To address frequency stability issues caused by reduced inertia level, weakened damping characteristics, and diminished frequency support capability in power-electronics-dominated power systems, this study proposes an adaptive coordinated frequency regulation strategy for MMC-HVDC to enhance frequency stability in asynchronous interconnected grids. For inertia support, a self-synchronized control strategy with adaptive SM switching is designed for MMC, which achieves synchronization without PLL while supporting receiving-end grid inertia, demonstrating significantly enhanced inertia support capability compared with conventional controls. For damping regulation and primary frequency control, a coordinated frequency support strategy considering frequency regulation dead band is proposed, where the dead-band parameters are designed based on the power regulation margin of the SG in the receiving-end grid, enabling adaptive switching between damping regulation and primary frequency control functions. This approach enhances frequency stability while minimizing the impact range of disturbances to the greatest extent possible. Furthermore, an equivalent circuit model of the MMC-SG dual-machine system is established based on electromechanical analogy principle, and the interaction mechanism between MMC and SG is quantitatively characterized, leading to the design of key control parameters. Finally, the real-time digital simulation model of MMC-HVDC system is built to verify the effectiveness of the proposed control strategy and model.
The dual-stator axial flux permanent magnet machines (AFPMMs) with sandwich rotor have attracted increasing attention owing to their advantages of low rotor eddy current loss and high operational efficiency. However, due to the three-dimensional (3D) structure of AFPMMs, the finite element method (FEM) imposes a substantial computational burden during the simulation process. To address this issue, a hybrid analytical model (HAM) that integrates subdomain method and magnetic equivalent circuit (MEC) approach is proposed in this paper to predict the magnetic field distribution in AFPMMs with sandwich rotor. By introducing virtual winding into the slot subdomain, the nonlinear behavior of the stator iron is effectively accounted for. The proposed HAM achieves higher accuracy in predicting the electromagnetic performance of AFPMMs with sandwich rotor under load conditions compared to the conventional subdomain model, while avoiding the complicated air-gap permeance modeling required by the MEC method. The accuracy and the effectiveness of the proposed HAM are validated through FEM and prototype experiments.
This study introduces a novel FOPD-PTID controller, which incorporates an enhanced filtering mechanism for load frequency control (LFC) in islanded AC microgrids. The proposed controller mitigates significant frequency deviations resulting from rapid load fluctuations, uncertainties in renewable energy sources (RESs), and failures in energy storage systems (ESSs). The analysis also examines the effects of electric vehicle (EV) integration. The Extended Stochastic Coati Optimization (ESCO) algorithm is employed to optimize controller gains. Comprehensive MATLAB/Simulink simulations evaluate the controller’s effectiveness and robustness under various operating conditions. Real-time OPAL-RT validation has been carried out for all operating scenarios to verify practical feasibility.
This paper proposes an active gate driver (AGD) circuit and a driving method for silicon carbide (SiC) metal-oxide-semiconductor field effect transistors (MOSFETs) that can reduce switching losses while maintaining low overshoot. Conventional AGD circuits suffer from complex sensing and driving circuits, difficulties in measuring the voltage and current during switching transients, and malfunction risks caused by control timing errors. To address these issues, an open-loop AGD based on a current-source gate driver is proposed. This study maximizes design efficiency by proposing a gate driving current selection method and timing model for determining current injection points to overcome limitations of the open-loop approach, specifically the need for extensive trial-and-error experiments to find optimal operating points. The proposed circuit and driving method are validated using double-pulse tests. The experimental results demonstrate a significant reduction in switching losses while effectively suppressing the voltage and current overshoot.
Focusing on switched reluctance motor (SRM) and synchronous reluctance motor (SynRM), this paper investigates and compares the harmonic content and torque generation mechanism of reluctance-type machines. Based on the air-gap field modulation theory, the general field modulation model for SRM and SynRM is proposed, which considers the winding width and magnetic circuit saturation. Then the modulating behavior and torque characteristics in SynRM are investigated by a new method carried out on the proposed rotor coordinate system. Based on this, the relationships between design parameters, spatial harmonics of MMF and torque, and time harmonics of torque are revealed. Furthermore, the field modulating behaviors in SRM and SynRM are investigated from the perspective of relative motion between original MMF and the modulator, rather than the traditional classification into synchronous and asynchronous modulations. Finally, the prototypes of the analyzed SRM and SynRM are manufactured and tested. The proposed modulation model is validated by both finite element analysis and experimental measurements.
