Abstract:Modern power systems are evolving due to increasing penetration of renewable energy sources, deeper participation of the demand side, and widespread deployment of advanced information and digital technologies. As a result, system operation and control are becoming increasingly challenging. Large language models (LLMs), with their advanced capabilities in semantic understanding and knowledge reasoning, offer a promising tool to support the operation, control, analysis, and decision-making of power systems. This paper provides a comprehensive review of LLM applications in power systems, encompassing four representative application domains: power grid, power equipment, demand side, and electricity market and policy-making. Based on the functional roles and implementations of LLMs, four major application strategies are identified: model adaptation, capability enhancement, multimodality integration, and multi-agent coordination. In addition, the core functions, representative methods, and evolving trends of LLM applications are reviewed across different domains. Finally, key challenges in applying LLMs to power systems are discussed, and future research directions are outlined with regard to ensuring physical feasibility, enhancing data efficiency and privacy, and improving interpretability and rationality.

Abstract:Electrical load forecasting (ELF) plays a critical role in the planning and operation of modern power systems. As energy demand patterns grow more complex, deep learning (DL) techniques, and more recently, foundation models (FMs), have emerged as powerful tools for modeling temporal dynamics and integrating heterogeneous inputs. In practice, the effectiveness of these models depends not only on their architectures but also on the learning paradigms that determine how they are trained, adapted, and deployed. However, most existing surveys focus solely on network architectures, with limited attention to the underlying paradigms. To this end, we survey the DL-based ELF from perspectives of learning paradigms and FMs. It organizes the literature into four orthogonal paradigms: task-tuned offline learning, adaptive DL, collaborative DL, and general-purpose DL. This paradigm-centric perspective enables a unified understanding of how DL methods evolve to meet the challenges of ELF. It also provides a natural framework to incorporate FMs as the latest advancement in this trajectory. Finally, key challenges are provided, and research opportunities are highlighted to inform future directions.

Abstract:High-frequency oscillations (HFOs) in modular multilevel converter based high-voltage direct current (MMC-HVDC) connected to AC networks have aroused concern in recent years. This paper reveals a new HFO phenomenon, where modular multilevel converter (MMC) exhibits impedance characteristics different from those in conventional HFO studies. To investigate this issue, a high-frequency dynamic model of MMC incorporating delay compensation is developed under different outer-loop control strategies. In addition, an AC line modeling method accounting for frequency-dependent parameters is proposed. Mechanism analysis demonstrates that the new HFO phenomenon arises from the adverse interaction between capacitive impedance of MMC and inductive impedance of AC network. Further, the causes of two types of capacitive regions in the high-frequency impedance of MMC are theoretically explained, which are linked to the relatively high proportional gains of current and power controls, as well as delay compensation. The effects of the frequency dependence of AC line parameters on HFO analysis are also demonstrated. Finally, the potential multiple HFOs and possible mitigation approaches are discussed for further study.

Abstract:Distributed generation by converter-interfaced renewable energy sources connected to distribution feeder nodes has been increasingly penetrating power grids, along with their expected contribution to frequency and voltage support services. This will change the perspective with which stability and dynamic behavior of power systems have been analyzed/simulated to date. Until a few years ago, only transmission systems were simulated in detail, while each distribution feeder was replaced by an aggregated load. This model reduction allows minimizing the CPU time of simulations and is deemed acceptable in the past. However, the ever-increasing share of distributed generation requires adopting accurate models of both transmission systems and active distribution feeders, which leads to a staggering increase in the total number of nodes/equations and CPU time. We propose a numerical method that is up to two orders of magnitude faster than existing methods in performing integrated transient stability simulations of transmission systems and distribution feeders in the three-phase frame. We show the numerical properties and efficiency of the proposed method by simulating a well-known transmission system connected to several distribution feeders with a high penetration of inverter-based resources.

