grounded in a Stackelberg game model. This strategy innovatively combines two conventional control methods: the day-ahead stage employs direct control for global pre-scheduling, leveraging its cost optimization capability; the intraday stage utilizes dynamic pricing to guide prosumers, tapping into DERs’ flexibility while accommodating their individual energy usage preferences. The Stackelberg game is resolved through a tiered solution methodology employing particle swarm optimization (PSO). To enhance solution efficiency, a Kriging surrogate model is introduced to replace the prosumers’ models, significantly reducing the computational burden of the PSO. Case studies demonstrate that the proposed strategy can balance operating costs and energy usage preferences, and the proposed solution approach can significantly enhance solution efficiency.

ObjectiveRecent advances in large language models (LLMs) have demonstrated breakthroughs in semantic understanding and reasoning-based generation, providing new technological pathways for multi-source information integration, complex scenario decision-making, and cross-agent coordination. However, a systematic understanding of the mechanisms and application frameworks of LLMs in climate-adaptive power systems remains lacking in both academia and industry. Against this background, the application potential and core issues of LLMs in enhancing the climate adaptability of power systems are systematically reviewed.MethodsFirst, starting from the multi-scale impacts of climate change on power systems, the key capability constraints of power systems in information integration, decision generation, and coordinated interaction are analyzed. Second, a unified analytical framework of "information understanding–decision generation–coordinated interaction" is established, and the typical methodological paradigms of LLMs in semantic understanding, reasoning-based generation, and multi-agent coordination are systematically summarized. Furthermore, key challenges of LLMs in terms of output reliability, consistency with physical constraints, real-time responsiveness, data security, and adaptability to non-stationary environments are analyzed based on engineering application requirements. Finally, future research directions and implementation pathways for LLM-empowered power systems are envisioned to address system evolution requirements under climate uncertainty.ConclusionFacing the challenges of climate change, the deep empowerment of LLMs will drive power systems from passive response toward proactive adaptation, advancing toward a new paradigm of coordinated development characterized by security, low-carbon development, and high resilience.

In order to enhance the operation economics and reduce the operation risks of large-scale integrated electric heat system (IEHS), as well as to generate the planning scheme with maximum benefits for each unit of investment, a cooperative planning method for IEHS considering multi-dimensional operation risks and benefit/cost ratio (BCR) is proposed. First, a district heat system (DHS) segmented virtual heat storage model considering regulation capacity quantification is developed, which can accurately quantify the electric power regulation capacity of DHS while considering heat storage characteristics. Then, a cooperative planning model for IEHS considering multi-dimensional operation risks and BCR is constructed, where BCR indicators considering the benefits of reducing operation cost and multi-dimensional risks including adequacy, flexibility and security are designed. Next, for the planning model containing fractional objective caused by BCR indicators, a solution method based on McCormick envelope and dynamic step sequential bound tightening is proposed for effectively enhancing the solution accuracy. Finally, an IEHS containing a 12-node 500kV transmission power system and two 10-node DHSs is used for case studies. Simulation results indicate that the proposed method outperforms other existing methods in terms of increasing BCR indicators of planning scheme, utilizing heat storage characteristics and electric power regulation capacity of DHS, and enhancing solution accuracy. Compared with the state before planning, the proposed method reduces the adequacy, flexibility, and security risks of IEHS by 72.68%, 62.11%, 49.99% respectively while reducing the operation cost by 12.39%.

Urban energy systems face critical challenges in renewable integration and resilience. This paper introduces OCTOPUS, a bio-inspired framework mimicking octopus distributed intelligence for urban energy management. The system integrates Tentacle Energy Networks, Distributed Cognitive Nodes, and a Central Intelligence Hub. Simulations across 750,000 residents demonstrate 52