With the rapid integration of renewable energy sources and flexible loads into the power grid, distribution networks face significant operational challenges. These include heightened operational risks and difficulties in achieving coordinated optimization between the planning and operation stages. To address these challenges, this paper proposes an integrated distribution network planning and operation framework for high renewable energy-penetrated distribution networks, based on dynamic causal traceability and multi-scale temporal correlation analysis. First, a multi-dimensional evaluation index system is established. Subsequently, the analytic hierarchy process (AHP) combined with the coefficient of variation (CV) method is employed to cross-validate planning and operational indicators, while a Granger causality network is applied to trace dynamic causal relationships among indicator deviations and identify root causes. Furthermore, a temporal convolutional network (TCN) is constructed to capture temporal dependencies and establish a closed-loop feedback mechanism for dynamically correcting planning schemes. Finally, the framework is validated on a practical distribution network. Simulation results indicate that the proposed methodology accurately assesses system performance and enhances planning-operation synergy. By employing this closed-loop framework, the prediction accuracy is improved to 0.955, and operational balancing costs are reduced by 38.9
This paper presents a novel 3D Flux Motor structure tailored for high-torque-density electric vehicle traction applications. The proposed topology integrates a toroidal winding and a four-way rotor arrangement to simultaneously guide both axial and radial magnetic flux within a compact and manufacturable design. This configuration expands the effective magnetic flux area, eliminates coil ends, and enhances torque generation through multidirectional flux paths. A key contribution of this study is the strategic suppression of leakage flux, particularly in the shaft direction. A radial inner rotor is introduced to increase the reluctance of undesired flux paths, thereby confining magnetic flux within the effective torque-generating loop. The magnetic equivalent circuit is formulated to analyze the internal flux distribution and quantitatively demonstrate the impact of flux path design on performance. The results confirm that the proposed configuration improves usable flux and torque density without increasing material volume or input current. This work proposes a design framework for compact, high-performance 3D flux machines with improved magnetic efficiency and leakage flux control, with practical feasibility to be further verified through experimental validation.
This paper proposes a low-complexity high-resolution direction of arrival (DOA) estimation technique based on adaptive grid definition algorithm. This method solves the problem of excessive computational amount of full multiple signal classification (MUSIC), which searches the entire angular range in a fine grid. The proposed technique firstly sets the region of interest (ROI) by evaluating the MUSIC spectrum with a coarse search for the entire angular range, and secondly performs a fine search only for the corresponding ROI to find a precise peak. In addition, unnecessary eigenvalue decomposition (EVD) calculation was prevented by calculating the noise subspace matrix calculated in the subspace decomposition process only once and reusing it for coarse and fine search. This reduces the computational complexity of the conventional full MUSIC while maintaining the same resolution as full MUSIC. As a result of the simulation, the proposed technique greatly reduced the amount of computation and delay while maintaining the root mean square error (RMSE) and resolution performance and proved its applicability in real-time and low-power environments.
The current power landscape has seen a significant increase in the integration of renewable sources, particularly photovoltaic (PV) and wind-based systems, within microgrids. While these technologies offer environmental and economic benefits, their inherent variability due to fluctuating solar insolation and wind speeds introduces considerable protection challenges. Such sporadic variations can lead to the malfunctioning of conventional relays based on pre-specified threshold level of current. So, in order to enhance the robustness of protection approach with immunity weather-related uncertainties, this work integrates a meteorological data-driven joint probabilistic model capturing the stochastic nature of solar insolation and wind speed and performs categorization task through Convolutional neural network (CNet). The adoption of CNet facilitates the identification of discriminative features from complex datasets while maintaining low computational cost. The proposed approach involves, recording the time-domain current-voltage data from the relaying buses and transforming them into image datasets which is further utilized to train the CNet models, for performing specific categorization or regression tasks. The comparative analysis with existing methods demonstrates that the proposed framework achieves superior speed (<1 cycle) and robustness against renewable variability while significantly reducing computational overhead. Furthermore, the approach has been validated in real time using an OPAL-RT OP4510 hardware platform, ensuring its practical applicability for next-generation microgrid protection.
This paper proposes a circulating current estimation method based on a Luenberger observer for use in subsequent protection decisions when an inter-turn short-circuit fault (ITSF) occurs in a motor winding. When an ITSF occurs, a circulating current is generated within the short-circuit loop, directly contributing to localized heating and further fault progression. Since this current is difficult to measure directly in practical motor drive systems, an estimation-based approach is required. In this study, assuming that the fault severity and faulted phase are provided through prior diagnosis, the circulating current is estimated using an ITSF model and a Luenberger observer. The proposed method is experimentally validated under healthy and two-turn short-circuit fault conditions. Under the healthy condition, the estimated circulating current remains near zero, whereas under the short-circuit fault condition, the estimated value generally follows the variation trend of the measured waveform. These results demonstrate that the proposed method can be utilized as a decision variable for subsequent output limitation in response to an incipient ITSF.
