The proliferation of distributed energy resources and multi-energy coupling technologies has transformed virtual power plants into complex MEVPP clusters. However, coordinating these heterogeneous clusters poses multifaceted challenges, particularly due to the deep stochastic uncertainties of renewables, the complex coupling of energy vectors with distinct dynamic inertias, and the severe computational bottlenecks inherent in traditional centralized control across multiple timescales. This paper proposes a novel hierarchical control strategy that leverages cloud-edge collaborative architecture and digital twin technology to optimize MEVPP clusters across electricity, heat, and gas networks. The framework employs a three-layer temporal decomposition: cloud-based day-ahead scheduling using mixed-integer linear programming for 24-h economic dispatch; edge-based intra-day rolling optimization with 15-min intervals for uncertainty mitigation; and real-time DMPC with 5-min resolution for dynamic balancing. A comprehensive digital twin framework integrates physics-based multi-energy flow models with data-driven techniques to enhance state estimation and prediction accuracy while maintaining computational efficiency for edge deployment. The DMPC algorithm coordinates multiple MEVPPs via decomposition and coordination, managing cross-coupling constraints and optimizing power distribution via tie-line sharing. Case studies on three heterogeneous MEVPPs demonstrate that the proposed strategy achieves a 1.07% reduction in total system costs through coordinated operation compared to independent optimization, with individual MEVPP cost reductions ranging from 8.77% to 28.84% during intra-day operation under renewable forecast uncertainties, while maintaining system stability and inter-MEVPP power exchange balance.

Research on low-carbon economic operation of park-level integrated energy systems (PIESs) has emerged as a significant focus in sustainable energy management. Effectively leveraging user-side flexible resources is critical to maximizing the system’s energy saving and emission reduction. However, users’ willingness to engage in demand response is influenced by their comfort, economic, and low-carbon preferences. This article proposes a bilevel low-carbon operation strategy for PIESs incorporating such heterogeneous user preferences. Firstly, energy and carbon flows within a PIES are traced based on an energy hub structure to derive the nodal carbon intensity. Subsequently, a bilevel low-carbon economic operation model is formulated considering diverse users’ interaction. The upper-level PIES operator incentivizes users by providing dynamic carbon intensity and prices, while lower-level users adjust consumption based on economic, comfort, and low-carbon objectives. This interaction enables end users to intuitively perceive the carbon emissions embedded in their energy consumption. Users comprehensively evaluate objectives through analytic hierarchy process, adjusting flexible loads without compromising personal preferences. Case studies demonstrate that the proposed strategy can effectively harness the low-carbon demand response potential of users with different consumption preferences and further reduce the system’s carbon emissions.

ABSTRACT The combination of electric‐heating interconnected microgrids (EHIMs) with grid‐forming converters (GFMCs) offers a highly promising approach to improving energy management within decentralised energy systems. Despite its potential, precisely modelling and analysing energy flow in such systems remains a considerable challenge due to the variety of energy sources, storage technologies and control strategies involved. This paper presents a novel methodology for the unified calculation of energy flow within EHIMs, utilising heterogeneous graph convolutional networks (HGCNs). In this framework, the microgrid components are represented as a heterogeneous graph, where nodes correspond to various system elements, such as photovoltaic panels, wind generators, solar thermal systems and energy storage units, whereas the edges represent the energy interactions and exchanges between these elements. By adopting this graph‐based structure, the model is better equipped to handle the variability of renewable energy sources and the adaptable nature of grid‐forming converters. Extensive simulation results show that the HGCN‐based approach surpasses traditional methods in terms of both prediction precision and computational efficiency, providing a more scalable solution for analysing real‐time energy flows in EHIMs. Ultimately, the proposed framework holds great promise for significantly improving the stability, operational efficiency and resilience of future smart grid infrastructures.

