Deep learning (DL) is a promising tool for enabling fast transient stability assessment (TSA) in power systems. However, the weak interpretability and confidence biases of DL may hinder its practical applications, where the latter could even mislead decision-making. This article proposes a calibrated explainable learning method to promote the credibility of DL-based TSA models, thus providing more reliable auxiliary information for operators. By combining tensorized learning units and a mixture attention mechanism, an explainable TSA model is first built, where attention factors can be dynamically assigned according to physical characteristics of input variables and faults, offering a visual way to understand model decisions. Then, an ensemble confidence calibration method is developed to modify the mapping relation between network logits and posterior probabilities by minimizing negative log-likelihood, making confidence estimations better reflect the true correctness likelihood. Test results verify the efficacy of the proposed credibility-enhanced TSA model. It can form expert knowledge-match attention patterns and output more reliable confidence estimates, helping to correct wrong decisions of confidence-dependent applications and contributing to practical decision-supporting.

The Cascaded H-bridge (CHB) multilevel inverter stands out as an optimal topology for integrating large-scale photovoltaic (PV) power systems into the grid. Due to its multiport characteristic, each submodule directly interfaces with distributed PV strings. Nonetheless, PV power generation often experiences mismatches among three phases and between arms of one phase due to factors such as nonuniform irradiation, partial shading, and temperature variations. Addressing these challenges, this study introduces a novel CHB multilevel inverter with a parallel structure. Consequently, the added power circulation paths facilitate phase power exchange through circulating current injection (CCI), effectively eliminating phase power mismatch and power mismatch between arms in one phase. By transitioning the power exchange mode from voltage-driven to current-driven, the prevalent voltage overmodulation issue associated with the traditional CHB multilevel inverter is resolved. Additionally, this study offers a quantitative assessment of the power balancing capability of the suggested inverter in PV applications. Evidence suggests that, by employing CCI, the proposed inverter significantly enhances power balance capabilities compared to the traditional CHB multilevel inverter that utilizes zero-sequence voltage injection (ZSVI). Both simulation and experimental results validate the proposed approach.

Due to low inertia and restricted frequency regulation resources, isolated system frequency may deviate from nominal operating conditions easily. The conventional solution is to install energy storage or curtail PV power, which is quite costly. This paper proposes a closed-loop feedback control strategy to provide quick frequency support based on conservation voltage reduction effect of voltage-sensitive loads. The power of different loads is regulated coordinately and continuously considering the difference in CVR effect and voltage state between each load. Thus, a two-stage adaptive control method based on the adjustable margin of each load is proposed to achieve online optimization of feedback coefficients. Voltage quality of each load can be guaranteed in the control process while system transient nadir can be minimized. The proposed scheme enables fast and continuous power support through the supply-side to respond to fast power fluctuation caused by PVs. Simulation is performed on the RTDS based on a real system in China, and results verify effectiveness of the proposed method.

Due to its large capacity and rapid adjustment capability, the electrolytic aluminum load (EAL) possesses significant potential for flexibility and can participate in demand response (DR) to alleviate peak regulation pressures on the power system. Within the frameworks of carbon emissions trading (CET), green-certificate trading (GCT), and renewable portfolio standards (RPS), the electricity-carbon benefits of the EAL acting as a multi-market participant are tightly coupled. Therefore, it is imperative to develop an electricity-carbon collaborative optimization strategy for EALs participating in DR. First, based on the dynamic carbon emission characteristics of EAL, a GCT-CET coupling model is constructed, and the ladder CET cost structure is enhanced to better characterize carbon trading expenses. Then, the production regulation model, regulation benefit model, and DR model are formulated for the EAL. Finally, considering the uncertainty of on-site photovoltaic (PV) generation, a stochastic electricity-carbon optimization model is proposed, incorporating GCT, CET, and DR, with the objective of maximizing the comprehensive net benefits. A real-plant case study demonstrates that the proposed model increases the EAL's comprehensive net benefit while substantially reducing average energy consumption and total CEs. Sensitivity analyses with respect to the compensation price coefficient, carbon-quota auction ratio, and carbon-quota reduction ratio further verify robustness. Overall, the results indicate that electricity-carbon co-optimization enables EAL participation in DR to deliver both economic and environmental benefits.

