Electric Power

An improved Transformer day-ahead electricity price forecasting model based on Patch mechanism and channel-independent structure

CHEN Zihong, HUANG Ningxin, LAI Zhihang, LAI Xiaowen, CHEN Xiaoting, CHEN Shuonan, GAO Feng

2026, 59(5): 1-8. DOI: 10.11930/j.issn.1004-9649.202506072

Asbtract ( )

PDF (1324KB) ( )

Figures and Tables | References | Related Articles | Metrics

To address the common problems of insufficient temporal feature extraction and poor adaptability to special day scenarios in day-ahead electricity spot market price forecasting, this paper proposes an improved forecasting model based on the Transformer architecture. The Patch mechanism is introduced to enhance local temporal feature extraction, and the channel-independent structure is combined to improve the learning efficiency of multivariate features. In addition, the multi-head attention mechanism is adopted to capture the global price fluctuation patterns. The proposed method is verified based on the historical data of the Guangdong electricity spot market. Compared with the baseline Transformer model, the mean absolute error (MAE) of the proposed model is decreased from 32.95 to 23.88 in weekend scenarios, and from 78.33 to 70.33 in holiday scenarios. The model exhibits significantly better adaptability to the phenomenon of quantity-price deviation than the baseline model, and can accurately capture the upward trend of price floors when the bidding space exceeds 60000 MW. The proposed model achieves a significant improvement in prediction accuracy under different scenarios (especially special scenarios) and has good adaptability to quantity-price deviations.

Identification of loss-risk zones of coal-fired power projects under the capacity price mechanism

YUAN Shuguang, ZHAO Xiaoxi, WANG Huaqing, LIU Jianye

2026, 59(5): 9-19. DOI: 10.11930/j.issn.1004-9649.202410092

Asbtract ( )

PDF (1759KB) ( )

Figures and Tables | References | Related Articles | Metrics

Guided by the capacity price policy, it is of great significance for building a clean, safe, economical and flexible new power system to optimize the market-oriented mechanism of coal-fired power projects on the basis of fully evaluating the loss-risk zones of coal-fired power projects in various regions. Taking three typical coal-fired power units as research objects, this paper proposes an improved coal-fired power regional levelized cost of electricity (LCOE) model by introducing the concept of equivalent units, so as to calculate and compare the LCOE of provincial coal-fired power units before and after the introduction of capacity price. On this basis, an investment return rate model for coal-fired power units is constructed, and the return-risk zones of regional coal-fired power units are calculated according to the current on-grid electricity price combined with the capacity price policy. The results show that the introduction of the capacity price mechanism has exerted a significant impact on the regional LCOE of coal-fired power, driving an overall declines in this cost level. Meanwhile, the average return range of regional coal-fired power units has been adjusted upward. Notably, the return levels of coal-fired power projects in western and northeastern China remain relatively low, indicating that such projects still face the possibility of losses during operation and their operational risks require attention. The research results can provide theoretical support for the formulation of regional coal-fired power prices and also serve as a decision-making basis for the market-oriented reform of coal-fired power.

Coordinated control strategy of electric-hydrogen coupled energy storage DC microgrid considering hydrogen storage state and DC bus voltage stability

XU Hengshan, CHAI Sen, WU Yangyang, LI Chenyang, ZHANG Yajian, MO Ruqiao

2026, 59(5): 20-32. DOI: 10.11930/j.issn.1004-9649.202412042

Asbtract ( )

PDF (2399KB) ( )

Figures and Tables | References | Related Articles | Metrics

To address the issues of bus voltage fluctuations and state of hydrogen (SOH) limit violations in electric-hydrogen coupled energy storage DC microgrids, a coordinated control strategy for electric-hydrogen coupled energy storage DC microgrids considering SOH and DC bus voltage stability is proposed. Firstly, a droop control strategy based on fuzzy algorithm is used for hydrogen energy storage system (HESS) to dynamically optimize the output power of HESS by comprehensively considering bus voltage fluctuations and SOH. Secondly, a control strategy based on analogous virtual synchronous generator (AVSG) is utilized for battery energy storage system (BESS) to simulate the charge-discharge characteristics of capacitors, so as to optimize the dynamic performance of DC bus voltage. Finally, considering the amplitude of bus voltage fluctuation, the operation modes of HESS and BESS are controlled to further coordinately optimize the stability of DC bus voltage. A simulation model of the electric-hydrogen coupled energy storage microgrid is built on the Matlab/Simulink platform to verify the effectiveness of the proposed strategy. The test results show that the proposed coordinated control strategy can improve the stability of DC microgrid bus voltage and alleviate the overcharge and overdischarge of the hydrogen energy storage system under source-load fluctuation scenarios.