Climate extremes and natural disasters can cause large-scale power outages, thereby compromising the stability and reliability of power systems. To address these challenges, resilience-enhancing approaches such as network reconfiguration and microgrids have received increasing attention. However, traditional power systems often face limitations in operational efficiency and available resources. In contrast, electric vehicles (EVs) offer significant potential for enhancing system resilience due to their bidirectional power flow capability and relative independence from conventional power infrastructure. This paper presents a comprehensive review of EV-enabled power system resilience, with a particular focus on transportation–power system integration structures. Existing studies are systematically classified according to EV resilience functionalities, integration approaches, and operational characteristics. Network-based and communication-based integration frameworks are comparatively analyzed with respect to planning-oriented and real-time resilience operation. In addition, representative resilience evaluation metrics are reviewed to analyze their applicability and limitations under different operational scenarios. Based on the literature review, a conceptual framework is proposed to provide a unified perspective on EV-enabled resilience enhancement. Finally, the paper discusses the major implementation challenges and future research directions associated with practical EV-enabled resilience operation, including scalable coordination, interoperability, and cybersecurity.
This study presents a fuzzy-logic-based weather-adaptive control strategy for photovoltaic (PV)-integrated energy storage systems (ESS) to improve economic efficiency in net-zero-energy buildings (NZEBs). Because solar radiation is highly weather-dependent, existing optimization-based and AI-driven ESS control methods often rely on accurate system models, extensive training data, or high computational effort, which limits their adaptability under changing weather conditions. To address this gap, the proposed method converts short-term solar-radiation forecasts into fuzzy membership values and determines a dynamic target state of charge (SOC) through defuzzification. An ESS scheduling model incorporating Korea Electric Power Corporation (KEPCO)’s time-of-use rates then generates charge–discharge commands. The method is validated using solar-generation and building-load data from the KAIST Naepo Mobility Research Institute from September 2024 to March 2025, under three representative weekly scenarios: high radiation in early September, mixed conditions in mid-November, and low-radiation snowfall in mid-January. Results show smoother ESS current profiles, reduced unnecessary cycling, up to 30
This paper proposes a probabilistic operational evaluation framework to quantitatively compare and analyze the system impacts of smart inverter control modes in distribution systems with high penetration of distributed energy resources. Based on historical data, the probabilistic characteristics of time-varying load demand and photovoltaic generation are modeled on a monthly and hourly basis, and probabilistic power flow analysis is performed by combining these models with operational scenarios that include variations in substation sending voltage. Fixed power factor, Volt–Var, Volt–Watt, Watt–Var control modes, as well as their coordinated operation, are applied under identical probabilistic conditions to enable a consistent comparison of system impacts among control strategies. For each scenario, probabilistic power flow analysis is conducted, and when overvoltage or generation-induced line capacity violations occur, a curtailment algorithm is applied to account for system constraints. The results for each scenario are aggregated using occurrence probabilities as weighting factors and are quantitatively evaluated using system impact indices including voltage violations, line overload, system losses, reactive power control effort, and curtailment amount. These indices are further integrated into a normalized composite score to support systematic comparison among control strategies. The effectiveness and applicability of the proposed framework are validated through case studies conducted on multiple practical distribution feeders with different operating characteristics, including a sensitivity analysis of Volt–Var control parameters based on IEEE Std. 1547–2018 to assess the influence of parameter variations on the probabilistic performance indices.
This study proposes and demonstrates an integrated decision-support platform that combines real-time trend analysis with AI-based menu recommendations to support small-scale dessert F B entrepreneurs facing limited information accessibility and analytical resources. The core technical contribution lies in the establishment of a framework that suppresses the hallucinations of generative AI by integrating a Retrieval-Augmented Generation (RAG) pipeline with trend quantification algorithms ( VPD, ER ), thereby proposing menu concepts based on objective metrics. Technical performance evaluations revealed that the system reduces trend analysis time by approximately 99.8
This paper proposes an iterative inertia-constrained unit commitment (UC) framework that explicitly incorporates system inertia into power system scheduling to ensure frequency-secure operation in low-inertia power systems. The proposed approach employs the System Frequency Response (SFR) model to perform dynamic frequency analysis based on hourly UC results and the largest generator contingency, without requiring linearization or convexification of the swing equation. An iterative procedure is developed to assess the adequacy of procured inertial energy and to generate inertia constraints when frequency nadir requirements are violated. The effectiveness of the proposed framework is validated on a modified benchmark power system, demonstrating its capability to mitigate weak-grid issues while revealing the trade-off between economic efficiency and system security. Additionally, the Rate of Change of Frequency (RoCoF) is evaluated as a complementary inertia-specific metric, confirming that the proposed framework simultaneously improves both frequency nadir and RoCoF across all scheduling periods.
