Electric Power Systems Research最新文献_第8页

Short-Term photovoltaic power forecasting based on K-means++ clustering, secondary decomposition and TCN-BiLSTM-Attention model 基于k -means++聚类、二次分解和TCN-BiLSTM-Attention模型的光伏短期功率预测

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112749

Jianwei Liang , Jie Yue , Yanli Xin , Shuxin Pan , Jiaming Tian , Jingxuan Sun

Accurate short-term photovoltaic (PV) power forecasting is crucial for grid operation and energy dispatch. To further enhance the accuracy of PV power prediction, this paper proposes a short-term prediction method based on secondary decomposition, temporal convolutional networks (TCNs), bidirectional long short-term memory (BiLSTM) and an attention mechanism. Firstly, the Spearman correlation coefficient (SCC) is used to screen key meteorological features and reduce input redundancy. Secondly, the K-means++ algorithm is applied to cluster historical PV power data into three weather types, i.e. sunny, cloudy and rainy days. Subsequently, to reduce the non-stationarity and complexity of PV power series, variational mode decomposition (VMD) is employed to perform primary decomposition. Then, complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is applied to the residual components for secondary decomposition, thereby obtaining more stable and informative subsequences. A TCN-BiLSTM-Attention hybrid model is further constructed to capture long-term dependencies, bidirectional temporal dynamics and salient features from the decomposed subsequences for PV power prediction, effectively handling its inherent intermittency and variability. Finally, the predicted values of all subsequences are superimposed to obtain the final prediction result. Simulation experiments with six benchmark models on two different datasets show that the proposed model can effectively improve prediction accuracy.

准确的光伏短期功率预测对电网运行和能源调度至关重要。为了进一步提高光伏发电功率预测的准确性，本文提出了一种基于二次分解、时间卷积网络（TCNs）、双向长短期记忆（BiLSTM）和注意机制的短期预测方法。首先，利用Spearman相关系数（SCC）筛选关键气象特征，减少输入冗余；其次，采用k -means++算法将历史光伏发电数据聚类为晴、阴、雨三种天气类型。随后，为了降低光伏功率序列的非平稳性和复杂性，采用变分模态分解（VMD）进行一次分解。然后，对残差分量进行含自适应噪声的完全集合经验模态分解（CEEMDAN）进行二次分解，得到更稳定、信息量更大的子序列。进一步构建了TCN-BiLSTM-Attention混合模型，从分解的子序列中捕捉光伏发电功率预测的长期依赖关系、双向时间动态和显著特征，有效处理其固有的间歇性和可变性。最后，对所有子序列的预测值进行叠加，得到最终的预测结果。在两个不同的数据集上进行了6个基准模型的仿真实验，结果表明该模型可以有效地提高预测精度。

{"title":"Short-Term photovoltaic power forecasting based on K-means++ clustering, secondary decomposition and TCN-BiLSTM-Attention model","authors":"Jianwei Liang , Jie Yue , Yanli Xin , Shuxin Pan , Jiaming Tian , Jingxuan Sun","doi":"10.1016/j.epsr.2026.112749","DOIUrl":"10.1016/j.epsr.2026.112749","url":null,"abstract":"<div><div>Accurate short-term photovoltaic (PV) power forecasting is crucial for grid operation and energy dispatch. To further enhance the accuracy of PV power prediction, this paper proposes a short-term prediction method based on secondary decomposition, temporal convolutional networks (TCNs), bidirectional long short-term memory (BiLSTM) and an attention mechanism. Firstly, the Spearman correlation coefficient (SCC) is used to screen key meteorological features and reduce input redundancy. Secondly, the K-means++ algorithm is applied to cluster historical PV power data into three weather types, i.e. sunny, cloudy and rainy days. Subsequently, to reduce the non-stationarity and complexity of PV power series, variational mode decomposition (VMD) is employed to perform primary decomposition. Then, complete ensemble empirical mode decomposition with adaptive noise (CEEMDAN) is applied to the residual components for secondary decomposition, thereby obtaining more stable and informative subsequences. A TCN-BiLSTM-Attention hybrid model is further constructed to capture long-term dependencies, bidirectional temporal dynamics and salient features from the decomposed subsequences for PV power prediction, effectively handling its inherent intermittency and variability. Finally, the predicted values of all subsequences are superimposed to obtain the final prediction result. Simulation experiments with six benchmark models on two different datasets show that the proposed model can effectively improve prediction accuracy.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112749"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039100","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

