• Probabilistic Prediction of Overhead Line Ampacity Based on Multi-error Characteristic Fusion

  XU Bin, LIU Ren, TANG Bo, WANG Shuai

  Available online:April 23, 2025  DOI: 10.7500/AEPS20240616001

  Abstract(6) HTML(6) PDF(5.13 M)( 14)

  Abstract:Accurate prediction of overhead line ampacity serves as the foundation for achieving dynamic line rating. Conventional single-point prediction models face challenges in effectively capturing the volatility and randomness of meteorological factors influencing ampacity (e.g., wind speed and ambient temperature), while ampacity prediction errors exhibit time-varying correlations and conditional distribution characteristics. To address these limitations, this paper proposes a novel framework employing serial ensemble learning-based Bagging integrated with parallel ensemble methods (AdaBoost, XGBoost, and LightGBM) to develop a robust point prediction model. This hybrid approach overcomes the generalization constraints of single models while enhancing prediction accuracy and stability. Through analysis of measured data, this paper demonstrate the time-varying correlations and conditional distribution patterns inherent in ampacity prediction errors. Subsequently, this paper introduces Sklar's theorem and dynamic Copula functions to establish a conditional distribution model for prediction errors. By integrating the proposed point prediction model with this error distribution framework, this paper develops a probabilistic ampacity prediction method that integrates serial-parallel ensemble learning and multi-error characteristic fusion. Compared with traditional methods, the conservativeness indicator of the proposed method is reduced by 28.69%.

• Data-driven Oscillation Mode Prediction Scheme and Resonance Mitigation Analysis

  DING Jiong, ZHU Jiebei, ZHANG Miao, BIAN Yinan, YU Lujie, JIA Hongjie

  Available online:April 23, 2025  DOI: 10.7500/AEPS20240717004

  Abstract(8) HTML(6) PDF(4.57 M)( 17)

  Abstract:To assess the small-signal stability of new power system rapidly and mitigate potential mode resonances, a data-driven oscillation mode prediction (DOMP) scheme based on measurement signals and operation scenario information is proposed. First, based on system oscillation modes of the multi-channel measurement signal identification system in historical operating scenarios, the issue of data sources during model training is resolved. Then, based on deep extreme learning machine algorithm, a system oscillation mode prediction model is established, which takes scenario information from historical data as input and identification results of oscillation modes as output to train and evaluate the accuracy of the prediction model, thereby improving the prediction accuracy of the DOMP scheme. Based on the mode prediction results of DOMP, the system control parameters are optimized to avoid mode resonances during operation, and improve the small-signal stability of the system. Through the IEEE 39-bus model, it is verified that the proposed DOMP scheme can quickly and accurately predict the oscillation mode of the system in future scenarios based on scenario information, and then mitigate the mode resonance generated during system operation through parameter optimization to improve system stability.

• Planning of Fractional Frequency Transmission for Offshore Wind Power Considering Resonance Stability and Reliability of System

  CAI Zhongqi, WEI Chengxiao, WANG Xiuli, YANG Kun, ZHANG Haitao

  Available online:April 23, 2025  DOI: 10.7500/AEPS20240408004

  Abstract(5) HTML(4) PDF(3.97 M)( 14)

  Abstract:Fractional frequency transmission provides a new option for large-scale, long-distance offshore wind power delivery by reducing the frequency and improving the line transmission capacity. Aiming at the impact of a high proportion of renewable energy and power electronics equipment on the resonance stability and reliability of the system, close to the actual engineering needs, a method of planning of fractional frequency transmission of offshore wind power is proposed. Considering economy, resonance stability, and reliability, the method plans fractional frequency grid connection plan for offshore wind power. Firstly, the fractional frequency transmission system of offshore wind power and industrial frequency power grid planning model, resonance stability evaluation model, and reliability evaluation model are introduced. Secondly, taking the offshore wind power development in a certain coastal area of China as a case study, wind power transmission and industrial frequency power grid expansion schemes are planed, and the engineering verification is carried out. Furthermore, the technical economy of three schemes of power frequency AC, flexible DC, and fractional frequency transmission are compared, and the optimal solution is selected. The results show that in the case scenario, the technical economy of fractional frequency transmission have certain advantages.

• Bidding Strategy of Virtual Power Plant Clusters for Joint Market of Electric Power Energy and Frequency Regulation

  LI Zhenkun, ZHANG Zhaoke, LI Jingyue, ZHANG Zhiquan, TIAN Shuxin

  Available online:April 22, 2025  DOI: 10.7500/AEPS20240910006

  Abstract(24) HTML(16) PDF(2.10 M)( 43)

  Abstract:Multiple virtual power plants participating in the joint market as a cluster can enhance their market competitiveness. To solve the resource allocation of the cluster in the multi-market coupling decision-making and balance the electric power energy and the frequency regulation revenues, this paper proposes a bidding strategy for virtual power plant clusters participating in the electric power energy market and frequency regulation market. Firstly, an internal transaction model between the virtual power plant and the cluster operator is proposed, and a pricing strategy for the cluster operator and a method for refining the response behavior of the virtual power plant when participating in the external joint market are designed. Secondly, an operation cost model of virtual power plant considering the electric power energy and frequency regulation capacity transaction is established to realize multi-product transactions in the cluster, providing a foundation for the operator to formulate bidding strategies in the joint market. Then, the price-quota curve is used to characterize the relationship between clearing results of the frequency regulation market and the bidding strategy, and a risk-averse optimal bidding strategy is proposed with the objective of maximizing cluster revenues. Finally, a simulation analysis is conducted on a cluster consisting of three virtual power plants. The case study results demonstrate that the proposed joint bidding strategy stimulates the response willingness of each virtual power plant and improves the economic benefits of the cluster.

• Adaptive Short-term Wind Power Forecasting for Extreme Weather Based on Transfer Learning and AutoEncoder

  LI Yujia, CHEN Fuhao, YAN Jie, GE Chang, HAN Shuang, LIU Yongqian

  Available online:April 22, 2025  DOI: 10.7500/AEPS20240429012

  Abstract(25) HTML(20) PDF(2.15 M)( 35)

  Abstract:Aiming at the problem of insufficient accuracy of wind power forecasting under extreme weather events such as cold waves, typhoons and icing, an extreme weather event discrimination method based on meteorological factors is proposed, and an adaptive short-term wind power forecasting method for extreme weather events based on transfer learning and autoencoder is proposed. Firstly, by analyzing the coupling characteristics between meteorological factors and wind turbine output, the extreme weather discrimination standards are defined to identify the types of weather events that will occur in the future. Secondly, based on the autocorrelation mechanism of autoencoder forecasting model to increase the utilization of long time series information, the “pre-training-fine-tuning” strategy of transfer learning is adopted, whereby sufficient samples under normal weather are used for pre-training the forecasting model, and then fine-tuning is carried out for the limited samples of extreme weather data, and the forecasting model is adaptively applied to the short-term wind power forecasts according to the discrimination of extreme weather events. The datasets from 12 wind farms are selected for analysis, and the effectiveness of the proposed method is verified by analyzing the forecasting performance of the model under extreme weather and all kinds of weather conditions. The experimental results show that the proposed method can accurately forecast whether an extreme weather event will occur in the future and substantially improve the accuracy of short-term wind power forecasting under extreme weather, and the average forecasting accuracy of the proposed method under three kinds of extreme weather is 87.52%, which is an average improvement of 2.47% compared with conventional methods.

• Flexibility Demand Decomposition and Operation Optimization of Power System Based on Hybrid Mathematical Morphology

  LUO Renjie, JU Jiaxin, LI Zhiyi

  Available online:April 22, 2025  DOI: 10.7500/AEPS20240906008

  Abstract(21) HTML(20) PDF(1.73 M)( 37)

  Abstract:Considering the rapidly increasing flexibility demand characterized by multi-timescale fluctuations in the new power system, the traditional separate quantization method for the regulation capacity such as frequency regulation and ramp may cause excessive capacity reservation. It is urgent to explore the resource regulation performance and quantify the system flexibility demand to improve the safe operation ability of power systems. This paper proposes a flexibility demand decomposition of power system and multi-scale regulation capacity reservation method based on hybrid mathematical morphology decomposition. Firstly, the kernel density distribution model is used to generate the set of flexibility demand curves within the prediction interval. Secondly, the hybrid mathematical morphology filter considering the output characteristics of regulation resources is designed to decompose the flexibility demand curves into regular and random components. The needed frequency regulation capacity is then calculated according to the set of random components. Finally, based on the regular components obtained by the flexibility demand decomposition, the stochastic optimization model is constructed for long- and short-term ramping capacity reservation of regulation resources. Numerical results show that compared with traditional methods, the proposed model can match the decomposition results with the output characteristics of regulation resources. Meanwhile, the reservation scheme can meet the flexibility demand of power systems and enhance the capacity utilization rate.

• Bi-level Optimization Strategy of Grid-side Multiple Energy Storage Stations Considering Available Capacity and Frequency Modulation Cost

  LI Junhui, SONG Qinglong, GUO Qi, YAN Jun, LI Cuiping, HAO Qianpeng

  Available online:April 18, 2025  DOI: 10.7500/AEPS20240327002

  Abstract(49) HTML(30) PDF(2.99 M)( 76)

  Abstract:Aiming at the problem that some energy storage stations in the region lose charge/discharge capacity in advance due to uneven distribution of frequency modulation instructions, resulting in insufficient response ability of subsequent frequency modulation demand, a bi-level optimization strategy of grid-side multiple energy storage stations considering available capacity and frequency modulation cost is proposed. The proposed strategy includes the frequency modulation power distribution layer of energy storage station and the frequency modulation power distribution layer of energy storage unit: the upper layer introduces the available capacity state coefficient and the adaptive coefficient about the adjustment ability of energy storage resources, and designs the optimal distribution model of frequency modulation power partition of energy storage station to realize the balanced distribution of frequency modulation power in each energy storage station. The lower layer includes the health of state (SOH) and resistance coefficient, and the balanced recovery of the state of charge (SOC) of the energy storage unit constitutes the objective function, so that the energy storage unit with high SOH has the priority to restore the SOC. The simulation results show that the proposed strategy can improve the response ability of the energy storage resources in the region and realize the maximum utilization of energy storage resources while ensuring a certain economy.