The wide-area coordination of Multi-Energy Microgrid (MEMG) clusters relies heavily on open communication networks. Consequently, the Cyber-Physical Systems within these clusters face severe challenges due to non-ideal communication conditions, including latency, packet loss, and noise. These impairments often compromise the convergence of traditional distributed strategies and deteriorate economic performance. To address these impediments, this paper proposes a Digital Twin (DT)-perceived cloud-edge collaborative resilient scheduling strategy for MEMG clusters. First, a communication topology adaptive reconfiguration mechanism based on Pareto optimality is established to enhance the system's structural resilience against cyber-attacks and network failures. Second, a hybrid-driven DT model, integrating a “mechanism model + residual correction network,” is constructed to achieve high-fidelity state prediction under communication constraints. These predictions serve as the initial values for the distributed consensus algorithm, thereby minimizing initial errors and mitigating aggregation bias and iterative oscillations induced by information truncation. Finally, a cloud-edge spatiotemporal decoupling mechanism is designed, employing tiered carbon pricing to guide optimal synergistic interactions between electricity and carbon across clusters. Case studies demonstrate that the proposed strategy maintains rapid and stable convergence even under severe latency and dynamic topology disturbances. Furthermore, it significantly enhances operational resilience while effectively reducing the total operating costs and carbon emissions of the clusters by up to 89.9% and 35%, respectively.

With the increasing penetration of renewable energy, power systems are facing greater uncertainty and volatility, which poses significant challenges for Virtual Power Plant scheduling. Existing research mainly focuses on optimizing economic efficiency but often overlooks system reliability and the impact of forecasting deviations on scheduling, leading to suboptimal performance. Thus, this paper presents a reliability-cost bi-objective cooperative optimization model based on a dual-swarm particle swarm algorithm: it introduces positive and negative imbalance price penalty factors to explicitly describe the economic costs of forecast deviations, constructs a reliability evaluation system covering PV, EVs, air-conditioning loads, electrolytic aluminum loads, and energy storage, and solves the multi-objective model via algorithm design of “sub-swarms specializing in single objectives + periodic information exchange”. Simulation results show that the method ensures stable intraday operation of VPPs, achieving 6.8% total cost reduction, 12.5% system reliability improvement, and 14.8% power deviation reduction, verifying its practical value and application prospects.

Multiple energy flow calculation (MEFC) serves as a fundamental technology within Integrated Energy Systems (IESs). Nevertheless, MEFC has been confronted with increasing challenges due to the complicated couplings of heterogeneous energies, information constraints, and extreme scenarios. This study proposes a data-driven approach for the MEFC task based on an edge-enhanced heterogeneous graph attention mechanism (EEHGAT). A heterogeneous graph structure is first employed to model the electricity — gas — heat coupled IES as three types of nodes and four types of edges. Then a dual — model architecture composed of a node model and an edge model is constructed. This architecture aims to fully map the MEFC information and simultaneously prevent error accumulation. Subsequently, a multi — task learning framework incorporating a variable — weight loss function is integrated to enhance the flexibility of the training process. Moreover, visualized interpretability is developed to demystify the 'black — box' nature of the neural network. Numerical results demonstrate that the Mean Absolute Percentage Error (MAPE) of the proposed node model and edge model can reach 0.1012% and 0.2060%, respectively. Meanwhile, under conditions such as measurement errors, information limitations, and extreme scenarios, the proposed method exhibits greater adaptability and robustness compared to the Newton — Raphson method and traditional data — driven methods.