A power quality enhancement using gradient descent-fuzzy logic controller in grid-connected PV system 基于梯度下降模糊控制器的并网光伏系统电能质量增强

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112746

Balaji Janapati, Nageswara Rao Pulivarthi

In recent years, the integration of renewable energy sources, particularly Photovoltaic (PV) systems into the power grid have introduced critical power quality issues. These issues primarily arise from the dynamic behavior of related power electronic converters and the intermittent nature of solar generation. Traditional standalone Maximum Power Point Tracking (MPPT) algorithms repeatedly suffered from poor adaptability, slow dynamic response under varying conditions, and insufficient harmonic reduction. These limitations degraded grid performance and inefficient energy withdrawal from PV systems. To overcome these challenges, this research proposes a novel hybrid approach that integrates Gradient Descent optimization with a Fuzzy Logic Controller (GD-FLC). The proposed controller is applied to a 100-kW grid connected PV system, incorporating an LC filter and MPPT-based boost converter to adjust voltage and improve power quality. The proposed GD-FLC attains fastest MPPT tracking time (0.29 s) under irradiance variation, and (0.28 s) under load variation, along with the lowest average rise time (0.0117 s). Similarly, it achieved output energy (160.72KJ) under step variation and 150.49KJ under ramp variation, clearly outperforming existing MPPT techniques. The GD-FLC controller offers an effective solution for grid-connected PV systems by attaining 89.82 kW for grid load and 80.09 kW for non-linear load respectively.

近年来，将可再生能源，特别是光伏（PV）系统并入电网带来了关键的电能质量问题。这些问题主要来自相关电力电子转换器的动态行为和太阳能发电的间歇性。传统的单机最大功率点跟踪（MPPT）算法存在自适应性差、动态响应慢、谐波抑制不足等问题。这些限制降低了电网性能，降低了光伏系统的能量提取效率。为了克服这些挑战，本研究提出了一种将梯度下降优化与模糊逻辑控制器（GD-FLC）相结合的新型混合方法。该控制器应用于100kw并网光伏系统，结合LC滤波器和基于mpt的升压变换器来调节电压并改善电能质量。在辐照度变化和负载变化条件下，GD-FLC的MPPT跟踪时间最短（0.29 s），平均上升时间最短（0.0117 s）。同样，它在阶跃变化下的输出能量为160.72KJ，在斜坡变化下的输出能量为150.49KJ，明显优于现有的MPPT技术。GD-FLC控制器为光伏并网系统提供了有效的解决方案，电网负荷和非线性负荷分别达到89.82 kW和80.09 kW。

{"title":"A power quality enhancement using gradient descent-fuzzy logic controller in grid-connected PV system","authors":"Balaji Janapati, Nageswara Rao Pulivarthi","doi":"10.1016/j.epsr.2026.112746","DOIUrl":"10.1016/j.epsr.2026.112746","url":null,"abstract":"<div><div>In recent years, the integration of renewable energy sources, particularly Photovoltaic (PV) systems into the power grid have introduced critical power quality issues. These issues primarily arise from the dynamic behavior of related power electronic converters and the intermittent nature of solar generation. Traditional standalone Maximum Power Point Tracking (MPPT) algorithms repeatedly suffered from poor adaptability, slow dynamic response under varying conditions, and insufficient harmonic reduction. These limitations degraded grid performance and inefficient energy withdrawal from PV systems. To overcome these challenges, this research proposes a novel hybrid approach that integrates Gradient Descent optimization with a Fuzzy Logic Controller (GD-FLC). The proposed controller is applied to a 100-kW grid connected PV system, incorporating an LC filter and MPPT-based boost converter to adjust voltage and improve power quality. The proposed GD-FLC attains fastest MPPT tracking time (0.29 s) under irradiance variation, and (0.28 s) under load variation, along with the lowest average rise time (0.0117 s). Similarly, it achieved output energy (160.72KJ) under step variation and 150.49KJ under ramp variation, clearly outperforming existing MPPT techniques. The GD-FLC controller offers an effective solution for grid-connected PV systems by attaining 89.82 kW for grid load and 80.09 kW for non-linear load respectively.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112746"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039097","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Optimal planning and charging/discharging control of retired batteries for demand-side applications under different PV integration schemes 不同光伏整合方案下需求侧应用的退役电池优化规划与充放电控制