Extreme weather events have become increasingly frequent in recent years, leading to widespread outages and substantial economic losses in power systems. While existing studies primarily focus on failure prevention, emergency response, and post-event recovery coordination, limited attention has been given to real-time optimal scheduling that incorporates probabilistic failure risk and resilience-oriented economic measures such as insurance. This paper proposes a data-driven hierarchical dispatch framework designed to enhance system resilience while reducing overall operational costs and load losses. First, an LSTM-Bayesian hybrid model is developed to estimate component failure probabilities under ice- and snow-related meteorological conditions. To accurately localize and quantify potential faults, a dynamic line vulnerability index is further introduced, capturing temporal variations in weather-induced stress. Building on these risk assessments, a resilience insurance purchasing strategy is formulated to mitigate the financial impact of extreme-weather-induced outages. The proposed framework is validated on the different node systems. Simulation results demonstrate that the coordinated dispatch strategy, integrating both probabilistic fault modeling and resilience insurance, significantly enhances system performance: load shedding is reduced by approximately 62.63% during fault conditions, dispatch costs are lowered, and insurance payouts decrease by 15% compared with an uninsured scenario. These findings highlight the effectiveness of combining data-driven risk analysis with resilience economics to support robust and sustainable grid operation under extreme weather events.

As the climate crisis worsens, carbon reduction becomes a global focus. Hydrogen fuel cell vehicles (HFCVs), as a low-carbon alternative to traditional vehicles, have developed rapidly in recent years, promoting the coupling between the electricity, transportation and hydrogen sectors. However, the hydrogen supply system remains underdeveloped, which poses a significant barrier to the widespread adoption of HFCVs. Therefore, this paper proposes a three-stage low-carbon planning framework to optimize the deployment of hydrogen supply infrastructures, with special consideration given to operational carbon emission constraints. The framework comprises three systematic stages: infrastructure planning, electricity-hydrogen-transportation system cooperation, and cross-sectoral carbon emission analysis. Especially, in the hydrogen delivery process, a hybrid delivery method is adopted, which incorporates both hydrogen pipelines and hydrogen tube trailers (HTTs). Furthermore, for operational carbon emission analysis, a cross-sectoral carbon emission flow (CSCEF) model is proposed to trace the carbon footprint throughout the electricity-hydrogen-transportation system. Then, a no-good cut based solution method is used for solving the proposed three-stage low-carbon planning model. Validated by case studies, the proposed framework significantly enhances the hydrogen supply efficiency by the hybrid delivery method, and reduces 81.8% of operational carbon emissions after low-carbon planning.

Climate change is progressively reshaping the spatiotemporal dynamics of renewable energy sources such as wind and solar, intensifying the complexity and uncertainty of long-term power system planning. Existing planning frameworks are largely focused on climate mitigation strategies but often overlook the critical dimension of climate adaptation, limiting their efficacy in managing evolving climatic risks. In response, this article proposes a multistage stochastic low-carbon planning framework that incorporates climate-related uncertainties into system planning decision-making. By embedding climate evolution trajectories into the planning horizon, the proposed approach determines optimal stage-wise planning pathways that jointly accommodate mitigation goals and adaptation imperatives under long-term climate uncertainties. First, a systematic climate uncertainty modeling approach is developed to capture both scenario uncertainty and climate response uncertainty through the construction of a representative scenario tree. Second, to reconcile the temporal mismatch between coarse-resolution climate projections and the finegrained requirements of power system planning, a climate-consistent temporal downscaling method is proposed to transform long-term climate projections into high-resolution, hourly level data. Third, to address the computational complexity inherent in the multistage planning problem, a tailored decomposition-based stochastic dual dynamic programming algorithm is developed, which operates on a stage-wise clustered scenario tree to leverage the tree’s structural compactness for accelerated convergence and scalable optimization under climate-related uncertainties. Numerical studies demonstrate that the proposed climate-adaptive planning framework enhances the power system’s ability to manage climate-induced risks while maintaining cost-effectiveness across a wide range of plausible climate futures.