Short-term load forecasting method for distribution networks based on transformer and ensemble learning

ZHANG Huaitian, JIA Dongli, WANG Shuai, HE Kaiyuan, REN Zhaoying, LIU Jiajing, HU Xuekai

2026, 59(5): 33-45. DOI: 10.11930/j.issn.1004-9649.202511063

Asbtract ( )

PDF (1864KB) ( )

Figures and Tables | References | Related Articles | Metrics

Against the backdrop of the new power systems, the penetration rate of distributed energy resources in distribution networks is rising steadily, and the load characteristics are becoming increasingly diversified. Existing short-term load forecasting methods thus fail to effectively capture the high-dimensional nonlinear temporal characteristics of load data. To address this issue, this paper proposes a short-term load forecasting method for distribution networks based on Transformer and ensemble learning. First, a multi-dimensional feature embedding layer is constructed to fuse the temporal and periodic characteristics of loads as well as environmental variables. Second, a multi-head self-attention mechanism is adopted to establish dynamic cross-time interval correlations, thereby extracting the spatiotemporal coupling characteristics of loads accurately. Third, a hierarchical randomized feedforward network is designed, with the Dropout technique integrated to enhance the multimodal representation capability of the model’s latent space. Finally, multiple differentiated Dropout-based models are ensembled, and Bayesian evaluation of forecasting uncertainty is realized through sampling with multiple forward propagations. Experimental results demonstrate that the proposed method outperforms state-of-the-art benchmark models in both forecasting accuracy and stability, and can thus provide effective technical support for the optimal dispatching of distribution networks.

Probabilistic load prediction based on gated spiking neural P system model

SUI Zeyuan, WANG Jun, PENG Hong, WANG Delin, SONG Ge

2026, 59(5): 46-56. DOI: 10.11930/j.issn.1004-9649.202505075

Asbtract ( )

PDF (1208KB) ( )

Figures and Tables | References | Related Articles | Metrics

Conventional deterministic load forecasting fails to provide uncertainty information of loads, while probabilistic load forecasting can generate probability distributions of predicted value uncertainty, thus providing comprehensive information for power grid dispatch decisions. In order to further improve the accuracy of probabilistic load forecasting, this paper proposes a model, which incorporates the least absolute shrinkage and selection operator (LASSO) and gated spiking neural P system (GSNP). Firstly, LASSO is used to extract the key features from external features such as minimum temperature, maximum temperature, average temperature, average humidity and precipitation. Subsequently, an improved GSNP model is developed to implement probabilistic load forecasting, enhancing the performance of long-term time-series forecasting. The case study using two long-term time-series datasets at different scales shows that the proposed model outperforms several other typical models in terms of both prediction accuracy and prediction interval quality.

Consider the fault recovery strategy of active distribution network for UAV recovery of temporary communication

TANG Jiajie, YUAN Song, LI Dian, ZHOU Rong, MEI Jiabao

2026, 59(5): 57-66. DOI: 10.11930/j.issn.1004-9649.202510004

Asbtract ( )

PDF (1293KB) ( )

Figures and Tables | References | Related Articles | Metrics

Aiming at the problem of large-scale information blind spot load recovery caused by communication interruption, an active distribution network fault recovery strategy considering the temporary communication restoration of unmanned aerial vehicle (UAV) is proposed. Firstly, under the constraint of limited UAV resources, a rapid recovery model of the communication network with the goal of maximizing the high-priority load coverage of the blind zone is constructed, and the optimal deployment location of the UAV and the configuration scheme of emergency communication equipment required in each blind area are optimized. On this basis, by coordinating the dispatchable resources in the region, the isolated island division and network topology reconstruction are implemented simultaneously to realize the collaborative optimization of UAV deployment scheme and emergency repair scheduling, so as to maximize the amount of fault load recovery. Finally, the effectiveness and superiority of the proposed fault recovery strategy are verified by simulation experiments, and the complete recovery time of the proposed policy is shortened by 60% compared with the fault recovery strategy of manual communication network maintenance.