• Transient Stability Assessment of Power System with Renewable Energy Based on Network Energy Trajectory Feature Learning

  ZHANG Yuchi, CAI Guowei, LIU Cheng, ZHANG Zedong, WANG Yingyue, YANG Jingying

  Available online:April 18, 2025  DOI: 10.7500/AEPS20240723007

  Abstract(40) HTML(24) PDF(1.99 M)( 65)

  Abstract:Rapid and accurate transient stability assessment is the key to ensure the safe and stable operation of power system. The high proportion of power electronic equipment connected to the grid leads to the non-smooth system characteristics of multi-control switching after the power system is disturbed, which brings challenges to the accuracy and generalization of the existing transient stability identification methods. A transient stability assessment method based on network energy trajectory feature learning is proposed for the grid-connected power systems of renewable energy and DC transmission. Firstly, the transient energy function of power system with renewable energy based on augmented network data is constructed, the mapping relationship between the temporal and spatial distribution characteristics of network energy and transient stability is studied, and the branch vulnerability index and stability discrimination index are proposed. Then, attention weights are guided by branch vulnerability index to enhance key branch features. Using stability evaluation function based on network energy and branch phase angle difference as input features, a transient stability assessment model is proposed. Finally, through simulation analysis of New England 10-machine 39-bus system and CEPRI-TAS practical system with 197 bus, the impact of power electronic equipment switching control on stability assessment methods is analyzed, and the accuracy and effectiveness of the proposed model are verified.

More+

Volume 49,2025 Issue 7

>专辑：大规模新能源基地稳定运行与直流送出技术

• Transient Control Method for Grid-forming Direct-drive Wind Turbine Unit Based on DC Voltage Synchronization

  DENG Zhenyan, WANG Han, ZHOU Shaoze, CAO Yunfeng, Wang Wei, CAI Xu

  2025,49(7):2-12, DOI: 10.7500/AEPS20240228004

  Abstract(112) HTML(81) PDF(3.43 M)( 86)

  Abstract:The power system with high proportion of renewable energy and high proportion of power electric equipment presents the characteristics of weak inertia and low short-circuit ratio, and the grid-following renewable energy cannot independently provide the transient frequency and voltage support required by the system. The adoption of grid-forming control is regarded as the key to solving this problem. Grid-forming converters manifest voltage source characteristics. However, due to the limited overcurrent capability of power devices, controlling transient currents during faults poses a challenge. Full-power direct-drive wind turbine units can employ DC voltage synchronization to achieve grid-forming, and an additional current controller is applied to suppress transient currents for fault ride-through. Nevertheless, this method still experiences impact currents during fault occurrence and recovery moments, risking the loss of transient support capability due to protection control strategy for converters. In response to this issue, by analyzing the short-circuit current generation mechanism and power response mechanism of grid-forming full-power wind turbine units using DC voltage synchronization during grid faults, an improved transient control method is proposed. The method employs a voltage vector limiter to achieve rapid matching of internal potential with fault voltage to limit the steady-state component of short-circuit current. Additionally, a nonlinear virtual resistor is introduced to achieve fast control of internal potential compensation to suppress the transient component of short-circuit current. Furthermore, during faults, the synchronization loop and reactive power loop are blocked to maintain the voltage source characteristics of the unit and achieve autonomous power response. The effectiveness and feasibility of the proposed method are verified through simulation and experimental validation.This pFoundation roject is supported by National Key R&D Program of China (No. 2022YFB2402703) and State Grid Corporation of China (No. 52272222001J).

• Quantitative Analysis of Transient Synchronization Stability of Virtual Synchronous Generator Considering Damping Effect

  ZHANG Jingyi, TIAN Zhen, LI Xilin, HUANG Meng, CAO Kan, ZHA Xiaoming

  2025,49(7):13-24, DOI: 10.7500/AEPS20231225001

  Abstract(97) HTML(50) PDF(3.75 M)( 81)

  Abstract:With the large-scale integration of renewable energy generators, the power system faces a severe risk of transient synchronization instability. The grid-forming converter controlled by the virtual synchronous generator (VSG) simulates the dynamic characteristics of the synchronous generator to achieve synchronization with the power grid, which has good frequency and voltage support ability. In order to accurately quantify the effects of virtual inertia and damping on the transient synchronization stability of VSG, this paper proposes a transient synchronization stability analysis method of VSG based on the iterative equal area criterion. Based on the principle of energy conservation, an iterative algorithm is designed to obtain the frequency-power angle distribution function of the converter under critical stability conditions. Thus, the effects of VSG inertia and the work done by the damping term on the acceleration and deceleration areas of the transient process are accurately quantified. By comparing the sizes of transient stability regions with different controller and grid parameters, the influence of parameters on the transient stability boundary is quantified. Finally, the MATLAB/Simulink simulation and the RT-LAB hardware-in-the-loop experiments are used to verify the effectiveness and superiority of the proposed stability analysis method. The conservatism of the proposed iterative equal area criterion is extremely low, and can realize the accurate and fast calculation of the transient stability boundary.

• Transient Synchronization Stability Analysis of Multi-virtual-synchronous-generator System Considering Controller Limiting

  LU Yiyuan, LI Yujun, YANG Dongmei, MU Tongpeng, LIU Jingrui, LIU Mengqi, HUI Jiaqi, DU Zhengchun

  2025,49(7):25-34, DOI: 10.7500/AEPS20240330001

  Abstract(106) HTML(56) PDF(1.79 M)( 83)

  Abstract:To protect power electronics with limited withstand voltage and current capability, virtual synchronous generators (VSGs) are typically configured with multiple controller limits, which are inevitably triggered during the transient process. This paper analyzes the transient synchronization stability of embedded power systems with multi-VSG incorporating current and frequency limiting, and proposes a control method to ensure the system stability. First, the dynamic model of a multi-VSG system with current limiting is established, and the transient energy function (TEF) is constructed. Using the nearest unstable equilibrium point (UEP) method, the maximum energy required to maintain stability in an untriggered limiting system is determined. Then, the influence of frequency limiting is analyzed. It is found that the two points with the lowest energy on the frequency limiting curve, which have only kinetic energy, are the leaving boundary points. And a control method is proposed to reset the system operation point to these two special leaving boundary points after triggering the frequency limiting. Finally, the maximum energy that prevents the multi-VSG system from entering the current limiting mode is determined by using an optimization method. The frequency limiting value at these special leaving boundary points is configured based on the minimum value, ensuring the global stability of the multi-VSG system with both current and frequency limiting. The correctness of the proposed theory is proven by simulation analysis and hardware test of the 9-bus power system.

• Optimization Strategy for Rotor-side Converter of Grid-forming DFIG-based Wind Turbines Under Unbalanced Weak Grid

  XU Shang, XIE Zhen, YANG Shuxin, LI Mengjie, YANG Shuying, ZHANG Xing

  2025,49(7):35-45, DOI: 10.7500/AEPS20240513004

  Abstract(88) HTML(38) PDF(2.52 M)( 69)

  Abstract:When doubly-fed induction generator (DFIG)-based wind turbines operate under an unbalanced power grid, the system has positive- and negative-sequence paths. With the weakening of power grid strength, the coupling between the turbine and the grid intensifies, and disturbances in unbalanced stator voltage will affect the normal operation of the system. In order to study the unbalanced operation and stability mechanism of doubly-fed wind turbine units under a weak grid, first, a small-signal state-space model of DFIG-based wind turbines is constructed, and the stability and unbalanced degree of positive- and negative-sequence rotor current control are analyzed based on the system model. Then, the optimal control strategy based on adaptive virtual impedance is proposed to realize the cooperative compensation for the unbalanced stator voltage and current under the weak grid. The theoretical analysis proves that the proposed strategy effectively improves the stability of the DFIG-based wind turbines under the unbalanced weak grid. Finally, through the hardware-in-the-loop experiment platform, the effectiveness of the proposed optimization strategy is verified.

• Switched System Modeling for Repeated Low Voltage Ride-through of Wind Turbine and Mechanism Analysis of Voltage Oscillations

  LAI Qiping, SHEN Chen, YANG Yanchen, LI Dongsheng, ZHANG Jun

  2025,49(7):46-56, DOI: 10.7500/AEPS20240401001

  Abstract(89) HTML(75) PDF(7.76 M)( 72)

  Abstract:The wind power connected sending-end grid is electrically far away from the main grid, lacking the support of synchronous generators, which shows distinct weak grid characteristics. Therefore, the voltage oscillation phenomenon is likely to occur, threatening the safe and stable operation of power systems. Its dynamic process and evolution mechanism need to be studied urgently. First, the conditions for voltage oscillation caused by repeated low voltage ride-through of wind turbines are analyzed through steady-state power flow calculation. Then, based on the switched system theory, considering the external connection impedance and internal control dynamics of the wind turbine, a switched system model for the grid-side converter of wind turbine is established. Next,by substituting relevant parameters, the dynamic evolution process of the terminal current, voltage, and power during the repeated low voltage ride-through period of the wind turbine is analyzed, and the mechanism of voltage oscillation is revealed. Finally, the voltage oscillation phenomenon of the wind turbine at the grid-connected point is reproduced through simulation, and the correctness of theoretical analysis results is verified. Furthermore, the main factors influencing the characteristics of voltage oscillations are explored as well.