Abstract:As the rapid growth of renewable energy sources (RESs) with fluctuating and uncertain power generation, reserve shortages become increasingly severe, limiting the further integration of RESs and making it essential to exploit additional reserve sources. The fast-growing data centers offer significant potential to provide reserves due to their spatial and temporal dispatch flexibility. However, large quantities of jobs in the data centers have different computing requirements, making it challenging to estimate the availability of data centers for providing reserves. Meanwhile, the regional resource mix leads to uneven reserve distribution and mismatches between reserve supply and demand. The cross-region reserve supply has the capability to mitigate such mismatches as it allows different regions to provide reserve to each other. This paper enriches the reserve sources in multi-region system dispatch to enhance renewable energy integration by exploiting the reserve supply capabilities of data centers and cross-region systems. The model for available reserves of data centers is proposed via analyzing the envelope of computing constraints of aggregated jobs. In addition, a comprehensive cross-region reserve supply model is presented by considering reserve supply and demand limitations and reserve transmission capability. Case studies are performed and the results show that the data centers can provide a substantial amount of reserves. Both the migration of jobs among data centers and the cross-region reserve supply transfer reserve resources among multiple regions. Simulations on a real-world power grid system further demonstrate that unlocking the reserve supply potential of data centers can reduce operating costs by 0.4% and the curtailment of RESs by 1.7%.

Abstract:This paper presents an industrial-grade hardware-in-the-loop (HIL) validation method for a wide-area monitoring system designed to detect electromechanical oscillations in power systems. The proposed method leverages dynamic mode decomposition (DMD) to extract spatiotemporal patterns from synchronized phasor measurements, enabling accurate identification of oscillation modes. Traditional methods are widely used but face limitations in accurately capturing complex system dynamics. Despite the improved processing capabilities of modern controllers, advanced data-driven methods such as DMD remain underutilized due to concerns about computational cost and implementation complexity. This paper demonstrates the feasibility of integrating DMD into industrial-grade controllers by employing efficient algorithms such as singular value decomposition and QR decomposition. A comparative analysis with the Prony method across multiple test systems, along with industrial-grade hardware-in-the-loop validation, confirms the accuracy and computational efficiency of DMD for real-time applications. Results show that DMD reliably identifies local modes, inter-area oscillations, multimodal behavior, and mode shapes. These findings support the integration of spatiotemporal methods into industrial-grade controllers to improve the performance of real-time monitoring on power system stability.

Abstract:The emergency control strategy for mitigating cascading failures plays an important role in the safe and stable operation of power systems with high integration of renewable energy sources (RESs). To mitigate the high control costs associated with active splitting, this paper proposes a novel tree-partitioning based emergency control strategy. Initially, a coherency grouping model for power system integrated with wind turbines is established, incorporating an improved phase motion equation to facilitate the identification of unit homology. Furthermore, a modified fuzzy c-means (FCM) clustering algorithm is proposed to achieve the grouping of the generators. Then, the cluster partitioning issue is converted into an eigenvalue solution problem. This allows for a rapid and accurate cluster partitioning of the power system, taking into account both the random fluctuations of renewable energy output and the unit homology constraints. Based on the modified Prim’s algorithm, the tree-partitioning method is used to select the optimal bridge, whereby other bridges are disconnected and the inter-cluster power imbalances remain stable. A power adjustment measure is proposed to determine the power adjustment amount within each cluster based on the power flow tracing. Simulations on the IEEE 118-bus system and the actual case system demonstrate that the proposed strategy reduces control costs by 55.60% and 43.26% compared with active splitting, while also accelerating the post-fault recovery. These results highlight the potential of the proposed strategy for mitigating cascading failures in real-world applications. Index Terms—Cascading failure, emergency control, tree-par‐ titioning, cluster partitioning, coherency grouping, spectral clus‐ tering.

Abstract:The significance of system inertia, especially its non-uniform spatial distribution, is becoming paramount in the power system. The scope of inertia estimation has traditionally been the estimation of overall or total system inertia. However, as frequency dynamics become increasingly localized with the increasing penetration level of inverter-based resources (IBRs), the need for higher spatial resolution (geographically localized estimation) and faster temporal resolution (online or continuous estimation) in inertia estimation becomes paramount. This paper proposes an analytical method to estimate the spatial inertia distribution down to the transmission bus level, i.e., nodal inertia. Depending on data availability, the proposed method is flexible and can be used in two ways ① to estimate nodal inertia under any given operating condition (or snapshot), or ② to continuously estimate nodal inertia under both ambient and transient conditions using available measurements from local phasor measurement units (PMUs). The novelty of the proposed method lies in its analytical formulation, which does not require rate of change of frequency (ROCOF) measurements or rate of change of power injections, making it immune to the noise associated with the estimation of these derived quantities. Additionally, the proposed method does not require defining near-zero ROCOF thresholds, which is a system-specific and non-trivial problem. The proposed method is mode-agnostic, which makes it more general than the dominant mode-based linearized methods. The applicability of the proposed method is demonstrated through simulation studies performed on the IEEE 39-bus and IEEE 68-bus test systems with varying penetration levels of IBRs. The robustness of the proposed method is numerically assessed against modeling and measurement uncertainties.