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112757

Rong-Ceng Leou , Kun-Che Ho , Yi-Hwa Liu

This paper evaluates the performance of a demand-side energy storage system (ESS) using retired batteries under different photovoltaic (PV) integration strategies. Mathematical models are developed to optimize ESS charging and discharging under both self-consumption and feed-in tariff (FIT) schemes, with the objective of minimizing electricity costs while accounting for electricity price variations, load and PV generation uncertainties, and battery aging. The impacts of FIT rates and electricity tariffs on PV integration scheme are examined, and an economic analysis model is proposed to determine the optimal ESS capacity by considering cost savings, system costs, and battery degradation. Sensitivity analyses are conducted on key parameters to assess their influence on system benefits. In addition, comparative results for different depth-of-discharge (DOD) operations between retired and new batteries indicate that retired batteries can effectively reduce environmental impact while providing a practical and cost-efficient solution.

在不同的光伏集成策略下，对使用退役电池的需求侧储能系统（ESS）进行了性能评估。在考虑电价变化、负荷和光伏发电的不确定性以及电池老化的情况下，建立了数学模型来优化自耗和上网电价（FIT）方案下的ESS充放电。研究了上网电价和电价对光伏一体化方案的影响，并提出了考虑成本节约、系统成本和电池退化的最优ESS容量的经济分析模型。对关键参数进行敏感性分析，评估其对系统效益的影响。此外，对退役电池和新电池不同放电深度（DOD）操作的比较结果表明，退役电池可以有效减少对环境的影响，同时提供实用且经济高效的解决方案。

引用次数: 0

Robust probabilistic photovoltaic power forecasting via multitask representation and cross-attention 基于多任务表示和交叉注意的鲁棒概率光伏功率预测

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112739

Boxiao Luo, Shenglei Pei, Jinchang Cheng, Shoupeng Zhang, Tao Luo

Photovoltaic (PV) power forecasting is essential for integrating high-penetration renewable energy into power grids. However, the inherent volatility of PV power and the occurrence of extreme values pose significant challenges. These issues not only degrade the accuracy of traditional models but also hinder reliable uncertainty quantification. To address these problems, this paper proposes DD-MDN, a robust probabilistic forecasting model. Unlike conventional methods that rely on linear feature selection, our model effectively captures nonlinear dependencies among meteorological features via deep nonlinear representation learning; this module is trained with uncertainty-weighted multitask learning to obtain robust, regime-aware features. Furthermore, a cross-attention mechanism is employed to capture the dynamic influence of time-varying meteorological conditions on power fluctuations, thereby further enhancing model robustness. Through a mixture density network, DD-MDN provides a robust probabilistic characterization of the uncertainty associated with extreme power output values. Compared with various baselines, our approach demonstrates superior performance under these challenging conditions: deterministic metrics show notable improvements, and it achieves the best overall probabilistic forecasting performance. Its robustness is validated through noise-injection tests. Experimental results on both public and industrial PV datasets confirm the model’s effectiveness and reliability.

光伏发电功率预测是实现高渗透可再生能源并网的关键。然而，光伏发电固有的波动性和极值的出现带来了重大挑战。这些问题不仅降低了传统模型的准确性，而且阻碍了不确定性的可靠量化。为了解决这些问题，本文提出了一种鲁棒概率预测模型DD-MDN。与依赖线性特征选择的传统方法不同，我们的模型通过深度非线性表示学习有效地捕获了气象特征之间的非线性依赖关系；该模块使用不确定性加权多任务学习进行训练，以获得鲁棒的、状态感知的特征。此外，采用交叉关注机制捕捉时变气象条件对电力波动的动态影响，从而进一步增强模型的鲁棒性。通过混合密度网络，DD-MDN提供了与极端功率输出值相关的不确定性的鲁棒概率表征。与各种基线相比，我们的方法在这些具有挑战性的条件下表现出优越的性能：确定性度量显示出显着的改进，并且它实现了最佳的总体概率预测性能。通过噪声注入试验验证了该方法的鲁棒性。在公共和工业光伏数据集上的实验结果验证了该模型的有效性和可靠性。