• Amplitude-Frequency Modulation Characteristics and Demodulation Method of Alternating Electrical Variables in Initial Stage of System Small-disturbance

  WU Penghui, WANG Sicheng, YUAN Xiaoming

  2025,49(7):57-67, DOI: 10.7500/AEPS20230914006

  Abstract(90) HTML(30) PDF(2.82 M)( 66)

  Abstract:During the dynamic process of system oscillation occurring after disturbances in the power electronics dominated system, the unbalanced power at device ports drives the adjustment of amplitude/frequency (A/F) of the the internal voltage, subsequently altering the AC voltage/current in the network, so each alternating electrical variable evolves in the form of its A/F being modulated. The A/F modulation process of the formation and evolution of the AC signal leads to profound changes in its oscillation characteristics. Therefore, this paper aims to clarify the oscillation characteristics of the AC signal based on the formation and evolution mechanism of the A/F modulation of the AC signal and proposes a demodulation method for the corresponding signal feature extraction. Firstly, according to the basic principle of system operation, this paper introduces the A/F modulation mechanism of AC signals during the dynamic process of the system small-disturbance, clarifying the A/F modulation characteristics that determine the formation and evolution of AC signals during the intial stages of the oscillation process. Subsequently, it proposes a demodulation method and its implementation measures for extracting the A/F modulation characteristics of AC signals. Finally, by combining different scales of simulations and experiments, the rationality of the modulation formation and evolution mechanism of AC signals during system dynamic processes is verified, along with the effectiveness of the proposed demodulation method for feature extraction.

• Medium-frequency Characteristic Analysis and Oscillation Suppression of Converter with Interaction Between Grid-following and Grid-forming Control

  MENG Changxin, WANG Yanxu, WANG Weiru, LIU Yunfei, DONG Hongda, LI Chenggang

  2025,49(7):68-79, DOI: 10.7500/AEPS20231230001

  Abstract(89) HTML(43) PDF(3.51 M)( 63)

  Abstract:The integration of grid-following and grid-forming converters to the power grid brings different influences to the medium-frequency stability of the system. To analyze the differences between the two influences in the medium-frequency band, this paper first establishes the small-signal sequence impedance models of grid-following and grid-forming converters based on the harmonic linearization method. The sequence impedance characteristics in the medium-frequency band of the grid-following and grid-forming converters are compared. Based on the sequence impedance model, the influence of the filter inductance of the converter, the current inner loop and the phase-locked loop of the grid-following converter, and the active power damping coefficient and the virtual rotational inertia coefficient of the grid-forming converter on the impedance characteristics of the converters in the medium-frequency band is analyzed in detail. The dominant factors affecting the medium-frequency band are obtained by using the impedance sensitivity analysis method. Then, the influence of the interaction impedance on the joint impedance characteristics of the converter is analyzed by using the superposition theorem. The stability criterion of the three-port network composed of the grid-connected system with grid-following and grid-forming converters is derived. The influence of the system parameters and the output power on the joint impedance of grid-following and grid-forming converters is analyzed. In addition, from the perspective of joint impedance, the influence of different power proportions of the two types of converters on the system stability is further explored. The positive resistance characteristic of the grid-forming converter can mitigate the negative damping characteristic introduced by the control link of the grid-following converter, so that the grid-following converter can transition from negative resistance and capacitance to positive resistance and inductance, which enhances the stability of the oscillation in the medium-frequency band under the grid-following control, and suppresses the system oscillation well. Finally, the correctness of the analysis results is verified by the electromagnetic simulation software.

• Construction Method and Control-interaction Analysis of Security Region for Control Parameters of Grid-connected Converter

  WANG Jinyuan, LIU Chongru, SU Chenbo, WU Wenchuan, FANG Jiashu

  2025,49(7):80-90, DOI: 10.7500/AEPS20240301001

  Abstract(93) HTML(57) PDF(3.37 M)( 70)

  Abstract:The interaction between the phase-locked loop (PLL) and the DC voltage control loop of a power electronic converter under weak grid conditions may trigger oscillations. In order to reduce the risk of oscillation caused by the grid-connected converters, a method for determining the boundary of the security region for the control parameter is proposed. By studying the relative constraint relationship of the control bandwidth, the influence law of each control parameter on the interaction between the main control links of the grid-connected converter and between the controller and the weak grid is quantitatively analyzed. First, from the perspective that the dynamic response characteristics of the overlapped region of the controller bandwidth are similar, a mathematical analytical expression model of the closed-loop transfer function of the outer DC voltage loop covering the grid strength and each control link is established. Then, based on the analytical expression of this transfer function, the boundary of the security region for the control bandwidth ratio is solved by using the Nyquist stability criterion and stability margin constraints. The result can guide the design of the PLL and DC voltage control bandwidths under weak grid conditions and map it to the security region for each control parameter. In addition, the security region for control parameters intuitively reflects the influence of differences in parameter settings of the grid-connected converter on system stability, thereby guiding the adjustment direction of control parameters. Finally, the effectiveness and accuracy of the proposed method are verified through PSCAD/EMTDC simulations.

• Optimal Configuration of Reactive Power for Large-scale Photovoltaic Cluster Sending-end System Considering Multiple Objectives and All Scenarios

  CAI Zihan, GUAN Lin, ZHANG Ye, TANG Wangqianyun, HUANG Lei, HU Shijun

  2025,49(7):91-102, DOI: 10.7500/AEPS20240413002

  Abstract(91) HTML(51) PDF(3.98 M)( 67)

  Abstract:Large-scale photovoltaic (PV) clusters collects through multi-level AC transmission lines and transmits via flexible DC lines. They face the challenges such as long-distance transmission, frequent power flow fluctuations, and the lack of support from conventional synchronous power sources. The reactive power generation configuration becomes one of keys to ensure the stable operation of systems. Based on this, aiming at the sending-end system of large-scale PV clusters without synchronous power source support, the optimal configuration problems of reactive power compensation in substations at all levels in a multi-voltage level, high-capacity, long-distance AC transmission system are studied. To meet the requirements for the all-scenario adaptability of reactive power sources in reactive power planning, an envelope-based scenario set extraction method of the representative days is proposed, and a scenario decomposition and combination method is adopted to form a reactive power optimization and compensation sequence of the representative days quickly. Additionally, to improve the dynamic regulation capability of the system after disturbances, a multi-objective optimization model that considers the reactive power-voltage control margin of converter stations is proposed. Addressing the switching-state coupling issues of static reactive compensation devices caused by the fluctuations in the power output sequence of renewable energy, a dynamic clustering smoothing (DCS) method for grouping the configuration of reactive power compensation is proposed. Finally, the case verification is conducted on a practical large-scale PV sending-end system in the western region of China, and the verificaiton results demonstrate that the proposed method can comprehensively cover all the reactive power-voltage demands of operation scenarios and effectively improve the system stability while significantly reducing the computational complexity, and has good generality and efficiency.

• System Strength Evaluation of Grid-following and Grid-forming Interconnected System Based on Eigen-subsystem from Dual Perspective

  LENG Ruohan, YIN Bo, HU Guang, CHEN Hangyu, CONG Cong, ZHUANG Kehao, XIN Huanhai

  2025,49(7):103-114, DOI: 10.7500/AEPS20240512001

  Abstract(92) HTML(45) PDF(2.51 M)( 62)

  Abstract:The significant interactions between devices in grid-following and grid-forming interconnected systems present challenges in accurately quantifying system strength. To address this, firstly, the matching relationship between the closed-loop stability of grid-following and grid-forming systems and the open-loop network sensitivity from a dual perspective is discussed. Based on this, a power grid strength index reflecting current strength is introduced.Furthermore, the low-dimensional grid-following and grid-forming eigen-subsystems equivalent to the stability of the high-dimensional interconnected system are analytically obtained. Based on this, a two-dimensional index of the strength of the interconnected power system is proposed. This index can reflects the stability margins of both the grid-following and grid-forming eigen-subsystems simultaneously, thus enabling the effective assessment of the multi-frequency-band small-disturbance stability of the interconnected system. Combined with this index, the inference is drawn that an appropriate interconnected form of the power system is conducive to instability risk reduction of an both the grid-following and grid-forming systems simultaneously. Finally, simulations verify the effectiveness of the proposed method and the correctness of the inference.

• Coordinated Configuration Strategy of Grid-forming Stations and Synchronous Condensers for Renewable Energy Bases in Gobi Desert and Other Arid Areas

  LIN Li, XU Ning, CUI Hao, TANG Chuanwei

  2025,49(7):115-125, DOI: 10.7500/AEPS20240515013

  Abstract(101) HTML(41) PDF(3.04 M)( 64)

  Abstract:The development of renewable energy bases in desert, Gobi and desertification land has given a strong push to the low-carbon transition of power systems. However, due to the lack of synchronous power source on site, these bases are characterized by low inertia and low short-circuit ratio. The reasonable configuration of grid-forming stations and condensers can effectively improve the inertia and voltage characteristics of the bases, but the existing research often focuses on only one of the two characteristics. Therefore, a coordinated configuration strategy of grid-forming stations and synchronous condensers considering both inertia and voltage support is proposed in this paper. First, the supporting capability mechanism of the grid-forming station and the synchronous condenser and the configuration relationship between them are analyzed. Then, based on the reactive power support coefficient, the priority order of grid-forming station reconstruction is built, and the preliminary configuration scheme of the grid-forming station is determined according to the order to meet the system inertia demand. Further, the synchronous condensers are configured to satisfy the constraint of the short-circuit ratio of multiple renewable energy stations, and the preliminary scheme is modified considering the inertia support provided by condensers. The configuration that maximizes the support capability of grid-forming stations and synchronous condensers is achieved eventually. Finally, the effectiveness of the proposed strategy is verified by the simulation with a renewable energy base located in the Gobi Desert and other arid areas in Northwest China which lacks the support of synchronous power source.