The sending-end hybrid cascaded (SE-HC) HVDC system formed by a DC-side series connection of a line-commutated-converter (LCC) and a modular-multilevel-converter (MMC) has become a crucial solution for large-scale renewable energy long-distance transmission due to its advantages of large transmission capacity, high operation flexibility, and strong grid support capability compared with conventional LCC-HVDC systems. However, SE-HC MMC may face DC overvoltage under sending-end (SE) AC system faults, which is caused by the HVDC system's active power surplus. To address this issue, this paper first analyzes the mechanisms and influencing factors of DC overvoltage based on the power balance model. Then, an overvoltage suppression strategy is proposed by coordinating the transmitted active power from SE and receiving-end (RE) converters, renewable energy bases (REBs), and energy dissipation devices. This method does not rely solely on the energy dissipation device to absorb surplus active power. Instead, by fully leveraging the control flexibility of the SE MMC, it enables the RE MMCs and REBs to detect SE AC system faults, thereby increasing active power transmission and reducing surplus active power during the fault. As a result, more effective suppression of DC overvoltage is achieved, and the required capacity of the energy dissipation device in the system is reduced. A monopolar model of the SE-HC HVDC system has been established in PSCAD/EMTDC, and the feasibility and superiority of the proposed control strategy have been verified through this model.

Distribution networks with high penetration of distributed photovoltaics and plug-in electric vehicles exhibit inherent intermittency and volatility, which frequently lead to voltage violations. Through incorporating the charging power attenuation characteristics of plug-in electric vehicles, as well as the static voltage characteristics of induction motors and ZIP loads, this paper proposes a two-stage multi-objective collaborative optimization method for volt/VAR optimization. At the day-ahead hourly stage, a stochastic optimization model coordinates the active and reactive power of multiple controllable resources with the objective of minimizing total network losses and the number of switching operations of on-load tap changers and capacitor banks. The proposed method achieves more accurate voltage and reactive power control by accounting for multiple types of load characteristics. At the intra-day 15-minute stage, a method for identifying pilot buses is proposed, taking load characteristics into account. A rolling optimization model for volt/var control is then established, incorporating dual time-scale coordination. The objectives include minimizing network losses, reducing control device operations, and minimizing voltage deviations at pilot buses. This intra-day model effectively mitigates voltage fluctuations caused by real-time variations in source and load, thereby further reducing power losses. Finally, simulation studies are carried out on a modified IEEE 33-bus system. The results demonstrate the effectiveness and novelty of the proposed two-stage collaborative optimization approach.

With the increasing penetration level of renewable energy sources (RES) interfaced with the grids by power electronic converters, the grid forming (GFM) control of converters is recognized as one of the promising solutions in dealing with challenges such as grid transient stability enhancement required for future power electronic converter-dominated power grids. However, the physical nature of the limited overcurrent capability of GFM converter constrains its application including grid transient stability enhancement during grid faults. To address this problem, this paper first analyzes the mechanism of grid transient stability and overcurrent of GFM converter during grid faults. Then, an adaptive power and virtual resistance cooperation control strategy of VSG-based GFM converter for grid transient stability enhancement is proposed, which can suppress the overcurrent of the GFM converter and enhance the grid transient stability simultaneously during grid faults. Meanwhile, a detailed quantitative analysis of parameter tuning of the proposed control is conducted, ensuring the robustness of the proposed control strategy in different fault depths and fault duration times. Finally, case studies are simulated in the PSCAD/EMTDC software to verify the effectiveness of the proposed control strategy.

When power disturbances occur in power electronic converter-dominated power systems, the inherent node frequency distribution characteristics, coupled with the dynamic disparities in the power/frequency response characteristics of grid-connected converters, render the system susceptible to power oscillations among different converters. These oscillations, in turn, pose a significant threat to the stable operation of the power system. To address this issue, this paper proposes a coordinated power oscillation suppression control method based on node frequency consensus for multiple grid-forming converters (GFMs) and grid-following converters (GFLs). Firstly, the differences in frequency response characteristics and power oscillation characteristics between GFM and GFL with active frequency support are analyzed, and a frequency response model for multi-GFM/GFL systems is established, which incorporates the response characteristic disparities among converters and system line parameters. Subsequently, the multi-GFM/GFL system is decoupled into a set of individual converter subsystems through modal decomposition theory, extracting the dominant factors influencing active power oscillations and their suppression under varying operating conditions. Then, leveraging the properties of power network and communication Laplacian matrices, a coordinated power oscillation suppression control method based on node frequency consensus is proposed, and the resulting stability enhancement is analyzed. Finally, simulation results demonstrate that the proposed method can effectively suppress power and frequency oscillations among grid-connected converter nodes by enhancing system damping, thereby supporting stable system operation and exhibiting strong adaptability.