The growing integration of distributed renewable generation has increased the need for flexible assets that can balance variability while ensuring economic efficiency across electricity markets. In this context, photovoltaic–battery energy storage systems (PV–BESS) have emerged as a key enabler for local energy management and price-risk mitigation. This study develops a data-driven techno-economic framework to evaluate PV–BESS performance under uniform-pricing electricity markets. The framework couples measured feeder-level demand, real PV generation, and hourly day-ahead market prices within the System Advisor Model (SAM) environment, providing an integrated assessment of operational dynamics and financial outcomes. By combining empirical data with SAM’s built-in dispatch and financial modules, the approach reproduces realistic market interactions without relying on synthetic inputs or external optimisation routines. Results demonstrate that system profitability and investment feasibility are jointly governed by the interplay of market price dynamics, self-consumption behaviour, and storage configuration. The proposed framework offers a transparent and reproducible analytical basis for evaluating distributed storage deployment and delivers quantitative insights to guide investment strategies and policy design toward resilient, cost-effective, and decarbonised electricity systems.
This paper proposes a speed-adaptive DC-offset-based open-switch fault diagnosis method for a dual three-phase inverter driving an asymmetric six-phase permanent magnet synchronous motor (PMSM). Open-switch faults interrupt half-cycle conduction of the faulty phase, thereby distorting the phase current and inducing torque ripple. This conduction loss produces a fault-induced DC offset in the faulty phase current and a second-harmonic ripple in the dq-axis currents. Based on these analytical characteristics, the proposed method directly extracts the DC offset from the measured phase currents using an integrator combined with a speed-adaptive high-pass filter (HPF), without requiring additional sensors or signal normalization. The cut-off frequency of the HPF is designed to be proportional to the electrical angular frequency, ensuring consistent detection performance over a wide operating speed range. Furthermore, the sign and magnitude of the accumulated detection variable enable both faulty phase localization and identification of upper- or lower-switch open faults. Theoretical analysis shows that the fault-induced DC offset can be observed within one electrical period; however, for reliable detection, the proposed method accumulates the detection variable over multiple electrical periods with a periodic reset operation. The effectiveness of the proposed method is validated through experimental results under both steady-state and transient operating conditions.
With the implementation of Chinese “dual carbon” policy, virtual power plants (VPPs) have become a key instrument in modern power systems for integrating distributed renewable energy, enhancing grid flexibility, and promoting low-carbon development. However, the operational efficiency of a VPP is highly dependent on optimal resource allocation during the planning phase, particularly in electricity–carbon coupled markets where economic and environmental objectives must be balanced. To this end, this paper proposes a novel market-oriented VPP resource allocation strategy that incorporates electricity–carbon coupling mechanisms. A bi-level framework is established, in which the upper level determines the VPP resource allocation strategy and the lower level simulates prosumer operations. The resulting model is solved by an improved Analytical Target Cascading (ATC) algorithm that innovatively introduces adaptive penalty factors to enhance convergence. Case studies demonstrate the feasibility and superiority of the proposed strategy in achieving synergistic optimization of economic benefits and low-carbon objectives.
Based on the optimal layout and identified insulation defect information, the method of locating insulation defects in high-voltage cables was studied in depth. Initially, an equivalent model of insulation defects was formed by establishing a mathematical model of the cable and analyzing the characteristics of insulation defects. Furthermore, a positioning method based on a double π-type equivalent circuit is introduced, which uses the voltage continuity of the insulation defect phase to achieve precise positioning of the defect by measuring the voltage and current at both ends. Only the defective phase cable is calculated, which reduces the calculation quantity and improves efficiency. The accuracy and feasibility of this method are verified through cases, showing good positioning accuracy under different arrangements, lengths, defect levels and load levels, with the maximum error not exceeding 1
The rising reliance on fossil-fuel engines has accelerated CO2 emissions, driving the shift to electric vehicles (EVs). While EV adoption reduces emissions, uncoordinated charging stresses active distribution networks (ADNs), causing voltage deviations, higher line losses, congestion, and reduced stability. To address these challenges, we propose a Bayesian network–based probabilistic framework for optimal EV charging station (EVCS) placement under multiple uncertainties. The model uses voltage deviation factor (VDF), location demand factor (LDF), and proximity to grid (PG) as conditional parameters to assess bus suitability. Distributed generation (DG) and demand response management (DRM) are integrated to enhance flexibility and reliability across scenarios. Concurrent optimization employs Rule-Based Fuzzy Logic (RBFL) with a Probabilistic Genetic Algorithm (PGA) to optimize DG sizing, placement, and DRM-based peak shifting. The method is validated on a real-time 23-bus Indian utility system and an IEEE 33-bus system. Results show the RBFL–PGA approach achieves up to 22.13