• Transmission Scheme for Large-scale Renewable Energy Bases in Gobi Desert and Other Arid Areas Based on Integration of Source, Grid, DC and Storage

  WANG Yu, JIANG Chongxue, SONG Shengli, YANG Pengcheng, XIONG Lingfei, LIU Mingsong

  2025,49(7):126-134, DOI: 10.7500/AEPS20240901001

  Abstract(48) HTML(51) PDF(1.90 M)( 64)

  Abstract:Aimed at the transmission scenario of large-scale renewable energy bases in the Gobi Desert and other arid areas, an ultra-high voltage direct current (UHVDC) transmission scheme for large-scale renewable energy base based on the integration of source, grid, direct current and storage is proposed. The large-scale renewable energy base is divided into zones, and is integrated through multiple low-voltage voltage source converter based high voltage direct current (VSC-HVDC) channels to the UHVDC sending-end converter station. The renewable energy side of the integration channels is equipped with energy storage devices featuring grid-forming control functions to mitigate power fluctuations and support the grid-connected voltage of renewable energy. The UHVDC system adopts line commutated converter based high voltage direct current (LCC-HVDC) at the sending end and VSC-HVDC at the receiving end. The characteristics of the proposed scheme are analyzed in terms of applicable scenarios, operation features, and economic performance. A coordinated control strategy for multiple power electronic converters is also proposed. The feasibility of the proposed scheme is verified through simulations of steady-state and AC/DC transient conditions based on the PSCAD/EMTDC electromagnetic transient simulation platform.

• Fast Transfer Capability Assessment for HVDC Considering Switching Control Characteristics of Renewable Energy Generation

  YAN Jiongcheng, LIU Tianhao, LIU Yutian, DU Zhengchun

  2025,49(7):135-147, DOI: 10.7500/AEPS20240922001

  Abstract(36) HTML(35) PDF(1.86 M)( 60)

  Abstract:The large-scale renewable energy generation bases mainly utilize the HVDC to transfer renewable energy generation. Online assessment for transfer capability of HVDC is a key technology for monitoring the operation security of renewable energy generation bases. Hence, considering the switching control characteristics of renewable energy generation during the fault ride-through process, an assessment method for HVDC transfer capability based on stacked target-related denoising autoencoder is proposed. First, the relation equations of the simplified model of the renewable energy generation transmission system through HVDC are derived and analyzed. The reasons that the switching control of renewable energy generation can aggravate the transient overvoltage levels and more severely restrict HVDC transfer capability are clarified. And the mathematical model of HVDC transfer capability calculation is constructed. Then, based on the stacked target-related denoising autoencoder, the assessment model of the maximum HVDC transfer capability is constructed. A switching weight selection mechanism is introduced to select the weight vector of the assessment model according to the sequential switching states of control modes, effectively taking into account the impact of switching control on HVDC transfer capability. Finally, an incremental updating method for the assessment model based on the orthogonal weight modification is proposed. The model weights are updated in the orthogonal direction of the previous sample feature space, which can decrease the adverse effect of catastrophic forgetting problem. Simulation results demonstrate that the proposed method can effectively adapt to the switching control characteristics of renewable energy generation, and achieve the fast and accurate assessment for the maximum HVDC transfer capability.

• Response-driven Assessment and Suppression Strategy for Transient Overvoltage at DC Sending End

  QIN Boyu, ZHANG Zhe, GAO Xin, DING Tao, ZHANG Yixing

  2025,49(7):148-157, DOI: 10.7500/AEPS20240407010

  Abstract(98) HTML(36) PDF(1.70 M)( 64)

  Abstract:For high-voltage direct current transmission systems with large-scale renewable energy, commutation failures may cause transient overvoltage problems at the sending end, leading to the disconnection risk of renewable energy. To ensure the safe and stable operation of the DC sending system under commutation failures, a response-driven assessment and suppression strategy for transient overvoltage is proposed. Firstly, integrating the squeeze-and-excitation (SE) module and the LeNet network, the overvoltage amplitude prediction model is established to explore the mapping relationship between the overvoltage level and key response features under normal operation conditions of the power grid. Secondly, the impact of the connection capacity of the condenser on AC bus voltage at the sending end is analyzed, and a simplified selection principle of condenser capacity is proposed based on overvoltage prediction results, preliminarily determining the number of condensers. Finally, dynamic reactive power characteristics of condensers during the commutation failure are analyzed, and an excitation control optimization strategy based on trigger angle is designed to weaken the anti-modulation effect and further suppress the transient overvoltage. The simulation results verify the accuracy of the proposed overvoltage assessment method and the effectiveness of the suppression strategy for a DC transmission system with a transient overvoltage problem.

>Basic Research

• Grid Charging Demand Forecasting for Private Electric Vehicles Considering Regional Differences

  WEN Xin, HUANG Xueliang, GAO Shan, LIU Yu, GU Yaru

  2025,49(7):158-168, DOI: 10.7500/AEPS20231207002

  Abstract(90) HTML(58) PDF(4.20 M)( 50)

  Abstract:To a certain extent, the difference in the location of geographical regions and functional attributes of regional grids result in the great difference in electric vehicle charging load within each grid. Aiming at the insufficient consideration of the difference in dynamic charging load distribution of electric vehicles in current research on electric vehicle charging demand forecasting, this paper proposes a data-driven grid charging demand forecasting method for private electric vehicles considering the differences of geographical regions and the diversity of user trips. Firstly, data mining on the travel tracks of private electric vehicle users, the urban traffic network, and other data types is conducted. Mathematical models are constructed to obtain the origin-destination information of multi-stage trips and the basic travel patterns of private electric vehicle users. Secondly, the latitude and longitude coordinates of each point of interest (POI) are mapped to the geographic network based on the geographic information system platform. Various POI quantity sets combined with users’ daily travel purposes in the geographical region grid are classified. The natural classification method is adopted to implement accurate grid division of the studied geographical region. The functional-area grid includes five categories: the work area, the business area, the living area, the residential area, and the mixed area. An origin-destination information probability matrix for each functional area is established during multiple periods. Combined with the obtained distribution results of private electric vehicles in each grid, this paper establishes an electric vehicle charging load forecasting model based on the Monte Carlo method to capture the continuous changes of electric vehicle electricity amount transferred between grids. Based on the actual historical data of electric vehicles in Suzhou, China, and taking a region of Suzhou as the application environment, the simulation of charging demand forecasting for private electric vehicles in each functional region is completed. The simulation results verify the rationality of regional grid division and the accuracy of charging demand forecasting.

• Optimal Power Flow of Integrated Hydrogen-Electricity Energy System Based on Reduced-dimension Projection

  PENG Hongyi, YAN Mingyu, ZHOU Yijia

  2025,49(7):169-178, DOI: 10.7500/AEPS20240828005

  Abstract(60) HTML(45) PDF(1.50 M)( 66)

  Abstract:An optimal power flow calculation method based on the reduced-dimension projection is proposed to solve the problem of optimal power flow in the integrated hydrogen-electricity energy system considering the uncertainty of wind power. Firstly, the centralized operation model of an integrated hydrogen-electricity energy system considering the uncertainty of wind power is established and equivalently transformed into a deterministic model. Then, the deterministic model is decomposed into subproblems for each subsystem. The reduced-dimension projection method is used to project the operation space of each subproblem into a lower-dimensional space, and the projected operation spaces of each subproblem are reconstrcted into a new convex hull.Next, by solving the optimal problem under the constraints of the reconstructed convex hull, the values of the common variables under the optimal cost of the integrated hydrogen-electricity energy system can be obtained. Furthermore, using the values of the common variables under the optimal cost to solve each subproblem, the power flow of each subsystem under the optimal cost can be obtained. This method conceals private information during the solution process and avoids iteration. Finally, the proposed method is tested on 6-6-node and 40-118-node integrated hydrogen-electricity energy systems, and the effectiveness, optimality, privacy protection, and other advantages of the proposed method are verified.

• Optimal Scheduling Strategy for Inertial Demand of New Power System Based on Stackelberg Game

  ZHAO Zhenghui, WANG Xianan, QI Buyang, PANG Pai, WANG Yang, ZHANG Qian

  2025,49(7):179-188, DOI: 10.7500/AEPS20240626009

  Abstract(43) HTML(49) PDF(1.33 M)( 69)

  Abstract:Compared with traditional power systems dominated by synchronous generators, new power systems exhibit distinct characteristics such as low inertia and low damping, which introduce new challenges to ensuring stable operation for the system. To ensure that the system maintains a high level of inertia during each time, this paper proposes an optimal scheduling strategy for the inertia demand of new power systems based on the Stackelberg game. The strategy consists of leaders and followers, where the leader framework solves the problem of minimizing scheduling economic costs, and the follower framework solves the problem of optimizing system inertia configuration. This model considers the minimum inertia demand of the system and configures additional inertia reserves to cope with the scenarios with insufficient inertia, while ensuring the economic benefits of additional supplementary inertia and the optimal overall scheduling economy. When evaluating the inertia demand of the system, the inertia provided by the inverter interface resources is easily affected by its variable operation conditions, which is not conducive to the minimum inertia evaluation and inertia configuration of the system. Therefore, the concept of virtual inertia reliability is proposed, and virtual inertia reliability coefficients are given according to different operation conditions to improve the accuracy of inertia evaluation in renewable energy power plants. Case studies validate the effectiveness of the proposed optimal scheduling strategy, demonstrating its ability to improve frequency stability during system operation while preserving economic efficiency.