Abnormal data detection method for distribution networks in data scarcity scenarios

ZENG Ruijiang, LI Zhiyong, HUANG Shu, WANG Weiguang

2026, 59(5): 67-75. DOI: 10.11930/j.issn.1004-9649.202510085

Asbtract ( )

PDF (1604KB) ( )

Figures and Tables | References | Related Articles | Metrics

In order to accurately detect abnormal voltage and current data in the distribution network and solve the problem of low accuracy of the detection model caused by the scarcity of abnormal data under normal operation of the distribution network, a method for detecting abnormal data based on an improved chaos optimization algorithm (ICEO) - dual attention mechanism Transformer (DAM Transformer) is proposed. This method first utilizes the strength controlled diffusion anomaly synthesis (SDAS) method to generate partial anomaly data, in order to alleviate the problem of insufficient model recognition accuracy caused by the scarcity of real anomaly samples; Secondly, an innovative DAM Transformer model was proposed, which integrates a dual attention mechanism to achieve collaborative modeling of complex patterns in different time scales and feature spaces, effectively improving the identification of multi-scale feature coupling relationships in the context of abnormal distribution network data; Finally, ICEO was used to iteratively optimize the hyperparameters of DAM Transformer, further improving the optimization efficiency and generalization performance of the model in complex scenarios. The results show that compared with traditional models, this method improves the accuracy of identifying abnormal voltage in distribution networks by 12.81% and the accuracy of identifying abnormal current by 12.22%. In data scarcity scenarios, the recognition accuracy is significantly better than traditional models. This method effectively solves the core bottleneck of sample scarcity and difficulty in modeling multi-scale features in abnormal data recognition of distribution networks, improves the accuracy of abnormal recognition and the stability of model operation, and provides key technical support for digital inspection, real-time fault warning, and operation and maintenance decision optimization of intelligent distribution networks. It has engineering application prospects.

Adaptive secondary frequency regulation strategy with coordinated virtual power plants

WANG Zesen, WANG Xuanyuan, KONG Shuaihao, SUN Bohao, JI Zhen, SUN Wei, ZHANG Yajian, ZHANG Jiafang, PENG Chen

2026, 59(5): 76-85. DOI: 10.11930/j.issn.1004-9649.202509016

Asbtract ( )

PDF (2845KB) ( )

Figures and Tables | References | Related Articles | Metrics

Virtual power plant (VPP) can effectively enhance the frequency resilience of the power systems by collaboratively scheduling energy storage resources through communication networks to participate in secondary frequency regulation (SFR). However, the dynamic access/exit of energy storage resources may lead to structural uncertainties in SFR system. A SFR strategy that considers dynamic access/exit of energy storage resources in VPP has been proposed. Firstly, the SFR system is modeled as a set of subsystems depending on the participation states of energy storage resources. Secondly, to reduce communication redundancy, a dynamic event-triggered communication scheme is proposed, which only triggers information exchange when the SFR performance degrades below a preset threshold. Thirdly, considering the asynchronous switching characteristics of communication delay causing control parameters to regulate behind the participation behavior adjustment of energy storage resources, design constraints and update criteria for control parameters have been obtained based on the average dwell time (ADT) technique. Simulation results have shown that compared with existing periodic triggering control schemes, the proposed scheme can save over 50.63% of network bandwidth occupancy and SFR device operation actions.