Abstract:The integration of renewable energy sources into power grids may introduce wideband oscillation risks, challenging the stability of modern power systems. Traditional artificial neural network-based data-driven impedance identification methods face significant limitations due to the black-/gray-box characteristics of wind power units (WPUs) and the scarcity of impedance measurement data. To address these challenges, this paper proposes few-shot data-driven online impedance identification and stability assessment for wind-integrated modular multilevel converter-based high-voltage direct current (MMC-HVDC) systems. By enhancing the backpropagation neural network (BPNN) with adversarial domain adaptation (ADA), the proposed online impedance identification leverages transfer learning to develop wideband impedance identification models for WPUs and MMCs, enabling online impedance identification with minimal data requirements and achieving the direct model transfer from WPUs to onshore MMC. A comprehensive case study of the Rudong offshore project in China demonstrates the effectiveness of the proposed few-shot data-driven online impedence identification and stability assessment, showing a 95% reduction in data requirements and significant improvements in model transferability compared with conventional methods.

Abstract:Advanced network analysis and control tools in energy management systems depend on a reliable real-time database provided by power system state estimation (SE). The processing of gross errors— bad data (BD)— is considered an essential SE step in such a context. It is widely recognized that the adverse conditions with weak network observability hinder the identification and correction of multiple BD, a process notorious for its time-consuming nature. This paper proposes a method for facilitating the processing of BD in supervisory control and data acquisition (SCADA) measurements using phasor measurement units (PMUs). The main objective is to strategically allocate a limited number of PMUs to allow the phasor-aided state estimation (PHASE), which offers the advantage of effectively handling single and multiple BD. Tests with the IEEE 14-, 30-, and 118-bus systems are conducted to demonstrate the practical application of handling BD, including adverse conditions such as network spots with low redundant measurements. Index Terms—Bad data, gross error, state estimation, energy management system (SCADA), phasor measurement unit , supervisory control and data acquisition (PMU).

Abstract:The timely detection of internal faults in permanent magnet synchronous generators (PMSGs), which are the key components of direct-drive or semi-direct-drive systems, is crucial for ensuring the long-term stable operation of wind turbines. Large-scale experiments are impractical for acquiring sufficient fault data, whereas simulation can effectively provide such data. Furthermore, the power electronic devices in wind turbines exhibit microsecond-level dynamic characteristics, necessitating electromagnetic transient (EMT) simulation. Moreover, the black-box models provided by commercial EMT simulation software do not support internal fault simulation. Additionally, existing modeling methods for internal faults in PMSGs can only simulate the generator itself, making it difficult to generate nodal equivalent circuits and preventing direct interfacing with external components such as converters in wind turbine systems. This paper proposes the EMT modeling and simulation method capable of representing various types of internal faults in PMSGs. By integrating two state variables, a nodal quivalent circuit is developed, effectively avoiding the calculation of time-varying partial derivatives. The proposed method can directly interface with converters and grid connections, enabling the fault characteristics to be reflected in the wind turbine system. A unified EMT model encompassing multiple fault types is developed through a standardized modeling procedure. The proposed method is implemented in PSCAD/EMTDC and compared with results obtained from MATLAB. The results demonstrate that the proposed method can accurately reflect the characteristics of internal faults, validating its effectiveness.

Abstract:The virtual synchronous generator (VSG) control is increasingly adopted in multi-paralleled photovoltaic generation systems (MP-PGSs) due to its enhanced grid-support capability. While the transient stability of VSG-controlled photovoltaic generation systems (PGSs) under symmetrical grid faults is well-studied, instability mechanisms under asymmetrical grid faults (AGFs) remain underexplored. This paper establishes a multiple coupling analysis model for transient stability analysis of VSG-controlled MP-PGSs under AGFs. Based on this model, the impacts of coupling dynamics including sequence coupling and mutual coupling on the transient stability of VSG-controlled MP-PGSs are investigated. Furthermore, the influence laws of key parameters on the transient stability are analyzed. To improve the low voltage ride-through capability of MP-PGSs under AGFs, a multi-objective stabilization control method is proposed, which satisfies both the grid codes and current limitation requirements. Finally, simulation results validate the correctness of the theoretical analysis and the effectiveness of the proposed control method.