{"title":"Robust probabilistic photovoltaic power forecasting via multitask representation and cross-attention","authors":"Boxiao Luo, Shenglei Pei, Jinchang Cheng, Shoupeng Zhang, Tao Luo","doi":"10.1016/j.epsr.2026.112739","DOIUrl":"10.1016/j.epsr.2026.112739","url":null,"abstract":"<div><div>Photovoltaic (PV) power forecasting is essential for integrating high-penetration renewable energy into power grids. However, the inherent volatility of PV power and the occurrence of extreme values pose significant challenges. These issues not only degrade the accuracy of traditional models but also hinder reliable uncertainty quantification. To address these problems, this paper proposes DD-MDN, a robust probabilistic forecasting model. Unlike conventional methods that rely on linear feature selection, our model effectively captures nonlinear dependencies among meteorological features via deep nonlinear representation learning; this module is trained with uncertainty-weighted multitask learning to obtain robust, regime-aware features. Furthermore, a cross-attention mechanism is employed to capture the dynamic influence of time-varying meteorological conditions on power fluctuations, thereby further enhancing model robustness. Through a mixture density network, DD-MDN provides a robust probabilistic characterization of the uncertainty associated with extreme power output values. Compared with various baselines, our approach demonstrates superior performance under these challenging conditions: deterministic metrics show notable improvements, and it achieves the best overall probabilistic forecasting performance. Its robustness is validated through noise-injection tests. Experimental results on both public and industrial PV datasets confirm the model’s effectiveness and reliability.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112739"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039048","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Total harmonic distortion and unbalance factor estimator by three-phase quadrature signal generator based on multiresonant linear oscillator 基于多谐振线性振荡器的三相正交信号发生器的总谐波失真和不平衡因子估计

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112730

Victor Aviña-Corral , Daniel Clemente-Lopez , Jose de Jesus Rangel-Magdaleno , Oscar Martínez-Fuentes , Jose Hugo Barron-Zambrano

This paper presents an advanced real-time framework for analyzing p-phase electrical networks by examining symmetric components across multiple harmonic levels. The proposed method is based on a linear oscillator and a bank of multiple resonant linear oscillators serving as quadrature signal generators, each tuned to specific harmonic frequencies, allowing simultaneous estimation of amplitude, phase angle, and phase shift for positive, negative, and homopolar sequences. A novel metric, the Total Harmonic Distortion and Unbalance Factor (THD-UF and THD-UF*), is introduced to quantify distortion and asymmetry per symmetrical sequence, providing a comprehensive indicator of power quality. The effectiveness of the proposed framework is demonstrated through simulations, embedded implementations, and measurements obtained from a real electrical grid. Specifically, the theoretical and practical analysis was carried out for a three-phase electrical network, focusing on the estimation of the fundamental component and multiple harmonic orders. From a practical perspective, the methodology was implemented on two embedded platforms to support experimental validation. In a first stage, a Raspberry Pi was used to characterize processing latency and timing behavior under constrained hardware conditions. Later, a Zynq-7000 system-on-chip was used for real-time acquisition and processing of the three-phase electrical network system, demonstrating the suitability of the proposed framework for real-time execution requirements in power-system monitoring applications.

本文提出了一种先进的实时框架，通过检查跨多个谐波水平的对称分量来分析p相电网。所提出的方法是基于一个线性振荡器和一组多个谐振线性振荡器作为正交信号发生器，每个调谐到特定的谐波频率，允许同时估计振幅，相位角，相移为正，负，和同极序列。引入了一种新的度量，总谐波失真和不平衡因子（THD-UF和THD-UF*），用于量化每个对称序列的失真和不对称，提供了一个全面的电能质量指标。通过仿真、嵌入式实现和从实际电网中获得的测量结果，证明了所提出框架的有效性。具体而言，对三相电网进行了理论和实践分析，重点是基波分量和多次谐波阶数的估计。从实践的角度来看，该方法在两个嵌入式平台上实现，以支持实验验证。在第一阶段，树莓派被用来表征在受限硬件条件下的处理延迟和定时行为。随后，Zynq-7000片上系统被用于三相电网系统的实时采集和处理，证明了所提出的框架适合电力系统监控应用的实时执行要求。