Coordinated optimization operation model of virtual power plant and distribution network in the electricity-carbon coupled market

LI Na, ZHANG Qiang, HAO Yi, WU Guannan

2026, 59(5): 86-96. DOI: 10.11930/j.issn.1004-9649.202512012

Asbtract ( )

PDF (1485KB) ( )

Figures and Tables | References | Related Articles | Metrics

To address the challenges of high-dimensional nonlinearity, multi-agent privacy protection, and computational efficiency in the collaborative optimization of VPP and distribution networks under an electricity-carbon coupled environment, this paper proposes a distributed collaborative operation model based on epigraph theory. First, a framework for electricity-carbon coupled market operations is established, incorporating the modeling of diverse distributed energy resources, including renewable energy units and carbon capture systems. Secondly, a quantitative bi-level mathematical model is formulated: the upper level aims to maximize the comprehensive benefits of the VPP within the electricity-carbon market, while the lower level focuses on minimizing system power purchase and carbon quota costs, with clearly defined constraints for each level. Finally, by leveraging epigraph theory, the original problem is converted into an efficient distributed solving form through function transformation and equivalent projection modeling. This approach effectively bypasses nonlinear constraint bottlenecks while safeguarding the privacy of participating entities. Simulation results on the IEEE 33-node system demonstrate that the proposed method achieves a 3.04% increase in total operating profit compared to traditional bi-level model solutions, validating its effectiveness in enhancing both economic performance and computational efficiency.

Design of global relay protection clock desynchronization intelligent verification system based on waveform recording data

SHI Hengchu, CHEN Xiaofan, YOU Hao, HU Xiao, XU Shoudong, GUAN Yuanpeng

2026, 59(5): 97-108. DOI: 10.11930/j.issn.1004-9649.202512042

Asbtract ( )

PDF (1381KB) ( )

Figures and Tables | References | Related Articles | Metrics

Time synchronization of protective relay devices is essential for post-fault incident analysis and action tracing in new power systems. To address the limitations of incomplete clock monitoring coverage, reliance on manual synchronization correction, and lack of a unified management platform, this paper establishes a global clock desynchronization intelligent verification system for protective relay devices based on waveform recording data. Leveraging the existing master-slave waveform recording architecture, the system deploys an online monitoring unit for relay protection clock desynchronization at the master station, which centrally accesses wave recording data from multiple substations. Two schemes are designed, namely time-phased remote startup monitoring and homologous data comparison monitoring, to realize online monitoring of clock-offset of global protection devices under both fault and non-fault conditions. On this basis, the system incorporates a Kalman filtering intelligent verification algorithm with clock-offset–drift rate as state variables, which filters and predicts multi-time observation results while considering the influence of measurement noise to generate more accurate deviation verification values. Field applications demonstrate that the system can effectively identify desynchronized relays and maintain the post-synchronization clock deviation of devices within 20 ms, which is conducive to post-incident analysis and operation maintenance of clock synchronization status under relay protection device clock desynchronization.

Capacitor voltage ripple suppression mechanism and control strategy for hexverter

SHU Jun, YI Yang, CHEN Zhengfeng, LU Zhongpeng, HUANG Yiwei, LIU Darui, ZHANG Yiming

2026, 59(5): 109-117. DOI: 10.11930/j.issn.1004-9649.202602001

Asbtract ( )

PDF (1826KB) ( )

Figures and Tables | References | Related Articles | Metrics

At low-frequency operating conditions, the submodule capacitor voltage ripple in Hexverter increases significantly, leading to higher capacitance requirements and resulting in increased converter volume and cost. To alleviate the design constraints on DC support capacitors, this paper proposes a capacitor voltage ripple suppression mechanism and control strategy. Based on the AC positive-sequence network model of the Hexverter, the charging and discharging power characteristics of the submodule capacitors are analyzed, and an analytical model of the capacitor voltage ripple is established. Furthermore, through the zero-sequence network model of the Hexverter, the quantitative relationship between the zero-sequence AC circulating current and the zero-sequence transferred power is derived, clarifying the voltage ripple suppression mechanism based on the injection of specific-order zero-sequence AC circulating currents. On this basis, combined with the conventional Hexverter control strategy, a compound control strategy for the Hexverter incorporating a ripple suppression loop is proposed to effectively suppress the submodule capacitor voltage ripple. Simulation results demonstrate that the proposed control strategy can significantly reduce the capacitor voltage ripple, verifying its feasibility and effectiveness.