Abstract:The increasing penetration of renewable energy sources will impose even more stress on the operational flexibility at multiple timescales. Energy storage (ES) is a promising option to provide multiple services, while various energy storage systems (ESSs) exhibit diverse economic performances at different timescales. However, efficiently and economically combining multi-timescale ESSs to meet flexibility requirements is challenging due to the gap between coarse-grained ESS representations and multi-timescale flexibility requirements. This paper presents a wavelet packet decomposition (WPD) based multi-timescale operational flexibility quantification method. Such requirements are clustered and then satisfied by an ES planning model covering multiple timescales from intra-hourly to seasonal using representative scenarios, while considering both short-term (operational) and long-term (technology cost) uncertainties. An empirical analysis of a provincial power grid in East China is performed to obtain the planning results in 2030 and 2060. Numerical results demonstrate the effectiveness of the multi-timescale ES planning model as well as its computational performance and economic advantages.

Abstract:Although virtual asynchronous machine (VAM) control has been proposed for virtual energy storage systems (VESSs), research into its secondary control applications is still limited. Thus, a distributed secondary frequency restoration control strategy based on VAMs is presented for VESSs. First, the VAM control is introduced, and a detailed electro-thermal coupling model of the VESS is developed. This model includes indoor-outdoor temperature differences, heat transfer through building envelope (walls, windows, and roof), solar radiations, ventilation losses, and electric boiler dynamics. It effectively captures the coupling between indoor temperature regulation and grid power balancing. Next, a distributed secondary frequency restoration control strategy based on VAM is proposed. It addresses parameter heterogeneity within a nonlinear multi-agent framework among VESSs. The nonlinear dynamics are converted into a linear reference model, which simplifies controller design and stability analysis. Using only local and neighboring information, the proposed strategy restores frequency and ensures active power sharing. Furthermore, the proposed strategy coordinates thermal power regulation to maintain indoor temperature balancing across VESSs within seasonal thermal comfort ranges. This improves thermal comfort without compromising dynamic response. Finally, the stability of the proposed strategy is verified using Lyapunov method, and simulation results from an islanded microgrid (MG) test system under parameter variations, communication imperfections, and winter/summer operating scenarios validate the effectiveness and robustness of the proposed strategy.

Abstract:Energy storage and generalized energy storage (GES) such as electric vehicles (EVs) and heating, ventilation, and air conditioning (HVAC) systems play a critical role in enhancing the flexibility of power systems. The shared system architecture can improve utilization rates of energy storage and reduce configuration costs. However, the heterogeneity and inherent uncertainty of GES pose significant challenges to the rational configuration and operation of energy storage. To address this, we employ a two-stage distributionally robust capacity configuration method for shared energy storage considering the flexibility and uncertainty of distributed resources. First, by integrating the operational characteristics of both physical and virtual energy storage, a shared system architecture is proposed for the GES. Second, to overcome the limitation of existing EV aggregation models, which are typically tailored for ideal battery behaviors, a more accurate aggregation model is introduced based on the parameter planning, termed the GES model. An uncertainty probability set modeling approach for demand-side resources is derived based on this model. Finally, considering the flexibility and uncertainty associated with EVs and HVAC systems, a distributionally robust optimization model for shared energy storage capacity configuration is proposed. Simulation results demonstrate that the proposed method increases the revenue of operators and the utilization of energy storage resources, while also reducing the configuration capacity.