• Power Restoration Method for Distribution Network Based on Optimal Configuration of Photovoltaic and Energy Storage in Extreme Disaster Scenarios

  TAO Wenjia, FENG Liang, PENG Ke, XIAO Chuanliang, ZHAO Xueshen, CHEN Jiajia

  2025,49(7):189-197, DOI: 10.7500/AEPS20231229002

  Abstract(107) HTML(70) PDF(1.27 M)( 75)

  Abstract:In recent years, outages of higher power sources in the distribution network triggered by extreme disasters have occurred repeatedly. When formulating power restoration strategies solely at the operation level, the distribution network often faces the problem of insufficient power resources, making it difficult to ensure a continuous and reliable power supply for important loads. To address the above problems, a power restoration method for the distribution network based on the optimal configuration of the photovoltaic and energy storage is proposed to achieve the optimal balance between efficient resource utilization and reliable load preservation. Firstly, the Beta distribution model is used to fit the light intensity distribution, and a photovoltaic output model considering temporal characteristics is established. Secondly, a bi-level power restoration optimization model is constructed for the distribution network, incorporating both the siting and sizing of the photovoltaic and energy storage, and the dynamic islanding partition in extreme disaster scenarios. Among them, the upper-level model considers investment and operation costs, line loss costs, and load loss costs, while the lower-level model aims to maximize the restored load as the objective function. Through the depth first search algorithm, the preliminary island partitioning is achieved, and the power margin and energy margin indicators are used to implement dynamic islanding partition, thereby achieving the maximum load restoration in extreme disaster scenarios. Finally, the feasibility and superiority of the proposed method are verified by comparing the restoration effects and costs of different strategies through numerical cases.

• Power Quality Disturbance Classification Method Based on Time-series Two-dimensional Transformation and Multi-scale Transformer

  WANG Shouxiang, LI Huiqiang, ZHAO Qianyu, GUO Luyang, WANG Tongxun, WANG Yang

  2025,49(7):198-207, DOI: 10.7500/AEPS20240412005

  Abstract(86) HTML(42) PDF(1.70 M)( 58)

  Abstract:With the increasing penetration of renewable energy, the power quality disturbance (PQD) problem faced by the power grid has become more complicated. The traditional classification method based on one-dimensional PQD signals has difficulty in extracting and identifying periodic and trend disturbances at the same time. To address this problem, this paper proposes a PQD classification method based on time-series two-dimensional transformation and multi-scale Transformer. First, the time-series two-dimensional transformation is used to convert the one-dimensional PQD time series into a set of two-dimensional tensors based on multiple periods, to deeply mine the characteristic information contained in the PQD signals in the two-dimensional space. Then, the multi-scale feature map of the PQD signal is extracted through the multi-scale Transformer encoder module. And the multi-scale Transformer decoder module is used to splice and fuse the multi-scale feature maps, for effectively merging the feature maps extracted on different scales. Finally, the PQD classification task is accomplished through a fully connected layer and a Softmax classifier. To verify the effectiveness of the proposed method, a dataset containing 24 kinds of PQD is established to test the model. The results indicate that the proposed method has a high classification accuracy and noise robustness for PQD signals.

• Generation Method for Transmission Grid Test System Based on Power Grid Evolution Model

  LI Bo, WEI Guangrui, ZHONG Haiwang, LIU Hui

  Available online:April 15, 2025  DOI: 10.7500/AEPS20240409003

  Abstract(337) HTML(38) PDF(1.75 M)( 134)

  Abstract:The IEEE test case has been widely used for simulation testing in various fields such as power system planning and operation. However, due to data privacy concerns, it isn’t easy to access publicly available datasets of actual power system generation and network structures. To address the issue, based on the evolution model of three-generation power grids, a new transmission test system generation method is proposed to build test cases that reflect actual power system network structures. First, a transmission expansion planning model considering N-1 security constraints is established. Secondly, the optimization objectives and constraints are proposed according to the characteristics of different stages of network development to simulate the evolution process of the power grids. To enhance the solving efficiency of the model, a binary representation method of transmission corridors is introduced to reduce the number of 0-1 variables. Finally, taking a provincial power grid as an example, an open-source power system test case is constructed. The validity of the constructed test case is verified based on the statistical properties of complex networks. Furthermore, the test case is applied to the optimization of transmission network structures to demonstrate its rationality.

• Anomaly Detection in Distribution Main Station Logs Based on Large Language Model

  Wang Shen, WEI Xingshen, ZHU Weiping, ZHU Daohua, GUAN Zhitao

  Available online:February 27, 2025  DOI: 10.7500/AEPS20240523001

  Abstract(314) HTML(338) PDF(1016.64 K)( 476)

  Abstract:Log anomaly detection is one of the key technologies to monitor the operation of distribution master station system and identify abnormal behavior. Existing log anomaly detection methods based on deep learning rely on a large amount of in-domain training data, and the scarcity of training data will lead to a significant decline in performance. Aiming at the above problems, based on the contextual reasoning characteristics of large language models, an adaptive hint strategy is designed and a training-free anomaly detection scheme for distribution master logs is implemented. Firstly, a demonstration example filtering algorithm is designed to dynamically select several high-quality demonstration examples from a small number of labeled local logs for different online logs. Then, combined with the task description and human experience knowledge, a text hint is automatically constructed to guide the large language model to complete the anomaly detection task of distribution master station logs. The experimental results on the general data set and the self-built distribution master station data set show that the proposed scheme has better performance than the existing methods, showing higher flexibility and generalization.

• Emergency Frequency Control Strategy of Microgrid Clusters for Extreme Survival of Critical Urban Loads

  WANG Ziyuan, XU Yin, WU Xiangyu, LI Jiaxu

  Available online:February 26, 2025  DOI: 10.7500/AEPS20241024004

  Abstract(126) HTML(84) PDF(1.97 M)( 411)

  Abstract:The proportion of electricity received outside the urban power grid is high, and extreme events leading to connectivity failures in the power transmission channels of the urban superior power grid may cause major power outages. In this extreme scenario, by flexible self-configuration operation and supply guarantee of microgrid clusters, the critical load survival can be achieved. However, the large transient frequency fluctuation of microgrids during sudden power shortages directly affects the success or failure of self-configuration operation and supply guarantee. Firstly, a framework for solving the extreme survival problem of microgrid clusters is proposed, and dynamic frequency constraints for microgrids considering multi-source collaboration are constructed based on the system frequency response model. Secondly, the frequency response model of microgrid containing nonlinear constraints such as ordinary differential equations and limiting links is differentially discretized based on the forward Euler method. Then, according to the generator tripping and load shedding during the transient frequency response of microgrid clusters, a microgrid cluster model for extreme survival is constructed considering inertia equivalent constraints, control action delay constraints and dynamic segmentation constraints. By solving a mixed-integer linear programming model, the emergency frequency control strategy and the segmentation state of microgrid clusters are obtained to coordinate multiple frequency regulation resources and minimize the load shedding volume while ensuring frequency safety. This model can be efficiently solved using mature commercial solvers. Finally, the effectiveness and superiority of the proposed emergency frequency control strategy of microgrid clusters are verified through numerical simulation analysis.

• Optimal Scheduling Method for Power Grid Considering Participation of Electric Vehicles in Flexible Regulation During Cold Wave Weather

  JIAO Zhijie, XU Yin, LIU Zhao, WANG Xiaojun, HE Jinghan, SI Fangyuan

  Available online:December 30, 2024  DOI: 10.7500/AEPS20240519001

  Abstract(65) HTML(156) PDF(3.06 M)( 819)

  Abstract:The arrival of a cold wave triggers a sudden drop in weather temperature, at this time, the energy consumption of the load increases, the output of renewable energy sharply decreases, the system backup and the power supply capacity of the higher-level power grid are insufficient, resulting in a significant power shortage problem within the power grid in a short period of time. With the increasing number of electric vehicles on the load side and the improvement in the responsiveness of flexible resources, the load-side flexible resources are adjusted to compensate compensate for the power shortage caused by the cold wave. This paper first clarifies that the flexibility of resources varies with changing scenarios, explores and constructs models of flexible resources such as electric vehicles during cold wave events. Secondly, considering that electric vehicles in flexible resources need to participate in grid scheduling through aggregation, a method for aggregating electric vehicles with uncertain connection times is developed. Subsequently, based on the flexible resources during cold wave events, non-residential loads, and residential loads, a rolling optimal scheduling method for the power grid during cold wave weather is proposed with the objective of minimizing social losses. The electricity adjustment of various resources in the grid during cold wave weather is determined. Finally, through case studies, the proposed method is shown to effectively address power shortages in the grid during cold wave events.

• Identification Method for Tree-touching Grounding Fault of Medium-voltage Line Based on Long-term Variation Features of Zero-sequence Current

  WANG Pengwei, XU Bingyin, LIANG Dong, WANG Lianhui, WANG Chao, ZOU Guofeng

  Available online:October 31, 2024  DOI: 10.7500/AEPS20240315002

  Abstract(71) HTML(56) PDF(6.79 M)( 157)

  Abstract:Distinguishing whether faults in medium voltage distribution lines are caused by lines touching trees is of great significance for clarifying the causes of forest fires and preventing line faults from causing forest fires. The zero-sequence currents of various high-impedance grounding faults are obtained through prototype experiments in the paper, and the long-term variation features of the zero-sequence current waveforms of high-impedance grounding faults are analyzed. Analysis shows that there are significant differences in the fluctuation, monotonicity, and sharpness of the waveforms of the effective value of the zero-sequence currents for line touching trees grounding faults compared to other high-impedance grounding faults. A multi feature fusion parameter set including standard deviation, discrete coefficient, kurtosis, skewness of the zero-sequence current effective value curve is designed, and a ientification method for tree-touching grounding fault of medium-voltage line based on support vector machine is constructed. The results showed that the proposed method achieved a fault recognition accuracy of 98%.