{"title":"Total harmonic distortion and unbalance factor estimator by three-phase quadrature signal generator based on multiresonant linear oscillator","authors":"Victor Aviña-Corral , Daniel Clemente-Lopez , Jose de Jesus Rangel-Magdaleno , Oscar Martínez-Fuentes , Jose Hugo Barron-Zambrano","doi":"10.1016/j.epsr.2026.112730","DOIUrl":"10.1016/j.epsr.2026.112730","url":null,"abstract":"<div><div>This paper presents an advanced real-time framework for analyzing <em>p</em>-phase electrical networks by examining symmetric components across multiple harmonic levels. The proposed method is based on a linear oscillator and a bank of multiple resonant linear oscillators serving as quadrature signal generators, each tuned to specific harmonic frequencies, allowing simultaneous estimation of amplitude, phase angle, and phase shift for positive, negative, and homopolar sequences. A novel metric, the Total Harmonic Distortion and Unbalance Factor (THD-UF and THD-UF*), is introduced to quantify distortion and asymmetry per symmetrical sequence, providing a comprehensive indicator of power quality. The effectiveness of the proposed framework is demonstrated through simulations, embedded implementations, and measurements obtained from a real electrical grid. Specifically, the theoretical and practical analysis was carried out for a three-phase electrical network, focusing on the estimation of the fundamental component and multiple harmonic orders. From a practical perspective, the methodology was implemented on two embedded platforms to support experimental validation. In a first stage, a Raspberry Pi was used to characterize processing latency and timing behavior under constrained hardware conditions. Later, a Zynq-7000 system-on-chip was used for real-time acquisition and processing of the three-phase electrical network system, demonstrating the suitability of the proposed framework for real-time execution requirements in power-system monitoring applications.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112730"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039049","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Global power sharing and virtual inertia control for the interlinking modular universal converter in standalone hybrid ac/dc microgrid 独立交直流混合微电网中互连模块化通用变换器的全局功率共享和虚拟惯性控制

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112764

Javier Gutiérrez-Escalona , Carlos Roncero-Clemente , Oleksandr Matiushkin , Mohammadreza Azizi , João Martins

Hybrid ac/dc microgrids (HMGs) have gained significant attention in recent years for their efficient integration of distributed energy resources and their capability to operate independently from the grid during islanded mode. The bidirectional interlinking converter (BIC) is the key element that coordinates power exchange between sub-grids to achieve precise power sharing. This work introduces a hierarchical control strategy providing adaptive bidirectional virtual inertia (BVI) support and precise global power sharing in a standalone HMG. Besides, the proposed control strategy was designed and validated for the recently introduced modular universal converter (MUC) topology with voltage boosting capability, demonstrating its effectiveness in performing advanced BIC functionalities such as inertia support. The hardware-in-the-loop tests conducted in real time show that the proposed strategy achieves zero power sharing error while enhancing the voltage and frequency stability through BVI transferring. Furthermore, the performance of the MUC as BIC with advanced functionalities is demonstrated, reaffirming its strong potential as a highly modular and standardized power converter solution for standalone HMGs.

近年来，交直流混合微电网因其高效集成分布式能源和在孤岛模式下独立于电网运行的能力而受到广泛关注。双向互联变换器（BIC）是协调各子电网间电力交换以实现精确电力共享的关键部件。本文介绍了一种在独立HMG中提供自适应双向虚拟惯性（BVI）支持和精确全局功率共享的分层控制策略。此外，针对最近推出的具有升压能力的模块化通用变换器（MUC）拓扑设计并验证了所提出的控制策略，证明了其在执行高级BIC功能（如惯性支持）方面的有效性。实时硬件在环测试表明，该策略在通过BVI传输提高电压和频率稳定性的同时，实现了零功率共享误差。此外，还展示了MUC作为具有先进功能的BIC的性能，重申了其作为独立hmg的高度模块化和标准化功率转换器解决方案的强大潜力。

{"title":"Global power sharing and virtual inertia control for the interlinking modular universal converter in standalone hybrid ac/dc microgrid","authors":"Javier Gutiérrez-Escalona , Carlos Roncero-Clemente , Oleksandr Matiushkin , Mohammadreza Azizi , João Martins","doi":"10.1016/j.epsr.2026.112764","DOIUrl":"10.1016/j.epsr.2026.112764","url":null,"abstract":"<div><div>Hybrid ac/dc microgrids (HMGs) have gained significant attention in recent years for their efficient integration of distributed energy resources and their capability to operate independently from the grid during islanded mode. The bidirectional interlinking converter (BIC) is the key element that coordinates power exchange between sub-grids to achieve precise power sharing. This work introduces a hierarchical control strategy providing adaptive bidirectional virtual inertia (BVI) support and precise global power sharing in a standalone HMG. Besides, the proposed control strategy was designed and validated for the recently introduced modular universal converter (MUC) topology with voltage boosting capability, demonstrating its effectiveness in performing advanced BIC functionalities such as inertia support. The hardware-in-the-loop tests conducted in real time show that the proposed strategy achieves zero power sharing error while enhancing the voltage and frequency stability through BVI transferring. Furthermore, the performance of the MUC as BIC with advanced functionalities is demonstrated, reaffirming its strong potential as a highly modular and standardized power converter solution for standalone HMGs.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112764"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039099","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Predictive cooperative terminal agents for decentralised self-healing in multi-terminal HVDC grids 多终端高压直流电网分散自愈的预测合作终端代理