Abstract:With high integration of distributed generators (DGs) and diverse topology reconfigurations, the dynamic states of active distribution networks (ADNs) become more complex, which poses great challenges to the dynamic equivalence of ADNs. In this paper, motivated by the similarities between Newton Raphson power flow calculation (NRPFC) and computation of convolution network (CNN), and based on the capability of recurrent neural network (RNN) to represent the differential algebraic equations (DAEs) of loads, we propose a multimodal machine learning based equivalent modeling method in order to track the dynamic behaviors of ADNs. First, the multimodal machine learning is divided into two modules, one of which is gated recurrent unit (GRU) + fully connected (FC) module to represent the DAEs for load modeling, and the other is CNN + attention module to simulate the NRPFC to extract spatial feature changes of power system caused by disturbances or topology reconfigurations. Then, the robustness and generalization abilities of the proposed method are evaluated by different test systems, where the IEEE 14-bus transmission network is connected to distribution networks with different scales (i.e., IEEE 33-bus distribution network and IEEE 57-bus distribution network). The simulation results reveal that the proposed method can accurately capture the dynamic behaviors of ADNs under different operation conditions. In addition, the precision and generalization of the proposed method is tested in comparison with other deep learning (DL)-based equivalent methods.

Abstract:The dispatch of demand-side resources is becoming increasingly essential for enhancing the resilience of distribution systems against extreme weather events. Traditional studies rely primarily on direct load curtailment to mitigate power shortages, thereby neglecting the power demand of different customers. To bridge this gap, a novel coordination method of transactive demand response (DR) and rolling outage management of diverse loads is proposed. First, the DR program is designed to incentivize voluntary load adjustments by coordinating participation of private consumers. Considering the diversity and characteristics of customers during DR, detailed models of diverse loads are established, including an energy-material flow model of industrial loads (ILs), an adjustable load model of commercial loads (CLs), and an outage-sensitive model of residential loads (RLs). When the supply-demand imbalance exceeds the adjustment capacity of DR, rolling outage measures are integrated into the proposed method to reduce losses incurred by load shedding. The coordination of transactive DR and rolling outage management is formulated as a bilevel optimization problem from system operators and responsive load, which is solved by the Stackelberg game-theoretic approach. Finally, the proposed method is tested on the modified IEEE 33-node distribution system. The results show that the proposed method can effectively ensure customer profit and reduce load interruption loss by dispatching demand-side resources during restoration.

Abstract:With the continuous development of renewable energy technologies and expansion of their system scales, a large number of renewable energy sources are being integrated into distribution networks. These renewable energy sources are increasingly managed in the framework of active distribution networks (ADNs). However, the expansion of the system scale and stochastic nature of user behaviors produce dynamic changes in the active and reactive power flows of these ADNs, which generate unpredictable operational stability problems such as voltage deviations. To address this challenge, a novel active and reactive power coordination control strategy is proposed for voltage control in ADNs. The strategy is based on a deep reinforcement learning framework, while introducing an attention mechanism to train an agent model for voltage control in an ADN in response to the system dynamics. To this end, this study first formulates an optimization problem for mapping between the active-reactive power and bus voltages in a system. Then, this problem is reformulated as a Markov decision process and solved using a local cross-channel interaction network-based soft actor-critic (LCCIN-SAC) algorithm. Simulation results on a modified IEEE 69-bus system demonstrate that the proposed strategy can successfully improve voltage deviation produced by rapid dynamic changes in a distribution network. The proposed strategy effectively ensures the stable and reliable operation of a distribution network with a high penetration of renewable energy sources.

Abstract:Distribution system reconfiguration (DSR) uses switching actions, which are available to distribution system operators (DSOs) without regulatory changes, to optimize the grid topology. Although distribution systems often support meshed operation, they are typically operated radially to simplify fault isolation and improve reliability. The minimization of active power losses not only improves efficiency but also helps prevent voltage and current limit violations. However, the DSR problem is computationally difficult due to its combinatorial and non-convex nature. This paper proposes a cycle-basis-informed heuristic method for radial DSR, aiming to reduce active power losses and stabilize the voltage. The proposed heuristic method works directly with the non-convex AC load flow model, considering switching actions as the only degrees of freedom. A large neighborhood search framework is used, constructing restricted mixed-integer nonlinear programming (MINLP) subproblems of controllable complexity. Radiality is ensured via a graph-theoretic method based on the cycle basis of the system, enabling systematic exploration of feasible configurations. The proposed heuristic method is evaluated on benchmark systems with varying load profiles, demonstrating robust performance, effective loss reduction, improved voltage profiles, and practical computation time. These results confirm its applicability to real-world DSR in radial AC systems.