This paper presents a robust emergency frequency control framework that leverages coordinated rapid regulation by wind turbines and loads to pre-generate control strategies for large hypothetical frequency sag faults and activate control to stabilize system frequency during faults. The underlying optimization problem considers multi-type resources with discrete and continuous control modes, while capturing uncertainties from environmental and modeling errors, nonlinearities, and non-analytical computations. Accordingly, this paper proposes a two-step successive solution approach to address the formulated mixed-integer nonlinear robust optimization problem efficiently. In step one, uncertainties are temporarily ignored, and the resulting mixed-integer nonlinear and non-analytical problem is solved using proposed simplified methods, such as differential discretization, which serves as a warm start for the original problem. In step two, the solution from step one is treated as a reference, and trajectory sensitivities are used to quantify the impact of uncertainties on frequency security and other constraints. The original problem is ultimately reformulated as a bi-level mixed-integer linear optimization with independent decision variables, enabling efficient solution. Finally, simulations on the modified IEEE 39-bus and 118-bus systems demonstrate that incorporating rapid regulation of wind turbines and loads significantly reduces control costs. Additionally, the proposed method ensures high solving efficiency and reliable control effects against modeling uncertainties.

This article proposes a variable voltage level base extended phase shift ((VL)-L-2-EPS) control for three-port dual active bridge (DAB) converter to achieve current stress optimization. By modifying the flying capacitor voltage balance control, the inner phase shift angle of DAB converter is released and can be used to realize extended phase shift (EPS) control. Furthermore, a variable voltage level can be constructed on the midpoint voltage of the primary side bridge arm by actively creating the unbalance of the flying capacitor voltage, so that the voltage excitation on the series inductor in the DAB converter varies within the inner phase shift interval thereby realizing a reconfiguration of the series inductor current, which reduces the high circulating power of the EPS control under light load conditions. In addition, by calculating the optimized trajectory of the (VL)-L-2-EPS control, the current stress of the three port DAB converter is substantially reduced while ensuring all switches of the three port DAB converter operate under soft switching conditions over the full load range. Finally, a 500 W rated prototype is built to verify the effectiveness of the proposed (VL)-L-2-EPS control.

Regional low-grade heat, such as ground source heat, domestic sewage waste heat, low-temperature industrial waste heat and data center waste heat, has considerable potential to contribute to the alleviation of energy crises and environmental pollution from a micro-grid perspective. To explore the method of park-level integrated energy systems (IES) that incorporates low-grade heat, a coordinated planning model is proposed in this paper. First, the working efficiency of heat pump (HP) is modeled using a combined method of thermodynamics and data regression. This HP model takes into consideration the heat upgrading process, which involves utilizing low-grade heat at various temperature levels. The conditional probability distribution of the HP model is also estimated according to the working temperature. Furthermore, the uncertainties arising from sources, loads, and HP models are addressed through chance constraints in a proposed stochastic planning model of IES. Specifically, the planning model is solved using an improved Grey Wolf Optimization (IGWO) algorithm, which enhances global search capability, local optimization accuracy, and stability. Simulation results show the advantage of integrating low-grade heat in IES with high-performance HP through multi-energy complementary mode. The planning model can coordinate multiple energy sources and provide a reference for decision-makers between risk and economy.