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112747

Sami Astan , Amin Hajizadeh

The large-scale integration of offshore wind generation demands protection systems for multi-terminal high-voltage direct current (MT-HVDC) grids that can operate reliably under stochastic power injections, converter-dominated fault dynamics, and asynchronous communication. This paper proposes a Predictive Cooperative Terminal Agent (PCTA) framework, a fully decentralised and self-healing protection architecture in which each voltage source converter terminal operates as an autonomous intelligent agent. Using locally measured electrical states, predictive threshold adaptation, and event-triggered peer-to-peer coordination, agents detect faults, isolate affected components, and restore power flow in real time. A centralised-training, decentralised-execution multi-agent actor–critic reinforcement learning algorithm optimises fault detection latency, selectivity, and renewable utilisation. Robustness is enhanced through Kalman-based denoising and validated using systematic noise-sweep and high-resistance fault studies. High-fidelity digital-twin and hardware-in-the-loop simulations demonstrate sub-8 ms fault detection, selectivity exceeding 93 %, stable post-fault self-healing, and scalability to dense meshed networks.

海上风力发电的大规模集成要求多端高压直流（MT-HVDC）电网的保护系统能够在随机电力注入、变流器主导的故障动态和异步通信下可靠运行。本文提出了一种预测合作终端代理（PCTA）框架，它是一种完全分散和自修复的保护体系结构，其中每个电压源变换器终端都作为一个自主智能代理运行。利用局部测量的电状态、预测阈值适应和事件触发的点对点协调，代理可以检测故障，隔离受影响的组件，并实时恢复潮流。一种集中训练、分散执行的多智能体actor-critic强化学习算法优化了故障检测延迟、选择性和可再生利用率。鲁棒性通过卡尔曼去噪增强，并通过系统噪声扫描和高阻故障研究进行验证。高保真数字孪生和硬件在环仿真表明，故障检测低于8毫秒，选择性超过93%，故障后自修复稳定，可扩展到密集的网状网络。

{"title":"Predictive cooperative terminal agents for decentralised self-healing in multi-terminal HVDC grids","authors":"Sami Astan , Amin Hajizadeh","doi":"10.1016/j.epsr.2026.112747","DOIUrl":"10.1016/j.epsr.2026.112747","url":null,"abstract":"<div><div>The large-scale integration of offshore wind generation demands protection systems for multi-terminal high-voltage direct current (MT-HVDC) grids that can operate reliably under stochastic power injections, converter-dominated fault dynamics, and asynchronous communication. This paper proposes a Predictive Cooperative Terminal Agent (PCTA) framework, a fully decentralised and self-healing protection architecture in which each voltage source converter terminal operates as an autonomous intelligent agent. Using locally measured electrical states, predictive threshold adaptation, and event-triggered peer-to-peer coordination, agents detect faults, isolate affected components, and restore power flow in real time. A centralised-training, decentralised-execution multi-agent actor–critic reinforcement learning algorithm optimises fault detection latency, selectivity, and renewable utilisation. Robustness is enhanced through Kalman-based denoising and validated using systematic noise-sweep and high-resistance fault studies. High-fidelity digital-twin and hardware-in-the-loop simulations demonstrate sub-8 ms fault detection, selectivity exceeding 93 %, stable post-fault self-healing, and scalability to dense meshed networks.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112747"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039103","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Admittance modeling and oscillation analysis of hybrid GFL and GFM DFIGs under weak grid conditions 弱电网条件下GFL和GFM混合DFIGs的导纳建模及振荡分析