Abstract:Extreme weather conditions, characterized as high-impact low-probability (HILP) events, pose significant threats to distribution network (DN). Soft open points (SOPs) have emerged for precise power flow regulation and voltage establishment, and hence can be used for post-fault load restoration to improve the DN resilience, incorporating scenarios generated from weather profiles. In this paper, an extreme weather risk-averse planning method for SOPs is proposed. A two-stage scenario-based stochastic programming (SBSP) model is established to minimize expectation and conditional value-at-risk (CVaR) of load shedding and network loss. Due to the integer variables introduced by DN reconfiguration constraints in the operational stage, the classic Benders decomposition algorithm is inapplicable. To address this challenge, we develop a novel generalized Benders decomposition (GBD)-based solution algorithm, designed to improve the computational efficiency on large-scale cases. Lift-and-project (L&P) cutting plane is employed to derive Benders cuts through the convex hull of mixed-integer second-order cone programming (MISOCP) subproblems. Finally, the effectiveness of the proposed method is demonstrated by numerical experiences on 33- and 123-node test DNs.

Abstract:To address the limited adjustable capacity of distribution networks (DNs) under the large-scale integration of flexible resources into microgrid (MG), an MG flexible operation region (MGFOR) model is proposed to quantify the adjustability of MG under aggregated regulation cost constraints. A coordinated dispatch framework of DN and MG is then developed based on MGFOR to enable cooperative flexibility allocation. The original boundary of MGFOR is determined through a radial iterative linear search, and its construction is completed using a convex-hull fitting approach. The simulation is conducted on the modified IEEE 33-node test system. It is demonstrated that MGFOR explicitly quantifies adjustable power boundaries, yielding 2.2% lower total operation cost for the proposed coordinated dispatch framework of DN and MG compared with hierarchical dispatch framework of DN and MG. The voltage safety margins at noon are enhanced by up to 56.46% at distribution feeder terminals. Furthermore, the impact of operation modes of distributed photovoltaic units on node voltage in DN is comparatively analyzed.

Abstract:The growing integration of renewable energy sources is driving microgrids (MGs) toward 100% inverter-based architectures, whose system stability and dynamic performance are tightly coupled with both the MG plant (i.e., resource type, number, and placement) and inverter control strategies. However, the existing MG design approaches typically overlook this coupling and fail to assess the dynamic performance during the design phase explicitly. Therefore, they often produce suboptimal configurations that do not satisfy the dynamic performance requirements once the design process is complete. To address this challenge, this paper introduces a plant-control co-design approach for 100% inverter-based MGs, which simultaneously optimizes the MG plant design and its control to minimize the system costs while ensuring the stable operation and enhanced dynamic performance. The proposed co-design approach systematically explores the feasible plant designs, evaluating their transient responses under disturbances using the optimal controllers. The MG plant design is formulated as a mixed-integer linear program, while MG dynamics are captured through nonlinear differential-algebraic equations. Electromagnetic transient simulation is used to validate the stability, compliance with IEEE 1547 standard, and dynamic performance, including power tracking in the grid-connected mode and disturbance rejection in the islanded mode. Case studies demonstrate the impacts of the number and placement of battery energy storage systems (BESSs) on the dynamic performance of 100% inverter-based MGs.

Abstract:With the increasing penetration of distributed energy resources, alternating current (AC) microgrids face persistent challenges in maintaining frequency and voltage stability, as well as achieving active and reactive power sharing under renewable intermittency and load variations. To address these issues, this paper proposes a predefined-time distributed secondary control (PTDSC) strategy that ensures fast and accurate frequency and voltage recovery as well as active and reactive power sharing within a user-specified convergence time, independent of the initial conditions and complex tuning of gain. A key component of the proposed strategy is a predefined-time distributed average-voltage observer (PTDAVO), which enables fully distributed observation of the global average voltage. By incorporating the observed average voltage into the secondary control, the proposed strategy achieves coordinated voltage recovery and reactive power sharing within a distributed predefined-time framework. Stability results are rigorously established by the Lyapunov stability theory. Extensive simulation validates the effectiveness, flexibility, and plug-and-play capability of the proposed strategy, along with its adaptability to time-varying communication topologies and superiority over existing strategies.