Multi-energy virtual power plant (MEVPP) faces significant challenges stemming from the inherent intermittency of renewable energy, the underutilized flexibility of distributed load-side resources, and the intricate complexities associated with optimizing multi-energy coupling and dynamic carbon emission management. This paper proposes a multi-timescale digital twin rolling optimization strategy incorporating vehicle-to-grid (V2G) interaction and stepwise dynamic carbon trading to address these issues. First, a digital twin dynamic aggregation modeling approach is developed by combining deep neural networks with an enhanced k-means spectral clustering algorithm. This enables accurate dynamic characterization of wind and solar generation uncertainties and equivalent aggregation of distributed resources, thereby unlocking greater flexibility on the demand side. Second, considering electricity-carbon coupling and the unique operational characteristics of electric vehicle charging and discharging, a real-time pricing strategy based on time-of-use tariffs and incentive/penalty mechanisms is introduced. Simultaneously, a stepwise dynamic carbon trading mechanism is proposed to facilitate more precise measurement of carbon emissions from various devices and enable coordinated optimization of electricity and carbon flows. Finally, a multi-stage rolling optimization model is constructed within a distributed model predictive control framework to mitigate deviations between day-ahead scheduling and intra-day operation caused by prediction errors and changing weather conditions. Multi-scenario comparative analyses demonstrate that the proposed strategy can effectively reduce the peak-to-valley difference of the electric load, lower the total system economic cost, and decrease carbon emissions while maintaining operational cost variance within 1.2%. These findings offer theoretical and practical guidance for the efficient and sustainable operation of MEVPP within next-generation power systems, particularly under evolving market conditions and increasing renewable penetration.

A novel method for localizing wide-frequency oscillation sources by combining Informer-based signal compression-reconstruction processes with Edge Graph Attention Networks (EGAT) to improve localization accuracy is proposed. At the substation level, the Informer network’s advanced temporal modeling and feature extraction capabilities are utilized to efficiently compress and encode the measurement signals from the power system. At the master station, the compressed signal features are integrated with the system’s network topology, and EGAT is employed to achieve accurate oscillation source localization. By adaptively adjusting the weights of the nodes based on edge feature information, EGAT captures inter-node correlations, which enhances the accuracy and robustness of the localization process. This approach proves effective in accurately identifying the locations of oscillation sources within complex power network topologies, even in the presence of noisy measurement data. Experimental results demonstrate that the proposed method maintains high accuracy and timeliness across various power system topologies, even under conditions of noise interference, highlighting its robustness and effectiveness.

This article investigates the low-carbon economic operation problem of commercial park-level integrated energy systems (PIES), with a focus on the refined modeling of diverse flexible resources in both supply and demand sides. In the supply side, the organic Rankine cycle (ORC) is integrated into the combined heat and power (CHP) system to realize flexible outputs, instead of being constrained at a fixed thermoelectric ratio. In the demand side, the building's central air conditioning (CAC) load is modeled by a thermal equilibrium equation, capturing CAC flexibility more accurately. Additionally, given the unique commuting patterns of commercial parks, battery swapping stations (BSS) and electric vehicles are also characterized as flexible demand-side resources. The BSS is modeled as a generalized energy storage device through Minkowski summation to reduce model dimensionality. To address carbon responsibility, a ladder carbon trading mechanism is implemented to constrain the park's carbon emissions. Simulation results show that the proposed model achieves a 22 % cost savings and a 5 % carbon emission reduction compared to traditional operation strategies. This demonstrates that the coordination of supply-demand flexibility significantly improves economic and environmental performance in commercial PIES operation.

In the park-level microgrids with photovoltaic system, uncertainties such as DC side voltage fluctuations and system parameter variations, as well as the presence of nonlinear loads, pose a serious threat to the voltage-quality of the photovoltaic inverter, making it difficult to ensure its stable output. To effectively enhance the robustness and the performance of the photovoltaic micro-inverter in dynamic process, an adaptive fuzzy sliding-mode control (AFSMC) framework is proposed in this paper. An adaptive fuzzy controller is designed to approximate the sliding mode control law under the rated model, and the parameter adaptation rate of the fuzzy controller is also designed. This strategy can significantly alleviate the chattering phenomenon and enhance the robustness against the system uncertainties and nonlinear characteristics in the photovoltaic inverter system. Finally, a simulation model was built in MATLAB/Simulink for verifying the effectiveness of the designed control strategy by conducting simulation experiments. The results derived from the simulations demonstrate that the high - quality voltage under both linear and nonlinear loads can be obtained. In addition, when facing load disturbances, it has strong robustness and superior dynamic performance. The proposed AFSMC strategy significantly improve the robustness and transient performance of the photovoltaic power generation system, which provide a new solution for enhancing the quality of the output by photovoltaic inverters.