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112759

Xiaoju Lv, Ziwen Liu, Ziting Mei, Li Zhang, Feng Wu

With the development of Grid-Forming (GFM) wind turbine technology, dynamic uinteractions between hybrid Grid-Following Doubly-Fed Induction Generators (GFL DFIGs) and GFM DFIGs are intensifying, posing stability challenges, especially in weak grids. Therefore, this paper investigates the dynamic frequency coupling characteristics and instability mechanisms of hybrid systems composed of GFL and GFM DFIGs. Based on the dq coordinate system, a multiple-input multiple-output (MIMO) frequency-coupled admittance model is established for both GFL and GFM DFIGs, considering the dynamics of stator and rotor flux linkage and various control loops. Using the generalized Nyquist criterion, system oscillation mechanisms under weak grid conditions are assessed through eigenvalue trajectories of the frequency-domain loop gain matrix. The impact of different GFM configuration proportions on hybrid system stability is also explored. Finally, simulations in PSCAD validate the theoretical analysis, providing a foundation for parameter optimization and coordinated operation of hybrid GFL and GFM DFIG systems.

随着电网成形（GFM）风力发电技术的发展，混合式随网双馈感应发电机（GFL DFIGs）和GFM DFIGs之间的动态不相互作用日益加剧，对稳定性提出了挑战，特别是在弱电网中。因此，本文研究了由GFL和GFM DFIGs组成的混合系统的动态频率耦合特性和失稳机理。在dq坐标系下，考虑定子和转子磁链的动力学特性以及各种控制回路，建立了GFL和GFM两种dfig的多输入多输出频率耦合导纳模型。利用广义奈奎斯特准则，通过频域环增益矩阵的特征值轨迹评估了弱电网条件下系统的振荡机制。探讨了不同GFM构型比例对混合系统稳定性的影响。最后，在PSCAD上进行仿真验证了理论分析，为GFL和GFM混合DFIG系统的参数优化和协调运行提供了基础。

{"title":"Admittance modeling and oscillation analysis of hybrid GFL and GFM DFIGs under weak grid conditions","authors":"Xiaoju Lv, Ziwen Liu, Ziting Mei, Li Zhang, Feng Wu","doi":"10.1016/j.epsr.2026.112759","DOIUrl":"10.1016/j.epsr.2026.112759","url":null,"abstract":"<div><div>With the development of Grid-Forming (GFM) wind turbine technology, dynamic uinteractions between hybrid Grid-Following Doubly-Fed Induction Generators (GFL DFIGs) and GFM DFIGs are intensifying, posing stability challenges, especially in weak grids. Therefore, this paper investigates the dynamic frequency coupling characteristics and instability mechanisms of hybrid systems composed of GFL and GFM DFIGs. Based on the dq coordinate system, a multiple-input multiple-output (MIMO) frequency-coupled admittance model is established for both GFL and GFM DFIGs, considering the dynamics of stator and rotor flux linkage and various control loops. Using the generalized Nyquist criterion, system oscillation mechanisms under weak grid conditions are assessed through eigenvalue trajectories of the frequency-domain loop gain matrix. The impact of different GFM configuration proportions on hybrid system stability is also explored. Finally, simulations in PSCAD validate the theoretical analysis, providing a foundation for parameter optimization and coordinated operation of hybrid GFL and GFM DFIG systems.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112759"},"PeriodicalIF":4.2,"publicationDate":"2026-01-20","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039098","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Research on the topology and startup control method of diode series boost offshore DC transmission system 二极管串联升压离岸直流输电系统拓扑结构及启动控制方法研究

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-20 DOI: 10.1016/j.epsr.2026.112716

Jianqiang Liu , Sixu Liu , Jinda Li , Xinyu Tan , Tengfei Guo , Haokai Xie

Diode rectifiers in offshore wind power DC transmission systems have economic and reliability advantages. This paper proposes a topology circuit for a diode series boost-type offshore wind power collection and transmission system, investigates its operational mechanism, and analyzes issues like uneven voltage sharing in series diodes and harmonic suppression in diode rectifiers. The harmonic characteristics and current distribution from the diode rectifier are examined, and parameter design methods are developed, including for the diode rectifier, high-potential converter, and LC filter. Also, to address reverse power flow during startup, a corresponding control strategy is developed. Through phased startup, the wind farm achieves a transition from black start to stable operation. Finally, simulations verify the effectiveness of the proposed topology, parameter design methods, and control strategy.