Network congestion is a frequent challenge that power systems need to handle in real-time operation. Though learning-based congestion event prognosis (CEP) is a promising way for early warning, its inadequate adaptability to topology alterations and lack of interpretability may hinder its practical applications. This article first proposes a feature combinatorial optimization (FCO) method to explore a variable set that can provide robust contributions for CEP on different topologies, where an information-assisted scheme is designed to facilitate the FCO efficiency and result. Then, using topologically robust variables, an explainable CEP model is built based on the tensorized learning network and mixture attention mechanism, where the contribution from individual variables can be explicitly tracked via variable-wise hidden states. Next, the data of static covariates are encoded into models to improve CEP performance. These components finally drive a CEP model with decent topology robustness and explainability. Numerical results validate the efficacy of the proposed method, indicating that it can effectively mitigate the degradation in CEP performance caused by topology alterations by at least similar to 32% and achieves similar to 95% average CEP accuracy even encountering unseen topological changes.

Islanded microgrids offer a reliable solution to maintain power supply and minimize the outages impacts when distribution networks disconnect from the main grid. This paper proposes a dynamic islanded microgrids formation method that ensures the frequency security of the microgrids, addressing challenges such as high renewable energy penetration and low system inertia simultaneously. The method enables energy storage systems (ESSs) to operate in virtual synchronous generator (VSG) mode, providing frequency support for the islanded microgrids. A frequency response model considering ESS-based VSG is developed to capture the dynamic frequency characteristic of islanded microgrids, ensuring that renewable energy sources (RESs) remain connected during the microgrid formation process. To achieve this, a mixed-integer nonlinear programming (MINLP) model is formulated, considering the frequency security constraints, network topology and RESs operational constraints of microgrids to be formatted. It dynamically optimizes ESS switching mode, branch switching, load shedding, and dispatch of distributed RESs for network reconfiguration to maximize the load supply capacity including critical loads after the distribution network disconnection from the main grid. Simulation results have shown that the proposed method is correctness and effectiveness.

Commutation failure (CF) during the fault recovery stage of line-commutated converter-based high-voltage direct current (LCC-HVDC) systems remains a critical issue under both rectifier-side and inverter-side AC faults. Although previous studies have examined CF mechanisms during fault recovery at the sending or receiving ends AC faults conditions, these scenarios have typically been treated as independent problems, lacking a unified suppression strategy. In this paper, based on simulation results of LCC-HVDC systems under sending-end and receiving-end AC faults, the mechanisms of CF during fault recovery are systematically analyzed. It is revealed that improper control mode switching of the inverter and the rapid advancement of the inverter-side commutation bus voltage phase angle are the primary causes of CF in both cases. To address this, a unified CF mitigation method based on adaptive extinction angle control and DC current control is proposed. By compensating for the advancement of the inverter-side commutation voltage phase and moderating the recovery speed of the DC current, the proposed method effectively suppresses CF during fault recovery at both the sending and receiving ends. Real-time digital simulation (RTDS) results confirm the effectiveness of the proposed approach in mitigating CF during fault recovery in LCC-HVDC systems. Compared with the benchmark control in the CIGRE benchmark HVDC system, the proposed method achieves reductions of over 90

To address the difficulty of source-load balance matching in power systems with high penetration of renewable energy, this paper proposes a multi-time scale automatic "scenario-strategy" matching method for power system dispatch. Firstly, a scenario matrix containing the time series of system states and temporal constraints of flexible loads is constructed, and typical scenarios are identified and prioritized using multidimensional criteria. Secondly, a multi-objective dispatch optimization model is designed to generate flexible load strategy sequences adapted to state evolution. Finally, a nonlinear mapping between scenarios and dispatch strategies is established based on the LSTM model and deep learning. Case studies show that, compared with traditional single-time scale strategies, the proposed method can effectively improve dispatch performance and provide important support for secure operation and dispatch of power systems under massive and complex operation scenarios caused by uncertainties of renewable energy.

Grid-forming control is currently a research hotspot for renewable energy grid-connected converters due to its inherent capabilities of providing inertia and voltage support, and it can serve as the primary control mode for renewable energy-synchronous machine hybrid transmission systems. During large grid disturbances such as faults, transient instability can readily occur in grid-forming converter-synchronous machine parallel systems. Concurrently, current saturation in the grid-forming converter alters the system's transient instability characteristics, necessitating specific control of the converter to enhance the parallel system's transient stability. To address the aforementioned issues, this paper proposes a transient stability enhancement control strategy based on dynamic current-limiting phase control for parallel systems, activated when the grid-forming converter enters current limiting. By dynamically regulating the current phase angle, the strategy enables active adjustment and dynamic allocation of active and reactive power from the grid-forming converter. This extends the fault critical clearance time of the grid-forming converter-synchronous machine parallel system, thereby enhancing the transient stability. By developing a grid-forming converter-synchronous machine parallel system simulation model on the PSCAD/EMTDC platform and applying a three-phase short-circuit fault to ground, the correctness and effectiveness of the proposed control strategy were validated.