The home energy management system (HEMS) has gained significant attention with the advancement of smart monitoring and Internet of Things technologies. Data-driven approaches, particularly reinforcement learning (RL), have shown promise in learning HEMS scheduling policies through environment interaction, but often suffer from poor sample efficiency and high computational overhead in environments with delayed rewards. To address these challenges, this paper proposes a hybrid framework integrating large language models (LLMs) and RL for efficient and safe HEMS scheduling. Unlike existing LLM-based energy studies that mainly use LLMs for forecasting, knowledge processing, or open-loop advisory support, the proposed framework deploys an LLM as a direct decision-making actor in a closed-loop HEMS control setting. By leveraging the LLM’s pre-trained reasoning capability as an implicit policy prior, the proposed method improves the sample efficiency of RL-based HEMS scheduling without requiring expert demonstrations. PPO-based fine-tuning with low-rank adaptation (LoRA) further enables efficient LLM–HEMS alignment by updating only approximately 0.1% of the model parameters. Moreover, a constrained action-sampling mechanism restricts LLM outputs to predefined admissible appliance actions, thus ensuring physically executable, constraint-compliant scheduling decisions. Experiments driven by real-world market and appliance data show that the proposed method improves home utility by 33.3%, reduces user discomfort costs by 28.6%, and requires over 20 times fewer training episodes than conventional RL baselines.

The rate of change of frequency (RoCoF) is a critical metric of power system frequency security. Post contingency nodal RoCoFs are not identical across a power system, with spatial characteristic significantly related to the fault size/location, network connections and parameters, as well as the system's inertia distribution state. However, most traditional security-constrained unit commitment (SCUC) strategies typically rely on aggregated frequency models, lacking necessary granularity to differentiate disparities among nodal RoCoFs. Consequently, certain generators may suffer from extremely large RoCoFs, potentially triggering unexpected pole slip protection and trip operations, even if the system's RoCoF estimated by an aggregated model remains within the allowable range. To address this issue, a nodal RoCoF security constrained unit commitment is proposed considering N-1 contingencies of any single generator trip in this letter. First, unified boundary conditions of generators are established to depict their disparate inertia response behaviors in online/offline states. Then, nodal RoCoF security constraints are further derived and embedded into an optimization problem for SCUC scheduling. Following the proposed SCUC strategy, nodal RoCoF security can always be guaranteed in any single generator trip contingencies, which is further verified by simulation results of the South East Australia Power System in different scenarios.

To mitigate the conservatism of scheduling schemes derived from the two-stage robust unit commitment model (TS-RUCM), this paper proposes a novel multi-state uncertainty set (MSUS) considering the uncertain wind power output (WPO). The MSUS formulates WPO uncertainties with higher resolution by introducing multiple discrete state values within the uncertain interval. Additionally, the MSUS limits extreme fluctuations of WPO uncertainties among state values through transition indicators embedded in multiple transition constraints. The state values and transition indicators are calculated using the conditional quantile regression technique and Markov chain analysis. This systematic and scientific approach allows the MSUS to effectively exclude low-probability WPO scenarios, thereby significantly mitigating the conservatism of the scheduling schemes. Moreover, to accelerate the computation of the TS-RUCM based on the MSUS, this paper proposes an improved inexact column and constraint generation (II-C&CG) method. The II-C&CG method employs an adaptive tolerance strategy and refined backtracking to solve master and subproblems inexactly while ensuring finite convergence, significantly reducing computational time. Case studies demonstrate the effectiveness and advantages of both the MSUS and the II-C&CG method.

The deep coupling between physical and cyber power systems exposes transmission systems to greater risks during extreme events, particularly under physical deliberate attacks (PDAs) characterized by spatial, temporal, and quantitative uncertainties. To ensure reliable load supply under such attacks, this paper proposes a resilience enhancement strategy for cyber physical power systems against PDAs with spatial-temporal quantitative uncertainties. First, a tri-level resilience enhancement framework for transmission systems considering cyber physical interdependence is proposed, simulating sequential interactions among the defense system operator (DSO), the attacker, and the transmission system operator (TSO) in cyber-physical coupled environments. Second, a two-stage multi-uncertainty DAD model is constructed to determine the DSO's optimal line hardening schemes by evaluating the attacker's most destructive targets and timing under different attack capabilities, along with the TSO's coordinated optimization of emergency dispatch and line repair scheduling. Third, an improved nested column-and-constraint generation algorithm is developed to efficiently solve this model. Finally, case studies on a modified IEEE RTS-79 system validate the effectiveness of the proposed strategy and demonstrate the advantages over both attacker-defender and traditional DAD models, while simulations on a modified IEEE 118-bus system confirm its scalability.