二极管整流器在海上风电直流输电系统中具有经济、可靠的优点。提出了一种二极管串联升压式海上风电集输系统的拓扑电路，研究了其运行机理，分析了串联二极管电压分担不均匀、二极管整流器谐波抑制等问题。研究了二极管整流器的谐波特性和电流分布，提出了二极管整流器、高电位变换器和LC滤波器的参数设计方法。针对启动过程中的反向潮流问题，提出了相应的控制策略。通过分阶段启动，实现了风电场从黑启动到稳定运行的过渡。最后，通过仿真验证了所提出的拓扑结构、参数设计方法和控制策略的有效性。

引用次数: 0

Voltage harmonic suppression powering non-linear load of single-phase inverter with impedance method 用阻抗法抑制单相逆变器非线性负载的电压谐波

IF 4.2 3区工程技术 Q2 ENGINEERING, ELECTRICAL & ELECTRONIC

Electric Power Systems Research

Pub Date : 2026-01-19 DOI: 10.1016/j.epsr.2026.112733

Jian Huang , Fei Wang , Chen Wang , Huiqin Xu , Lixin Liu

This paper proposes a novel control strategy that integrates the load current feed-forward method (LCFFM) with an impedance notch filter to mitigate output-voltage distortion in single-phase inverters supplying non-linear loads. The study emphasizes the synergistic combination of LCFFM and the notch filter as its core innovation, which significantly improves both steady-state accuracy and harmonic suppression. Through output impedance modeling, the characteristics of non-linear loads and their impact on output voltage are analyzed. The design of LCFFM and its influence on output impedance are elaborated, along with an analysis of sampling and PWM-update delays. Furthermore, the impedance notch filter is designed and implemented to enhance harmonic rejection around critical frequencies. The proposed integrated approach is also evaluated in terms of dynamic and steady-state performance, accompanied by stability analysis. Experimental results demonstrate that the combined use of LCFFM and the impedance notch filter reduces the output voltage total harmonic distortion (THD) from (7.7 ± 0.2)% to (4.3 ± 0.2)% and improves steady-state voltage accuracy from 99.39% to 99.71%, validating the effectiveness and innovation of the proposed method.

本文提出了一种将负载电流前馈法（LCFFM）与阻抗陷波滤波器相结合的新型控制策略，以减轻提供非线性负载的单相逆变器输出电压畸变。该研究强调LCFFM与陷波滤波器的协同组合是其核心创新，在稳态精度和谐波抑制方面都有显著提高。通过输出阻抗建模，分析了非线性负载的特性及其对输出电压的影响。阐述了LCFFM的设计及其对输出阻抗的影响，并分析了采样和pwm更新延迟。此外，设计并实现了阻抗陷波滤波器，以增强临界频率附近的谐波抑制能力。本文还从动态性能和稳态性能两方面对所提出的综合方法进行了评价，并进行了稳定性分析。实验结果表明，LCFFM与阻抗陷波滤波器的结合使用将输出电压总谐波失真（THD）从（7.7±0.2）%降低到（4.3±0.2）%，将稳态电压精度从99.39%提高到99.71%，验证了所提方法的有效性和创新性。

{"title":"Voltage harmonic suppression powering non-linear load of single-phase inverter with impedance method","authors":"Jian Huang , Fei Wang , Chen Wang , Huiqin Xu , Lixin Liu","doi":"10.1016/j.epsr.2026.112733","DOIUrl":"10.1016/j.epsr.2026.112733","url":null,"abstract":"<div><div>This paper proposes a novel control strategy that integrates the load current feed-forward method (LCFFM) with an impedance notch filter to mitigate output-voltage distortion in single-phase inverters supplying non-linear loads. The study emphasizes the synergistic combination of LCFFM and the notch filter as its core innovation, which significantly improves both steady-state accuracy and harmonic suppression. Through output impedance modeling, the characteristics of non-linear loads and their impact on output voltage are analyzed. The design of LCFFM and its influence on output impedance are elaborated, along with an analysis of sampling and PWM-update delays. Furthermore, the impedance notch filter is designed and implemented to enhance harmonic rejection around critical frequencies. The proposed integrated approach is also evaluated in terms of dynamic and steady-state performance, accompanied by stability analysis. Experimental results demonstrate that the combined use of LCFFM and the impedance notch filter reduces the output voltage total harmonic distortion (THD) from (7.7 ± 0.2)% to (4.3 ± 0.2)% and improves steady-state voltage accuracy from 99.39% to 99.71%, validating the effectiveness and innovation of the proposed method.</div></div>","PeriodicalId":50547,"journal":{"name":"Electric Power Systems Research","volume":"255 ","pages":"Article 112733"},"PeriodicalIF":4.2,"publicationDate":"2026-01-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146039106","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":3,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0