• Frequency Safety Assessment of Power Systems Based on Multivariable-sample Convolution and Interaction Network

  LIU Jie, SHI Fang, SONG Xuemeng, TIAN Shuoshuo, NIE Liqiang

  Available online:October 17, 2024  DOI: 10.7500/AEPS20231101004

  Abstract(61) HTML(58) PDF(2.34 M)( 252)

  Abstract:The existing intelligent assessment methods for transient frequency in power systems do not adequately consider the temporal characteristics of input data. Therefore, a frequency safety assessment method for power systems based on intelligent prediction of transient frequency response curves is proposed in the paper. A multivariate sample convolutional interactive network is designed to fully exploit the temporal characteristics of power system measurement data, thereby improving the prediction accuracy of power system transient frequency response curves; Key indicators such as maximum frequency deviation, occurrence time of maximum frequency deviation, and the metastability frequency are calculated based on the predicted frequency response curves, and the frequency safety of the system is comprehensively assessed. Simulation tests are conducted on frequency stability standard examples, and the results showed that the proposed method effectively improved the accuracies of frequency response curve prediction and system frequency safety assessment compared to classical methods such as deep learning.

• Selectivity Enhancement Scheme for Zero-sequence Current Protection Based on Distance Relay for High-resistance Grounding Fault

  LI Yong, LI Yinhong, LIU Huanzhang, LIU Yang

  Available online:October 09, 2024  DOI: 10.7500/AEPS20240228008

  Abstract(39) HTML(26) PDF(2.76 M)( 173)

  Abstract:The last section of zero-sequence current protection of AC line adopts 300 A, which has the risk of disordered tripping.Therefore, a new principle of high-resistance grounding distance relay based on zero-sequence reactance line and non-fault phase polarization is proposed. The relay adopts the technical route of phase selection before measurement. The phase selection element combines zero-sequence reactance line and non-fault phase polarization method to form a variety of combined criteria to complete the phase selection. Due to the phase difference between the zero-sequence current at the protection installation and the zero-sequence current at the fault point, the zero-sequence reactance lines of the single-phase grounding fault phase and the advance phase of the inter-phase grounding fault have aliasing region when the fault point is near the setting point. The large variation of the operation voltage of the non-fault phase is not conducive to distinguishing the two types of faults in the aliasing region, and thus the phase selection element is divided into low-resistance module and high-resistance module. The low-resistance module adopts the zero-sequence reactance line with the downward bias, which is used to identify the near end and low-resistance short circuit. With the assistance of the low-resistance module, the high-resistance module only needs to deal with the faults near the set point, which reduces the difficulty of distinguishing the two types of faults . After phase selection, the operation voltage before fault is obtained by non-fault phase polarization method, so as to determine the operation characteristics of the relay. The ability of high-resistance distance relay to withstand the transition resistance is far beyond the requirements of the regulations, which improves the selectivity of ground backup protection to high-resistance faults.

• High-resistance Grounding Fault Detection Method for Distribution Network Considering Capacitor Switching Disturbance

  HAN Zhaoru, SHI Fang, ZHANG Hengxu, JIN Zongshuai, YUN Zhihao

  Available online:September 29, 2024  DOI: 10.7500/AEPS20240116008

  Abstract(26) HTML(98) PDF(4.03 M)( 468)

  Abstract:The accurate and reliable detection of high-impedance grounding fault (HIGF) is a challenging issue in fault handling for distribution networks, and normal capacitor switching operations can cause interference. Addressing this problem, a disturbance-resistant detection method for HIGFs based on zero-sequence Lissajous curve analysis is proposed in this paper. Firstly, the zero-sequence electrical quantities of HIGFs and capacitor switching disturbances are theoretically derived. There is no regular difference in the traditional time-frequency domain feature aspect between the two, thereby clarifying the cause of the interference. Further, the zero-sequence current and voltage waveforms are reconstructed into zero-sequence Lissajous curves. A quantitative index for the distortion complexity of the Lissajous curve trajectory shape based on the mathematical morphology theory is proposed, and an adaptive starting criterion is designed in combination with the probability distribution law of the zero-sequence Lissajous curve area. The disturbance-resistant detection algorithm for high-impedance grounding faults in the noise scenario is presented. Finally, the effectiveness and reliability of the proposed method are verified through electromagnetic transient simulation examples and real fault tests in the distribution network.

• Development and Application of Controllable Line Commutated Converter

  HE Zhiyuan, GAO Chong, YE Hongbo, YANG Jun, WANG Chenghao, SHENG Caiwang

  Available online:August 29, 2024  DOI: 10.7500/AEPS20240506007

  Abstract(16) HTML(6) PDF(2.88 M)( 171)

  Abstract:Controllable line-commutated converter(CLCC) is a new type of DC converter equipment proposed to solve the commutation failure problem of conventional DC transmission converter. In June 2023, the world"s first set of controllable phase converter valves for the ±500kV/1200MW Gezhouba to Shanghai Nanqiao high-voltage DC transmission system renovation project (hereinafter referred to as "Genan renovation project") was successfully put into operation. The technical requirements and principles of CLCC are analyzed, as well as its technical and economic benefits. Combined with the input conditions of the Genan renovation project system, the electrical parameters and structural design scheme of the controllable line-commutated converter valve were proposed, and the equipment development was completed. The type test scheme and test parameters were proposed, and the type test assessment was completed. According to the technical characteristics of the controllable line-commutated converter valve, field tests such as low-voltage test, open line test, and artificial short-circuit test were carried out to ensure the smooth operation of the project. The operation performance since the project put into operation was introduced, and the correctness of the control sequence during the AC fault was analyzed based on the field recording. Finally, the application prospects of the controllable? line-commutated converter valve in UHVDC projects, provincial AC liaison line renovation and other scenarios were prospected and analyzed, providing a reference for the further promotion and application of the controllable commutation converter valve.

• Ultra-short-term Regional Load Forecast Based on Spatio-Temporal Graph Attention Network

  ZHAO Ziyu, CHEN Yuanrui, CHEN Tingwei, LIU Junfeng, ZENG Jun

  Available online:April 28, 2024  DOI: 10.7500/AEPS20230914003

  Abstract(66) HTML(40) PDF(3.29 M)( 185)

  Abstract:A regional-level ultra-short-term load forecasting model based on a spatio-temporal graph attention network is proposed in this paper. Firstly, based on the existing regional-level load, cell partitioning is carried out to construct a graph topology that considers cell correlation. Secondly, effective features are extracted from the spatial, feature, and temporal dimensions through the graph attention network, one dimensional convolutional network and gated recurrent unit, connecting the fully connected layers to output the results. Finally, simulation validation is conducted based on real power load data from the New England region of the United States, and model attention weights are extracted to analyze spatial dependencies between cells. The results show that, compared with traditional models, the proposed model provides higher accuracy and stability with different prediction steps, effectively exploiting the spatial dependence of regional spatial load.

• Pricing Strategy of Charging Station in Power-Transportation Coupling Network Considering User Travel Cost Budget

  XIE Longtao, XIE Shiwei, CHEN Kaiyue, ZHANG Yachao, CHEN Zhidong

  Available online:January 04, 2024  DOI: 10.7500/AEPS20230628010

  Abstract(61) HTML(158) PDF(2.23 M)( 209)

  Abstract:With the large-scale development of electric vehicles, it is of great significance to study how to effectively consider the travel behavior mechanism of users and formulate rational charging prices for charging stations for the collaborative optimization and scheduling of power-transportation networks. To solve this problem, this paper proposes a pricing strategy for charging stations in the power-transportation coupling network considering the user travel cost budget. Firstly, a transportation user equilibrium model considering the travel cost budget is established, and the equilibrium state is equivalently described through variational inequalities, so as to characterize the travel demands and charging behaviors of electric vehicles. Secondly, a second-order cone optimization model for distribution networks considering power reduction is constructed. The charging station pricing problem has been transformed into an optimization problem with variational inequality constraints, and an alternating iteration algorithm combined with an extra-gradient algorithm is designed to solve the problem. Finally, the effectiveness of the proposed model and methods is verified through a case, and the results show the necessity of considering the travel cost budget for charging pricing in coupled networks.

• Non-Intrusive Real-Time Monitoring System for Electric Bicycle Based on Wavelet Detection and Feature Graph Decision

  LI Xiang, Liu Yuhang, ZHANG Qi, WU Xin

  Available online:September 12, 2023  DOI: 10.7500/AEPS20230425001

  Abstract(67) HTML(70) PDF(2.53 M)( 195)

  Abstract:The illegal charging behavior of electric bicycles (EBs) in households has temporal randomness and spatial concealment,which poses significant safety hazards and is difficult to effectively manage. A non-intrusive real-time monitoring system for EB charging based on wavelet detection and feature graph decision is proposed, utilizing the characteristics of real-time autonomous execution and easy promotion of non-invasive monitoring systems. Considering the physical structure and charging characteristics of EB loads, the typical common characteristics of EB loads are analyzed from both transient and steady-state perspectives. The EB proprietary feature map with strong distinguishability and universality is constructed in advance to realize consistent and structured expression of EB steady-state common features. In the actual monitoring process, in order to reduce the computational power demand and data transmission pressure of the system, EB specific transient phenomena with high-frequency components are accurately located based on wavelet transform to complete EB like event detection. Finally, extract event waveforms and train efficient classifiers through graphs for load identification and real-time upload. By monitoring actual users, the effectiveness of the monitoring system has been verified, which can effectively solve the problem of charging EBs in buildings and households.