Abstract:Dynamic microgrids (MGs) possess flexible system topologies and power flow control capabilities, enabling adaptive responses to outages caused by extreme events. This paper proposes a multi-region and multi-stage coordinated control method for the resilience enhancement of dynamic AC/DC MGs, aiming to address limitations arising from rigid electrical boundaries and insufficient coordination of heterogeneous resources during distribution network restoration. The proposed method introduces smart switches (SSWs) to establish dynamic electrical boundaries, thereby partitioning the system into multiple self-operable mini-AC/DC MGs. A novel multi-mode coordinated control scheme is designed to govern grid-forming (GFM) distributed generators (DGs), grid-following (GFL) DGs, and coupling converters (CCs). Specifically, GFM DGs operate in voltage-source mode, maintaining system frequency and voltage via droop control; GFL DGs operate in current-source mode, actively contributing to system regulation through inverse droop control; and CCs manage bidirectional power transfer between AC and DC subgrids based on inverse droop characteristics. Furthermore, a distributed secondary control layer enables the coordinated operation of GFM DGs, GFL DGs, and CCs across multiple regions and stages, ensuring zero-deviation restoration of voltage and frequency, proportional load sharing, and seamless topology reconfiguration. Notably, GFL DGs and CCs remain actively engaged throughout the entire restoration process, thereby maximizing the utilization of available resources and alleviating the capacity burden on GFM DGs. Simulation studies conducted on a 12-bus hybrid AC/DC test system validate the effectiveness of the proposed method in enhancing the resilience and adaptability of hybrid AC/DC MGs under extreme operating conditions.

Abstract:The integration of renewable energy sources and the rising load demand has introduced significant challenges to the secure operation of interconnected power systems. To address these issues, a novel grid-connected microgrid control strategy based on higher-order sliding mode observer (HOSMO) and fractional-order sliding mode controller (FOSMC) is proposed for a grid-connected microgrid incorporating photovoltaic (PV) system, wind energy conversion system (WECS), and battery energy storage system (BESS). The proposed control strategy ensures effective DC bus voltage regulation by optimally managing the PV, WECS, and battery converter units. The HOSMO is designed to accurately estimate the current of PV or WECS converter, enhancing system robustness, while the FOSMC minimizes the chattering in DC bus voltage deviations. The performance of FOSMC is rigorously evaluated through extensive MATLAB/Simulink simulation under various uncertainties, including generation fluctuations, DC load variations, and DC faults. Stability analysis and comparative studies have been presented to demonstrate the superiority of the proposed control strategy. For comparative analysis, advanced control techniques such as finite time disturbance observer (FTDO)-based fixed time terminal sliding mode controller (FTTSMC) and nonlinear disturbance observer (NDO)-based back-stepping sliding mode controller (BSMC) are considered. Additionally, real-time validation of the proposed FOSMC is executed using the OPAL-RT (OP4510) platform to confirm the feasibility and reliability under practical operation conditions.

Abstract:Hydrogen - electricity coupled DC microgrids (HE-DCMGs) represent a promising and sustainable solution for off-grid power supply. However, achieving high economic performance while ensuring DC bus voltage stability is a critically challenging task. This study proposes a hierarchical energy management and control strategy for HE-DCMGs that integrates an adaptive mutation Harris hawks optimization (AMHHO) algorithm at the system level with a fractional-order sliding mode controller (FOSMC) at the device level. A multi-objective optimization problem is formulated to minimize hydrogen consumption and reduce degradation of proton exchange membrane fuel cells and lithium-ion batteries. The AMHHO algorithm, augmented with differential evolution and Lévy flight mechanism, determines the optimal power allocation among distributed sources, while the FOSMC provides robust DC bus voltage regulation. The proposed strategy is validated on a 750 V HE-DCMG experimental platform capable of 168 hours of off-grid operation. Experimental results show that the proposed strategy reduces long-term operating costs and improves energy-utilization efficiency, achieving an overall system efficiency of 80.49%-97.37%. The DC bus voltage is maintained with a response time of 0.02 s, a low overshoot of 3.7%, and a voltage fluctuation rate of 3.08%, all of which comply with the requirements of IEEE Std 1547‒2018.