In the wind power industry,the algorithm library plays an important role in the development of control systems as a carrier for the application and accumulation of industry knowledge.In order to solve the problem that the wind power algorithm library of foreign brand controllers is not open source and realize the independent development of wind power al-gorithm library,this paper analyzes the specific needs of the wind power industry for the algorithm library,combines the commonly used control system program design standards in the industrial control industry,proposes a design method of wind power algorithm library from the perspective of functional design and interface design,and describes the process of realizing it through IEC61131-3 language and C language.The algorithm library developed by this method has been ap-plied to large-scale wind turbines.

With the continuous expansion of the coverage of the power system, its overall architecture and planning have become increasingly complex. As a crucial component of the operation of the power system, the distribution network has been affected, and the accuracy of fault location has significantly declined. Therefore, this paper proposes a method for fault location in the distribution network based on the improved binary particle swarm optimization algorithm. During the iterative process of binary particles, the particle positions are first adaptively mutated. At the same time, an adaptive strategy is introduced in the setting of the inertia weight to construct a binary particle swarm optimization algorithm with dual adaptive characteristics. The simulation results show that, whether it is a standard radial distribution network or a similar distribution network with distributed power sources, the improved algorithm can accurately locate the fault section. Compared with the traditional binary particle swarm optimization algorithm and the genetic algorithm, the improved algorithm demonstrates more stable convergence performance. It will not fluctuate due to different fault types, and it has extremely high reliability. Therefore, it can better adapt to the complex and changeable distribution network environment and efficiently complete the task of fault location in the power system.

After the proposal of net zero emission policy, virtual power plant has attracted substantial attention due to its capacity to integrate distributed flexible resources. This paper presents a deeper investigation into a multi-energy virtual power plant (MEVPP) with four types of coupling networks (electricity, heat, gas and cold), considering carbon trading mechanism and vehicle-to-grid (V2G) interaction. Here we implemented penalties for carbon emissions and compensations for grid peak-valley difference reduction, transferring unquantifiable environments and grid stability effects into quantifiable economics effects, satisfying the economic, environmental and grid stability objective simultaneously. Then, a two-stage Particle Swarm Optimization (PSO), including day-ahead scheduling based on forecast data and intraday rolling optimization based on real-time data, is applied to address the optimization problem. The case study shows that considering carbon trading mechanism, carbon emissions can be reduced by 9.8%. And V2G interaction can decrease the peak-valley difference by 45.38%.

High-Distributed Energy Resources (DERs) distribution networks challenge real-time digital twins (DTs) since conventional AC power-flow solvers struggle to converge within sub-second budgets. We propose Logic Tensor Network-Digital Twin (LTN-DT), a hybrid framework that couples a semidefinite-relaxed OPF, adaptive Newton refinement and a neuro-symbolic LTN constraint layer. On IEEE 13-, 37-, 123-bus, and a real-world 10 kV feeder, LTN-DT keeps voltage-magnitude errors below 0.5 % (0.28–0.46 % MAPE) and preserves solutions within 1 % of full-AC benchmarks. Newton iterations drop from ten to about three, while the surrogate trims per-iteration cost by 60 %, yielding an overall ⩾5 × solver speed-up latencies on medium feeders (0.415 s on 123-bus).