Ultra-short-term power load forecasting based on dynamic weighting mixture of experts

ZHOU Zhuan, WANG Jie, BIAN Jiayu, YU Zhiyong, YUAN Tiejiang

2026, 59(5): 118-132. DOI: 10.11930/j.issn.1004-9649.202506007

Asbtract ( )

PDF (1356KB) ( )

Figures and Tables | References | Related Articles | Metrics

Ultra-short-term power load forecasting is a key supporting technology for real-time scheduling of new-type power systems, and its accuracy directly determines the consumption capacity of new energy, the economics of units combination, and the charging and discharging efficiency of energy storage system. To address the challenges of load data—including its strong temporal dependency, sensitivity to meteorological conditions, sensitivity to calendar effects, and anomalous fluctuations, a dynamic weighting mixture of experts (DW-MoE) model is proposed for ultra-short-term load forecasting. Firstly, the model captures the periodic temporal characteristics of the load sequences through BiLSTM, characterizes the nonlinear correlation between meteorological factors, date factors and loads using XGBoost, and achieves accurate detection of abnormal load patterns using GAN. Then, a dynamic weighting mechanism based on sliding window error feedback is designed to achieve adaptive fusion of multiple expert outputs; Finally, an online update mechanism is introduced to incrementally optimize the model’s parameters based on the latest sampled data, enhancing the dynamic response capability of the model to non-stationary load fluctuations. The experimental results demonstrate that compared to single models and traditional hybrid methods, the DW-MoE model exhibits significant advantages in both prediction accuracy and convergence speed for ultra-short-term load forecasting, and notably, it achieves a substantial reduction in prediction error under anomalous load scenarios, validating the model's robustness to abrupt load variations.

Study on stress distribution of distribution network arresters under complex operating conditions based on electro-thermal-mechanical coupling

WANG Zhenyu, SHI Junhao, WANG Wei, BAI Yinhao, AI Xueyong

2026, 59(5): 133-141. DOI: 10.11930/j.issn.1004-9649.202509024

Asbtract ( )

PDF (1680KB) ( )

Figures and Tables | References | Related Articles | Metrics

Zinc oxide arresters are the main devices for preventing lightning strike faults in distribution network lines. In the actual operation process, the operating conditions of arresters are complex and diverse. When subjected to various unbalanced forces caused by factors such as non-vertical installation methods, wind force and horizontal tensile force, the thermal stress superimposed by lightning impulse is an important cause of aging and failure of arresters. Considering the influence of various unbalanced forces and multiple lightning strikes, based on the principle of multi-field coupling of electricity, heat and force, a simulation model for calculating the operating parameters of arresters was established. The internal distribution of electricity, heat and force of the entire 10kV arrester under different working conditions was analyzed. The calculation results show that multiple lightning strikes increase the stress at the interface of zinc oxide resistance sheets inside the arrester by about 15% compared with a single lightning strike. Under the combined action of unbalanced force and lightning strike, the stress at the interface of zinc oxide resistance sheets increases by about 25% compared with the sum of the stresses when unbalanced force and lightning strike act separately. The mutual coupling of thermal stress and mechanical stress The phenomenon of nonlinear stress enhancement occurred. The research results provide a reference for analyzing the fault causes of surge arresters under complex operating conditions from the perspective of stress.

Online incremental power forecasting method for centralized photovoltaic power plants based on deep learning

YANG Chaoying, LI Huipeng, ZHAO Jun

2026, 59(5): 142-149. DOI: 10.11930/j.issn.1004-9649.202510014

Asbtract ( )

PDF (906KB) ( )

Figures and Tables | References | Related Articles | Metrics

Under the "dual carbon" goals, centralized photovoltaic (PV) power stations have become a crucial support for the new energy power syst however, PV power generation is strongly intermittent and volatile due to factors such as seasons and weather. Addressing the issue of model performance degradation caused by the dynamic evolution of input data in practical acations, and the susceptibility of traditional updating methods to catastrophic forgetting, this paper proposes a deep learning-based online incremental power prediction model. The model introduces the Deep Experience Replay (DER) incremenng mechanism to construct a dual-core framework of "block feature extraction and online knowledge retention". It captures multi-scale periodic features via a patch token strategy, utilizes self-attention mechanto mine multivariate dependencies, and combines experience replay techniques to alleviate catastrophic forgetting. Experimental results based on real-world data from a PV power station indicate that the cumulative accuracy degradation rate of the prsed model is significantly lower than that of traditional models, demonstrating stronger adaptability and generalization capabilities, and providing an effective solution for online dynamic power prediction in centralized.