• Spatial Feature Based Detection of False Data Injection Attack on Synchronous Grid Measurements

  ZHENG Yao, ZHANG Jie, YAO Wenxuan, QIU Wei, TANG Sihao

  Available online:March 16, 2023  DOI: 10.7500/AEPS20220813001

  Abstract(42) HTML(202) PDF(2.51 M)( 49)

  Abstract:As the power system gradually moves toward a new ecosystem of energy interconnection and the deep coupling of the network layer and physical layer, the threat of network attacks on the power system continues to rise. The source identity (ID) spoofing attack, as a new and complex, strong stealthy false data injection attack, can cause the grid control system to misjudge and cause system paralysis. To address this problem, a spatial feature-based method is proposed for detecting false data injection attacks on synchronized measurements of power grids. It has extracted different spatial features of the synchronized measurement devices at different locations by variational modal decomposition (VMD) and improved discrete orthonormal Stockwell transform (IDOST), so as to extract the authentication information of the measurement data without losing the spatial features of the measurements. Combined with light convolutional neural network (LCNN) to evaluate the likelihood of measurement data being attacked by source ID to enhance the speed of detection response. The effectiveness of the method is verified by the detection results of actual multi-point synchronous measurement data.

• Short-term Wind Power Interval Prediction Based on Combined Mode Decomposition and Deep Learning

  XIAO Bai, ZHANG Bo, WANG Xinwei, GAO Ningyuan

  Available online:February 20, 2023  DOI: 10.7500/AEPS20220807002

  Abstract(236) HTML(618) PDF(3.51 M)( 167)

  Abstract:Wind power prediction is very important for the economic dispatch of power systems containing wind power. Aiming at the problem that point prediction is difficult to describe the uncertainty of wind power, a short-term wind power interval prediction method based on combined mode decomposition and deep learning is proposed. Firstly, the original wind power sequence is decomposed into multiple modal components by using the improved complete ensemble empirical mode decomposition with adaptive noise, and the high-frequency strong non-stationary components are decomposed again by using the variational mode decomposition. On this basis, the sample entropy is used to calculate the complexity of each component and reconstruct them into trend components, oscillation components and random components. Then, the three components are input into the Bayesian optimization bidirectional long short-term memory neural network to establish their respective prediction models, and the point prediction values of the three components are obtained. The mixed kernel density estimation method is used to estimate the error distribution of the prediction results of oscillation components and random components, and the overall interval prediction results are obtained by combining the point prediction values. Finally, the actual examples show that this method has higher prediction accuracy than other models.

• Named Entity Recognition for Electric Power Industry Based on Enhanced Text Features

  LIU Wensong, HU Zhuqing, ZHANG Jinhui, LIU Xuejing, LIN Feng, YU Jun

  Available online:September 27, 2022  DOI: 10.7500/AEPS20210323003

  Abstract(40) HTML(42) PDF(1.43 M)( 54)

  Abstract:Considering the characteristics of small scale, nested entities and abbreviated entities for electric corpus, the named entity recognition (NER) based on enhanced vectors of text features is proposed. Firstly, by the way of the low grain word segment and the preset dictionary, the semantic information in Chinese words is properly utilized, and the transmission errors caused by word segment are decreased. Secondly, the features of inner structure of a single Chinese word is learned by the word-level bidirectional gated recurrent unit (Word BiGRU). Together with the features of the part of speech for words and word length, the enhanced word vector is built by concatenating these features vectors with word vectors. Finally, the NER model is designed with BiGRU, attention mechanism and conditional random field (CRF). The proposed method is verified using electric corpus and F1 measurement reaches 87.02%, which proves the effectiveness of NER for electric power industry.

• Identification of Abnormal Operation Condition for Environmental Protection in Enterprises Based on Power Quality Monitoring Data

  ZHANG Yi, YAO Wenxu, SHAO Zhenguo, ZHANG Liangyu

  Available online:August 18, 2022  DOI: 10.7500/AEPS20211203007

  Abstract(88) HTML(136) PDF(4.68 M)( 219)

  Abstract:Aiming at the problems of abnormal operation condition monitoring for environmental protection in polluting enterprises at present, such as difficult implementation, large identification errors and easy tampering with the results, this paper proposes an identification method of abnormal operation conditions for environmental protection based on power quality monitoring data. The multi-dimensional power quality data obtained from non-invasive load monitoring at the public power entrance of enterprise equipment are used to train the model of condition classification, to realize abnormal condition identification, which is different from the existing scheme of power consumption monitoring with a separate meter installed for each device. First, the time series change-point detection and the clustering calculation for the characteristic data representing the production conditions are carried out to realize the division of production operation conditions of enterprises. Then, combined with the operation of environmental protection equipment, the categories of environmental protection operation conditions for classification are obtained. Furthermore, the operation condition scenarios related to environmental protection are classified and learned by the Stacking learning model. Finally, the trained classification model is used to identify the abnormal operation conditions for environmental protection in the enterprise. The effectiveness of the proposed method is verified by the simulation test data and the actual enterprise data.

• Causes and Suppression Measures of Resonant Overvoltage in Offshore Wind Power AC Transmission System

  YANG Daye, SONG Ruihua, XIANG Zutao, LIU Dong, CHAO Wujie, YAN Yuesheng

  Available online:May 26, 2022  DOI: 10.7500/AEPS20220126001

  Abstract(32) HTML(18) PDF(3.69 M)( 44)

  Abstract:Since the capacitance per unit length of cable is more than 20 times that of overhead lines with the same voltage level, more and more offshore wind power is connected to the grid through AC cable, which reduces the natural resonant frequency of the transmission system and increases the resonant risk of the system. Aiming at the resonant overvoltage phenomenon in the process of grid connection of an offshore wind farm, the fault recording data are analyzed. Based on the impedance model of the transmission system, the mechanism is analyzed, and through the electromagnetic transient simulation, it is verified that the natural resonant frequency near the double frequency of the transmission system is the fundamental reason for the resonant overvoltage caused by the operation of no-load transformer and no-load line in the wind farm. Combined with the engineering practice, the joint suppression measures of changing the operation mode of the transmission system, optimizing the control and protection system of the static var generator and increasing the access load are proposed. Finally, the effectiveness of the proposed measures is verified by simulation and field tests.

• Calculation of Power Supply Reliability for Distribution Network Based on Neo4j Graph Database

  JIANG Wei, WANG Minghua, CHEN Jinming, LIU Jiangdong, PU Shi, XU Zhiqi

  Available online:May 13, 2022  DOI: 10.7500/AEPS20211031001

  Abstract(118) HTML(214) PDF(1.74 M)( 101)

  Abstract:In order to improve the efficiency of power supply reliability calculation for the complex distribution system, a reliability calculation method based on the Neo4j graph database is proposed. Firstly, the topology structure of the distribution network is stored in vertex-edge form through the graph database. Meanwhile, the feeder classification and load partitioning in the complex distribution system is completed by using the characteristics of different types of edges in the Neo4j graph database, and the distribution network diagram model based on the graph database is built. Secondly, the subgraph division of the distribution network diagram model is combined with path search, and the model simplification is completed based on each subgraph. Finally, the power supply reliability analysis of the distribution system is realized based on the minimal path reliability algorithm combined with the efficient shortest path query and other functions of the Neo4j graph database. The effectiveness of the proposed method is verified by comparing the Roy Billinton test system and an actual 10 kV distribution network in China for algorithm verification.

• Flexible Ring Network Controller Without Interface Transformer for Medium Voltage Distribution Network

  LI Zheng, CHEN Wu, HOU Kai, SHI Mingming, MOU Xiaochun, ZHU Jinsong

  Available online:April 26, 2022  DOI: 10.7500/AEPS20210806005

  Abstract(137) HTML(172) PDF(2.92 M)( 102)

  Abstract:When the flexible ring network controller cancels the interface transformer, the transmission of the zero-sequence voltage component cannot be prevented when the AC side fails, thus increasing the fault range. Therefore, this paper uses classical circuit analysis and positive and negative sequence analysis algorithms to explain the basic principles of the formation and transmission of zero-sequence voltage components. A topology of the flexible ring network controller without the interface transformer is proposed. The AC side converters are all modular multilevel converters with traditional half-bridge sub-modules. And the full-bridge sub-module valve strings are connected in series on the positive and negative polarity busbars. Utilizing the ability of the full-bridge sub-module to output positive and negative voltages, the DC side voltage fluctuations are suppressed, and the fault range is prevented from expanding. By using the MATLAB/Simulink software, the characteristics of zero-sequence voltage suppressed during the fault are simulated and analyzed. The simulation results verify the correctness of the theoretical analysis and the effectiveness of the proposed topology.

• Electricity Theft Detection Based on Multiple Different Learners Fusion by Stacking Ensemble Learning

  YOU Wenxia, LI Qingqing, YANG Nan, SHEN Kun, LI Wenwu, WU Zeli

  Available online:March 30, 2022  DOI: 10.7500/AEPS20210731001

  Abstract(238) HTML(154) PDF(1.38 M)( 116)

  Abstract:Aiming at the problems that the consumer power consumption data categories are unbalanced in electricity theft detection, and the ensemble learning method using voting as a combination strategy can not give full play to the advantages of multiple different learners, a model using Stacking ensemble learning to fuse multiple different learners is proposed and applied to electricity theft detection. Firstly, starting from the factors affecting electricity metering, six electricity theft behavior modes are simulated according to five common electricity theft methods; Secondly, synthetic minority oversampling technique (SMOTE) is used to process the unbalanced power consumption data, and k-fold cross-validation method is used to divide the balanced training sets to alleviate the overfitting caused by repeated learning; Then, the evaluation indicators and diversity metrics are employed to optimize different primary learners and meta-learners of the model, and a Stacking combination learning electricity theft detection model integrating the advantages and differences of different learners is constructed; Finally, the comparative analysis results of examples show that the proposed electricity theft detection model can effectively solve the imbalance of power consumption data categories, give full play to the advantages of different learners, and the evaluation index is good.