Non-Intrusive Load Monitoring (NILM) has become a key technique for enhancing residential energy awareness, efficiency, and management. However, many existing NILM approaches aim to disaggregate explicit load profiles based on the smart meter data, neglecting the impact of behind-the-meter Photovoltaic (PV) systems. As most standard smart meters only record the net household load, these methods may not perform well in scenarios where PV generation is present. The absence of dedicated devices to accurately capture PV output further limits the real-world applicability of such methods. To address this, we propose a BERT-based NILM method that utilizes only hourly meter data and weather information to simultaneously disaggregate appliance-level loads and PV export. Compared to other conventional approaches, our method incorporates an embedding module to learn latent variables representing the context of the residence. Experimental results using real-world hourly data from Texas show that our method can reduce the root-mean-square error by 34.6%, lower the normalized disaggregation error and mean-absolute error 43.8%, 38.7% separately, and improve disaggregation accuracy by 2.7% compared to vanilla BERT model.

Power-to-gas technology uses temporary surplus electricity to create either renewable hydrogen or renewable natural gas, which can then be stored in natural-gas pipelines and used when needed. As a result, P2G transforms conventional one-way coupling of a power/heat/natural-gas system into two-way coupling. Furthermore, its operating characteristics makes it possible to more effectively utilize wind-power. This paper describes a new optimal dispatch model for integrated electricity/gas/heat energy systems. The model considers the effective use of surplus wind-energy with electricity-to-gas equipment. First, a multi-energy network model is built, taking into account both equipment and network constraints. Then, we apply a novel two-layer optimization method, which uses P2G, to “absorb” wind power. While the top-layer model is used for the day-ahead dispatch of the natural-gas network containing P2G, the bottom-layer model describes the day-ahead economic dispatch of the electricity/heat system, which includes wind power. Based on the Karush-Kuhn-Tucher conditions of the bottom-layer model, the two-layer model is transformed into a single-layer model, and we linearize the nonlinear equation to convert the nonlinear model into a mix-integer linear programming problem, which is solvable using the General Algebraic Modeling System. Finally, numerical case-studies are performed to evaluate the accuracy and effectiveness of the proposed method.

Energy efficiency and carbon mitigation are important issues in modern energy systems research. Considering the couplings between energy and carbon, this paper investigates a carbon-embedded energy coordination problem in a park-level integrated energy system (PIES). Firstly, based on energy hub, a multilateral interactive transaction framework is introduced, consisting of energy hub operator (EHO), building users with photovoltaics, and an electric vehicle (EV) charging agent. Secondly, a time-varying carbon emission measurement model for the outsourced electricity is designed, taking into account the dynamic composition of the supply-side generators. Based on the time-varying carbon factor and virtual carbon emission flow, an energy-carbon pricing strategy is proposed to balance the carbon responsibility among different participants. Then, a multi-agent interactive trading model in the PIES is constructed, where the energy consumption plans of building users and EVs are guided by the dynamic energy-carbon integrated prices. Through the iterative interaction among different agents, the optimal trading results are finally obtained. Simulation results illustrate the effectiveness of the proposed energy-carbon integrated pricing method in reducing carbon emissions and promoting energy sharing.

Conventional model-driven methods are hard to handle large-scale power flow with multivariate uncertainty, variable topology, and massive real-time repetitive calculations. With the ability to deal with non-Euclidean graph-structured power system data, graph deep learning shows great potential in modern power flow calculation. However, general graph deep learning based power flow calculation has limited adaptability because of its sole mapping of node information and black-box attributes. In this paper, an edge graph attention network based power flow calculation (EGAT-PFC) model is proposed with improved adaptability for power flow analysis of complex system scenarios. First, the dual-model structure of the node model and edge model is constructed to realize a complete power flow mapping covering all information in power systems. Second, an improved learnable attention coefficient mechanism fusing node and edge features is proposed to ensure global information can be completely considered. Third, mechanisms of extended first-order neighborhood, dynamic normalization, and regularization-based loss function are designed to improve training performance. Finally, visualized interpretability is developed to show valuable information of vulnerable nodes and lines of power system operation. The numerical simulation verifies that EGAT-PFC has high accuracy, fast mapping, as well as excellent adaptability to variable topologies.