Modern agriculture is under increasing pressure to decarbonize, and replacing diesel tractors with their electric counterparts has gained traction as a promising pathway. The wider adoption of electric tractors faces the challenge of lacking infrastructure solutions for high-power electric tractors operating across vast, seasonal farmlands. The battery swapping stations equipped with standardized and “Lego-like” battery cells meet the high-volume charging demands and reduce downtime compared to conventional charging. To support the future infrastructure requirements of heavy-duty electric tractors, this paper addresses a critical gap in agricultural electrification by proposing a location selection and capacity configuration model for battery swapping stations and demonstrates its feasibility in a real-world scenario. By analyzing the spatiotemporal distribution pattern of the charging demand across multi-regional farmlands under a fully electrified scenario, the optimal deployment location for installing the battery swapping station is determined. Then, the capacity configuration model optimizes the number of charging bays and battery cells with the objective of minimizing the annualized investment expenditure. The case study validates the effectiveness of the deployment location selection and capacity configuration models.

The increasing penetration of distributed photovoltaics (PVs) and stochastic loads aggravates voltage violations and feeder overloading in flexible interconnected distribution networks (FIDNs). To address these issues, this paper proposes a novel flexibility-oriented distributed cooperative control framework to achieve fast voltage regulation and load balancing with low communication overhead. A hierarchical control architecture is established for FIDNs with distributed generators and soft open points (SOPs), and flexibility adequacy indices are constructed and used as event-triggering thresholds. At the intra-subsystem level, distributed linear-quadratic optimal controllers with observers restore node voltages and equalize the utilization of PV inverter reactive power capability. At the in ter-subsystem level, an event-triggered multi-objective dynamic consensus algorithm is employed to coordinate transformer loading ratios, and a distributed coordination controller is designed to issue active and reactive power commands to SOPs. Case studies on a four-terminal FIDN and a practical 87-node FIDN demonstrate that the proposed method effectively mitigates voltage violations and load imbalance, enhances renewable energy hosting capacity, and improves system resilience against severe contingencies.

Stable interactions among electricity, carbon allowance, and fossil fuel markets are essential for sustainable energy transition, because excessive cross-market risk transmission may affect energy affordability, carbon-price credibility, and low-carbon investment signals. This study provides comparative evidence on dynamic connectedness, tail-state shock responses, and return-based complexity in representative Chinese and European benchmark markets. Using daily market data from the Wind database for November 2021-January 2026, the empirical framework combines time-varying parameter vector autoregression (TVP-VAR), quantile vector autoregression and quantile impulse response functions (QVAR/QIRFs), and rolling multifractal detrended fluctuation analysis (MFDFA). The results show that the European benchmark system has a higher absolute connectedness level than the Chinese benchmark system: the full-sample mean total connectedness index (TCI) is 18.75 in Europe and 5.63 in China, while the crisis-period mean TCIs are 25.19 and 12.12, respectively. Post-peak adjustment depends on the reversion metric used: China shows a faster initial half-life decline from the crisis peak, whereas reversion to lower region-specific connectedness thresholds depends on the selected benchmark. Natural-gas-shock QIRFs indicate stronger upper-tail persistence in Europe, whereas China is characterized mainly by short-run directional divergence; supplementary coal-, oil-, and carbon-shock checks show that response patterns are shock-source-dependent. Electricity-return multifractal spectrum width (MFW) does not show stable full-sample explanatory power for TCI, but it provides stage-dependent auxiliary diagnostic information. These findings provide a comparative diagnostic framework for monitoring cross-market systemic risk and supporting sustainability-oriented energy-market governance under low-carbon transition.