• Topology Optimization of Power Systems Combining Deep Reinforcement Learning and Domain Knowledge

  YAN Ziming, XU Yan

  Available online:December 01, 2021  DOI: 10.7500/AEPS20210510001

  Abstract(92) HTML(68) PDF(1.43 M)( 186)

  Abstract:While the flexibility of power systems operation can be improved by topology optimization, the dimension of system-level discrete decision variables, including the connections of lines and substation busbars, is prohibitively high. Thus, the topology optimization problem of power systems can hardly be solved by the conventional mixed-integer optimization method. Aiming at this problem, a reinforcement learning based method is proposed combining asynchronous advantage actor-critic and power system domain knowledge, which transfers the computational burden of online optimization to the offline agent training stage. The defined reward function is adopted to minimize the violations of power transmission line flow limits. Forced constraints verification is employed to reduce the searching space and improve the efficiency of the reinforcement learning. The fast computation of the topological structure optimization of power system operation is realized,and the operation security of power systems is enhanced. The effectiveness of the proposed method is validated by simulation testing results.

• A Rapid Fault Tracing Method for Large-Scale Complex Distribution Networks Based on Dynamic Graph ProjectionZHANG Yujia1^, YUAN Ye1^, ZHOU Suyang2^, ZHU Hong3^, ZHOU Aihua3^, CHEN Qingquan⁴

  Zhang Yujia, YUAN Ye, ZHOU Suyang, ZHU Hong, ZHOU Aihua, CHEN Qingquan

  Available online:  DOI: 10.7500/AEPS20240112003

  Abstract(139) HTML(0) PDF(0.00 Byte)( 0)

  Abstract:With the rapid growth of the distribution network and the high penetration of distributed resources, the topology of distribution networks has become increasingly complex, posing significant challenges to fault location analysis. When applying matrix algorithms and intelligent optimization algorithms to fault location, it is necessary to construct network matrices or establish optimization models based on changing topology information. This greatly increases the computational burden and complexity, leading to low efficiency in data processing and computation. Therefore, this paper first constructs a graph data model for the distribution network topology. Utilizing graph projection techniques, it extracts optimized subgraphs tailored for fault tracing scenarios from the panoramic power grid graph. On this basis, the Yen"s shortest path search algorithm is employed to find potential fault paths from the power source to the abnormal nodes. By traversing the line nodes and assessing their overcurrent information, the fault section is identified, thereby resolving the issues of accurate representation and rapid search of the power grid topology. This enables quick and precise fault localization in large-scale complex distribution networks, greatly enhancing fault search efficiency while ensuring the accuracy of fault tracing.

• Identification Method of Multi-perspective Time-frequency Characteristic Synthesis for Transformer Excitation Inrush Current

  Chen Chun, ZHAN Luxin, CAO Bozhong, CAO Yijia, Li Yong, LIU Junle

  Available online:  DOI: 10.7500/AEPS20240723005

  Abstract(160) HTML(0) PDF(0.00 Byte)( 0)

  Abstract:The continuous integration of power electronic devices in distribution networks has led to an increasing level of harmonic currents, posing challenges for traditional transformer secondary harmonic restraint differential protection. Simultaneously, single-feature identification methods are influenced by distributed energy resource types and closing angles, making it difficult to accurately distinguish fault currents and excitation inrush currents in different scenarios. To enhance the accuracy of excitation inrush current identification, this paper proposes a multi-angle time-frequency analysis method that comprehensively integrates time-domain, frequency-domain, and time-frequency-domain features. It utilizes Bayesian optimization of XGBoost (extreme gradient boosting) classification parameters to improve model generalization, enabling accurate identification of fault currents and excitation inrush currents under various capacities and types of distributed energy resource integration. The SHAP (shapley additive explanations) value analysis method is employed to reveal the contribution of each feature value in the identification model. The proposed method was verified through PSCAD/EMTDC simulation data and field measured data. Within the data samples provided in this article, the Bayesian-XGBoost algorithm under multi-angle time-frequency analysis has an accuracy of identification of excitation inrush current close to 100%, which is better than several common classification algorithms compared in this paper.

• Construction Scheme of Voltage Supported Renewable Energy Station Based on Distributed Synchronous Condenser——(II) Coordinate Control Strategy and Test Verification

  LIANG Hao, QIN Chuan, XIE Huan, LIANG Beihua, WU Tao, WANG Xuanyuan, WU Long

  Available online:  DOI: 10.7500/AEPS20240807002

  Abstract(169) HTML(0) PDF(0.00 Byte)( 0)

  Abstract:It is an effective measure to improve the whole process voltage support ability of the power supply side to deploy the distributed synchronous condenser (SC) in renewable energy station. However, the current "reactive outer loop + voltage inner loop" cascade strategy is adopted to integrate the SC to the automatic voltage control system (AVC) of the station to restrict its low-frequency voltage source characteristics. This paper first describes the existing problems of SC access to the AVC of the station, and proposes the requirement of constant voltage in the whole process of SC. Then, based on the topology of the renewable energy station with SC, the reactive voltage conversion coefficient and the reactive power shunt influence factor are analyzed. Based on this, a new scheme of "constant voltage + reactive power shunt suppression" for integrating the SC to the AVC is proposed. The program was developed in a domestic mainstream manufacturer"s equipment model, and the effectiveness of the scheme was verified by in-loop simulation of the SC and AVC dual controller at the station. Finally, the engineering application was completed in an actual new energy station. The field operation results show that the scheme realizes the steady-state regulation of AVC multi-type reactive power equipment of the station, effectively reduces the voltage fluctuation amplitude of PCC bus of the station, gives full play to the whole process voltage control of the SC, and guarantees the voltage stability margin of the system.

• A New Single-Ended Fault Location Algorithm for Grounding Electrode Lines Based on Broadband Excitation Injection and Short-Time Matrix Pencil Method

  TIAN Xincui, CHEN Kaiwen, SHAN Jieshan, ZHANG Yining, YU Jinyun, LI Qiang

  Available online:  DOI: 10.7500/AEPS20240904005

  Abstract(94) HTML(0) PDF(0.00 Byte)( 0)

  Abstract:The fault information in grounding electrode lines is weak and well-hidden, making detection and fault location very difficult. Based on this, a new single-ended fault location algorithm for grounding electrode lines is proposed, utilizing broadband excitation injection and the short-time matrix pencil method (STMPM). First, a Gaussian signal excitation with an “oscillatory decay characteristic” is injected into the grounding electrode line in differential mode, ensuring that the injected excitation does not leak into the DC system side through the neutral bus and minimizing waveform distortion during the propagation of the signal along the grounding electrode line, thereby improving the detection efficiency of fault traveling waves. Secondly, sliding short-time windows are used to perform singular value decomposition (SVD) on the fault traveling waves. The eigenvalues obtained from the decomposition are used to distinguish between interference signals and fault signals, effectively amplifying the weak fault signals while suppressing the interference signals. Finally, the damping factor of the fault traveling wave within the short window is determined, establishing a one-to-one mapping relationship between the zero-crossing moment of the damping factor and the arrival time of the fault traveling wave, and the fault distance is then determined. Extensive simulations show that the distance measurement algorithm can effectively detect fault traveling waves and achieve high fault location accuracy.

• Weakly Correlated Measurement Anomaly Recognition of Low-Voltage Distribution Grid Based on Neural Network with Weights Variability

  JIN-Yangxin, XU Yongjin, HU Shuhong

  Available online:  DOI:

  Abstract(32) HTML(0) PDF(0.00 Byte)( 0)

  Abstract:As an important data source of low-voltage distribution grid， the measurement anomaly control of smart energy meter concerns electricity trading fairness. A line loss recognition model of weakly correlated measurement anomaly was devised, whose inputs are frozen active power and voltage recorded by energy meter, as well as partial archival information. The principals of the model is a residual neural networks with variable weights, which is utilized to forecast the measurement anomaly index. Aimed at the accuracy degradation owing to insufficient inputs, the network weights variability algorithm was proposed, which modifies the tradition that weights are fixed after training, whereas regards them as superposition of several eigenstate weights. Via the node correlation state that is not able to be input directly, the superimposing parameters of eigenstate weights were calculated in feed-forward operations, with model recognition performance improved significantly. Finally the validity of residual neural networks with variable weights was verified with the enhanced training and validation datasets; furthermore, the model was applied to the control of high-loss range in a city in Zhejiang to confirm its detection effect under actual conditions.

• Data-driven Disturbances Classification for Distribution Network Based on Synchronized Measurements

  CHEN Zhilin, LIU Hao, BI Tianshu

  Available online:  DOI:

  Abstract(25) HTML(0) PDF(0.00 Byte)( 0)

  Abstract:To address the challenges posed by the integration of a large number of distributed renewable energy resources, synchronized phasor measurement technology has been introduced into distribution networks, providing a more comprehensive data basis for system safety and stability analysis. To further enhance situational awareness in distribution networks, this paper proposes a disturbance identification method driven by synchronous measurement data. On one hand, a disturbance source identification method based on Granger causal analysis is introduced to determine whether the detected disturbance originates from the transmission side or the distribution side. On the other hand, an algorithm misjudgment disturbance identification method based on kernel support vector machines (SVM) is proposed to distinguish whether the disturbance is caused by errors in the instantaneous frequency estimation algorithm. This comprehensive approach ultimately enables the identification and extraction of genuine distribution network disturbances. The effectiveness of the proposed method is tested and validated using IEEE123 simulation data and field synchronous measurement data.