A robust coordinated planning method for transmission system and storage considering analytical frequency dynamics constraints

CHEN Honglin, LUO Shuxin, XU Wei, WU Guangda, WANG Shichao

2026, 59(5): 150-163. DOI: 10.11930/j.issn.1004-9649.202511031

Asbtract ( )

PDF (1364KB) ( )

Figures and Tables | References | Related Articles | Metrics

To address the challenges posed by the uncertainty of offshore wind power output to power system frequency security and transmission network planning, this paper proposes a transmission-storage coordinated robust planning method considering analytical dynamic frequency constraints. First, a bi-level optimization framework is formulated, consisting of upper-level transmission network planning and lower-level operation scheduling; the upper level aims to minimize the total system investment and operation cost while planning transmission lines that satisfy load rate constraints, and the lower level considers the coordinated operation of the transmission network and energy storage system (ESS) with frequency response modeling to guarantee dynamic frequency security under disturbances. Then, the bi-level model is transformed into a tractable single-level model via strong duality theory. Furthermore, an adversarial scenario generation (ASG) method is employed to construct extreme risk scenarios, the probabilistic uncertainty of offshore wind power and load is characterized by a 1-∞ norm hybrid uncertainty set, and conditional value-at-risk (CVaR) is adopted to quantify the corresponding risk loss, based on which a transmission-storage coordinated distributionally robust optimization model is established and solved by the column-and-constraint generation (C&CG) algorithm. Simulation results on the IEEE 24-bus system demonstrate that the proposed method can enhance system frequency security, effectively alleviate line overloading, and yield planning schemes with favorable economic efficiency and robustness.

A grid-forming energy storage primary frequency regulation method considering SOC constraints and active frequency modulation coefficients

LI Yongbin, GESANG Jinmei, WU Jie, ZHU Ling, XU Guang

2026, 59(5): 164-175. DOI: 10.11930/j.issn.1004-9649.202512032

Asbtract ( )

PDF (1436KB) ( )

Figures and Tables | References | Related Articles | Metrics

With the decline of system inertia in modern power systems, frequency issues following disturbances have become increasingly severe, and the deployment of grid-forming energy storage has gradually emerged as an important means for primary frequency regulation. However, improper parameter settings of grid-forming energy storage may aggravate transient stability risks, and existing studies generally neglect the state of charge (SOC) constraints of the storage system itself. To address this issue, this paper proposes a primary frequency regulation method for grid-forming energy storage considering SOC constraints and active frequency modulation coefficients.First, a detailed mathematical model is established to describe the coordinated participation of grid-forming energy storage and synchronous generators in primary frequency regulation, and the system frequency response transfer function is derived. Second, based on trajectory sensitivity analysis, the key dominant parameters affecting transient frequency stability are identified, and a parameter optimization model is constructed with the objective of minimizing the maximum system frequency deviation. The model incorporates constraints on the SOC safe operating range, output limits, and the rate of change of frequency (RoCoF), and is solved using the sequential quadratic programming (SQP) algorithm. Finally, simulation results demonstrate that, under SOC and RoCoF constraints, the proposed method reduces the maximum frequency deviation by 43.3% and significantly raises the frequency nadir, thereby effectively enhancing the transient frequency stability of the power system.

Wind farm power forecasting by physical data fusion

ZHAO Jun, ZHANG Shifeng, SONG Jinge

2026, 59(5): 176-182. DOI: 10.11930/j.issn.1004-9649.202510007

Asbtract ( )

PDF (1775KB) ( )

Figures and Tables | References | Related Articles | Metrics

Power forecasting is a fundamental research topic in the wind power industry. Existing wind power forecasting methods predominantly rely on either data-driven or physics-driven approaches, with few studies combining physical models and data-driven techniques despite their significant complementary potential. A data-driven model is established using K-means clustering, empirical mode decomposition, and parallel weighted long short-term memory networks. A novel integrated approach combining physics-driven and data-driven methods was developed for wind farm forecasting. Validation using real-world data from a Chinese wind farm demonstrated that the proposed integrated method achieved 21.67% higher prediction accuracy than data-driven methods and 35.17% higher accuracy than physics-driven methods. These results confirm the superiority and reliability of physics-data fusion methods in wind farm ultra-short power forecasting.

Table of Content