Demand response (DR) based on customer directrix load (CDL) is a new incentive-based DR scheme, typically invited by regional utility grid companies. In this context, a framework is proposed for a prosumer community to participate in CDL-based DR, comprising two key steps: power optimization to balance electricity costs and DR benefits for the overall prosumer community, and cost allocation for prosumers within the community through two hypothetical sub-steps, i.e., cooperation without and with DR. In the power optimization step, the Convex-Concave Procedure (CCP) algorithm is used to relax the non-convex parts of the optimization objective to be linear ones, gradually approximating the optimal solution by iteratively correcting the upper bound function. In the cost allocation step, the Owen value method is applied, which accounts for priority coalitions and eliminates the unrealistic coalition possibilities that are inherent in the calculation of Shapley value. Comprehensive simulations of a test case validate the rationality of the CDL-based DR framework and the effectiveness of the iterative method.

With the global energy transition towards more decentralized and sustainable systems, peer-to-peer (P2P) energy trading has emerged as a prominent energy exchange model, at tracting increasing attention from academic and industry fields. P2P energy trading not only facilitates the efficient utilization of distributed energy resources (DERs) but also enables autonomous energy transactions. This paper provides a comprehensive review on P2P energy trading, starting with an explanation of the P2P energy network, P2P market structures, and block chain in P2P energy trading. The main body reviews existing research from six key research domains: trading platform, market clearing method, security, uncertainty, modeling of users' behavior, and experimental validation. Subsequently, the paper discusses key factors in deployment of P2P energy trading systems. Finally, opportunities and challenges in the future are out lined, and conclusions are drawn.

The integration of artificial intelligence (AI) is fundamentally reshaping energy systems, offering unprecedented capabilities for renewable integration and autonomous operation. However, this transition introduces complex attack surfaces spanning data, model, system, and physical layers. This paper addresses this critical security paradox by reviewing the inherent vulnerabilities of frontier AI paradigms, including deep learning, large language models (LLM), and reinforcement learning (RL) within the energy context. Building on this, we propose a four-layer threat model that goes beyond established risks to analyze how modern paradigms create novel attack vectors. Specifically, we examine how LLMs enable prompt injection and tool abuse that can manipulate physical processes, how RL agents face reward hacking, and how complex data ecosystems are susceptible to sophisticated poisoning attacks targeting foundational knowledge bases. This paper synthesizes cross-disciplinary evidence and provides a structured threat-modeling and defense-in-depth blueprint applicable to both conventional energy AI pipelines (e.g., forecasting and optimization) and emerging LLM-based operational assistants. To counter these evolving threats, we establish a closed-loop defense-in-depth framework that organizes mitigation measures into four pillars: technical protection, data privacy, system architecture, and governance. We map specific defense technologies to corresponding threat layers to provide a practical protection strategy. Looking forward, we envision pathways toward next-generation trustworthy energy AI, highlighting frontier research directions such as online adaptive defense, verifiable AI, and hardwareenhanced confidential computing. We specifically propose the development of red-blue adversarial training and deception defense systems. Ultimately, this paper advocates for deep interdisciplinary collaboration to ensure AI serves as a reliable driver for decarbonization rather than a security liability.

Urban energy ecosystems require real-time adaptability, resilience, and decentralized intelligence to manage fluctuating demand and renewable integration. This paper introduces FALCON (Flight-Adaptive Learning and Cooperative Optimization Network), a bio-inspired optimization framework modeled after the aerial precision, adaptability, and cooperative hunting strategies of falcons. The FALCON architecture integrates Flight-Adaptive Neural Controllers (FANCs) and Cooperative Energy Clusters (CECs) that emulate falcon-inspired flight dynamics for fast convergence and predictive decision-making. A dynamic Trajectory-Based Optimization Function (TBOF) enables agents to anticipate energy flow fluctuations, perform distributed reinforcement learning, and cooperatively reallocate resources across urban nodes. Extensive simulations across five metropolitan districts demonstrate that FALCON achieves a 57.8