When a single-line grounding (SLG) fault occurs in a small current grounding system (SCGS), the weak fault characteristics pose a challenge to the fault line selection (FFS). To improve the reliability of FFS, a new FFS method based on harmonic active injection of converter interfaced distributed generator (CIDG) is proposed in the paper. First, by analyzing the distribution law and attenuation characteristics of the harmonic current actively injected by CIDG in the system, the active injection characteristic harmonic current signal that can meet the needs of FFS is selected, and then, a characteristic harmonic current signal injection method based on CIDG active control is designed. On this basis, an FFS method based on the amplitude-phase relationship of harmonic current is proposed, enabling the method to simultaneously meet the requirements of fault line selection and phase selection requirements. PSCAD/EMTDC simulation analysis verifies the effectiveness and superiority of the proposed method compared to existing methods.

Gravity energy storage charges and discharges power/energy in the energy storage system by using height differences to raise and lower solid heavy energy storage media. The gravity energy storage system’s electric motor’s single machine capacity is limited, thus in order to satisfy the grid’s charging and discharging needs during significant power fluctuations, the system must employ many motors operating in parallel. In order to fulfill the power grid’s demands for energy charging and discharging, it is crucial to examine the power distribution strategy of each machine in the gravity energy storage system when many motors are operating in tandem. This study examines the operation state switching control technique of a gravity energy storage multi motor system that satisfies the grid’s power charging and discharging criteria in order to address this problem. This article builds a vertical matrix type gravity energy storage multi motor grid linked system model, two control strategies for varying the operating state of gravity energy storage multi-machine systems that adjust to various power fluctuation characteristics of the power grid are suggested, taking into account the power fluctuation characteristics of the power grid as well as the power/energy charging and discharging characteristics of the gravity energy storage system; In order to confirm the efficacy of the suggested control technique, a vertical matrix gravity energy storage system model was built using PSCAD/EMTDC, and the simulation results under two control strategies were compared and examined.

The integration of a large number of distributed resources into an active distribution network presents significant challenges, including high control dimensionality, strong output uncertainty, and low utilization of renewable energy. This paper introduces a distributed optimization strategy for networked microgrids based on network partitioning to alleviate the computational burden, reduce operating costs, and enhance the utilization of renewable energy. The active distribution network is partitioned into networked microgrids, and a two-layer distributed optimization model is developed for their management. The first layer focuses on intra-day distributed optimal dispatch, balancing power and load by managing various flexible resources and the exchange power between virtual microgrids. The second layer, real-time distributed power tracking optimization, coordinates flexible resources within virtual microgrids to mitigate photovoltaic power fluctuations and track intra-day dispatch instructions. Simulation results demonstrate that the proposed network partitioning method reduces dispatch costs by 5.3 % and increases the utilization of distributed PV by 3 %, compared to the NP method that only considering modularity. Moreover, calculation times for intra-day dispatch and real-time power tracking are reduced by approximately 26 % and 50 %, respectively, compared to centralized control.

Microgrids (MGs) typically exhibit low inertia, and interconnecting them into a microgrid cluster (MGC) can enhance both the stability and economic efficiency of system operation. To accommodate the increased system scale, hierarchical distributed control strategies are widely adopted for MGC management. While distributed control inherently supports plug-and-play functionality, limited research has addressed how to minimize the system fluctuations caused by MG integration and removal. This paper proposes a soft plug-and-play mechanism based on dynamic droop coefficients. By gradually adjusting the droop coefficient, a smooth transition in MG output power is achieved, thereby avoiding abrupt grid connection and disconnection. Building on this mechanism, an improved hierarchical distributed control strategy is developed for MGC. Simulation results demonstrate that the proposed strategy effectively mitigates the impact of microgrid plug-and-play on overall system stability.