Abstract:In the landscape of low-carbon transition in the power system, it is imperative for the system operator (SO) to implement electricity-carbon joint management. Currently, the increasing integration of renewable energy sources (RESs) is facilitating the achievement of emission reduction targets. However, the inherent uncertainties of RES power output pose challenges on power system operation, highlighting the needs for developing a power prediction model that serves as the prerequisite of better scheduling decisions. Nevertheless, existing accuracy-oriented prediction may not necessarily guarantee better decisions. Besides, due to the inevitable prediction errors, SOs have to adjust power outputs of thermal generators (TGs) during the intraday redispatching, leading to unexpected emission variations for each generation companies (GENCOs). Under the current centralized emission trading scheme (ETS), GENCOs with lower emissions are unable to fully utilize their emission allowances, while those exceeding their limits may face high penalties. However, these two groups of GENCOs exhibit inherent complementarity in terms of emission allowance consumption. To address the above challenges, this study proposes a novel multi-stage electricity-carbon joint management framework, where the power prediction model is decision-oriented to focus more on cost-saving. Moreover, bilateral trading contracts for emission allowances among GENCOs are incorporated into the proposed framework to promote the sufficient utilization of allocated emission allowances and prevent emission exceedances, thereby enhancing the total social welfare. Extensive simulations on a modified IEEE 30-bus system statistically verify the effectiveness of the developed decision-oriented predict-then-optimize method in terms of reducing operation cost. The welfare improvement of GENCOs brought by the designed bilateral trading contracts is also verified through simulation studies.

Abstract:Power systems are undergoing a policy-driven transition to decarbonization, which increases uncertainty and fluctuation due to the high penetration of renewable energy. Concurrently, advancements in communication and control technologies are unlocking greater demand-side flexibility. However, existing conventional long-term resource planning models for power systems do not adequately integrate demand response (DR) resources, operation simulations that account for uncertainty, and social factors such as government policy. To address these shortcomings, this paper introduces a robust planning framework for power system decarbonization pathways, specifically considering DR load as a critical flexible resource along with other techniques and impacting factors. Firstly, the long-term impacts of government policies and other social factors on system resource investment costs and development constraints are quantitatively considered, and the adjustable and transferable DR loads as key flexible resources are incorporated to smooth the fluctuations caused by renewable energy generation. Then, a three-stage robust planning model is developed by integrating long-term development planning, day-ahead unit commitment simulation, and intra-day power dispatch simulation to obtain the optimal solution with the lowest total cost over the planning period. Moreover, the column and constraint generation algorithm is modified to solve the planning model, specifically designed to address the decision-dependent uncertainty arising from new asset investments. Case study based on a provincial power system shows that the proposed framework enhances the integration of renewable energy by applying DR resources in line with real social development trends. The combination of long-term system planning, refined short-term operation simulation, and strategic DR integration ensures that the resulting decarbonization pathway is not only economically feasible but also highly robust and reliable.

Abstract:Virtual power plants (VPPs) promote the high-level integration of distributed energy resources (DERs) through complementation and aggregation. To avoid the privacy leakage and enhance the market efficiency, this study aims to achieve the non-iterative market participation of VPPs by accurately characterizing their external bidding functions. To this end, an improved multi-parametric linear programming (MPLP) method is developed to derive the bidding function representation of VPPs. The proposed method projects the detailed internal model of VPPs onto the point of common coupling (PCC). An algorithm for determining a precise initial parameter space (IPS) is proposed, thereby avoiding the redundant solutions that typically occur in traditional MPLP method. The IPS is then partitioned into several critical regions (CRs), within each of which the marginal cost remains constant, facilitating a piecewise linear mapping between the trading power and bidding price. Case studies on modified IEEE 33-bus and IEEE 123-bus systems demonstrate that the proposed method accurately characterizes the bidding functions of VPPs and substantially improves the efficiency of their non-iterative market participation while ensuring data privacy.

Abstract:Multi-energy retailer (MER) is a new type of retailer participating in both the electricity and gas markets. In a cooperative and competitive market environment, how to find the optimal coalition structure (OCS) model for MERs is an unsolved problem. Thus, this paper proposes an OCS model for MERs in electricity and gas markets. First, an MER optimization model is built to maximize profit, and the market-clearing constraints are considered. A coalition game model for MERs including coalition value and OCS model is introduced, which describes the behavior of forming a coalition among MERs considering coalition cost and scale limitation. Moreover, a hybrid algorithm including the diagonalization algorithm (DGA) and dynamic programming algorithm (DPA) is designed to solve the OCS model. Finally, the case studies are performed on a system with IEEE 14-bus electric network and 7-bus gas network. The numerical results show that the proposed model can effectively obtain OCS.