IEEE Transactions on Sustainable Energy最新文献

{"title":"Component Sizing and Energy Management of Electric-Hydrogen Hybrid Energy Storage Systems for Solid-State-Transformer-Based Meshed Networks","authors":"Qian Xun;Yuzhen Tang;Marius Langwasser;Fei Gao;Marco Liserre;Hengzhao Yang","doi":"10.1109/TSTE.2025.3644440","DOIUrl":"https://doi.org/10.1109/TSTE.2025.3644440","url":null,"abstract":"This paper proposes a meshed distribution network architecture based on solid-state transformers (SSTs) to integrate various distributed energy resources (DERs) such as photovoltaic (PV) systems, battery energy storage systems (BESSs), and hydrogen energy storage systems (HESSs) composed of fuel cells, electrolyzers, and hydrogen tanks. Moreover, a co-design framework is developed to optimize the component sizing and energy management of an electric-hydrogen hybrid energy storage system (ESS) including a BESS and an HESS. The objective of the optimization framework is to minimize the total cost of the hybrid ESS categorized as the investment cost associated with component sizing and the operating cost related to energy management. In particular, this optimization framework explicitly considers the losses of the BESS, the HESS, and the distribution lines to more comprehensively evaluate the total cost of the hybrid ESS. In addition, convex transformations are introduced to reformulate the nonlinear equality constraints and the discrete inequality constraints into convex forms. Therefore, convex programming can be utilized to solve the optimization framework efficiently to obtain a globally optimal solution. The meshed network and the co-design framework are evaluated using a modified 59-node low-voltage AC (LVAC) grid model in the German SimBench dataset. Comprehensive simulations are performed using a 12-day dataset and a 366-day dataset of the year of 2016. Simulation results show that the meshed network leads to a better economy and a better voltage stability compared to the radial network. The optimization framework properly determines the power distribution between the BESS and the HESS based on the constraints and bounds of the ESS states such as the battery state of charge (SOC). In addition, the optimization framework is scalable and can be used to address the component sizing and energy management issues in large-scale distribution networks.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2133-2152"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148518394","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Deriving the Robust Operating Region for Multistage Dispatch of Power Systems: A Decision-Dependent Uncertainties Based Modeling Approach","authors":"Xin Zeng;Xiaodong Zheng;Haoyong Chen;Xin Yin;Zhengmao Li;Yiping Chen","doi":"10.1109/TSTE.2026.3663924","DOIUrl":"https://doi.org/10.1109/TSTE.2026.3663924","url":null,"abstract":"The flexibility of energy storage systems (ESS) is essential for mitigating the uncertainties associated with variable renewable energy (VRE) sources in power system operations. While ESS scheduling for high-renewable systems has been widely studied, accurately coordinating energy-limited ESS with intertemporal coupling under uncertainty and non-anticipative operation remains a challenging robust scheduling task in practice. In this context, a multistage robust scheduling model is developed that captures the energy-limited characteristics of ESS while enforcing non-anticipativity. The model schedules upper and lower boundaries of ESS state of energy (SOE) and generator outputs in the day-ahead timeframe and derives a robust operating region (OR) via decision-dependent uncertainty (DDU). These robust ORs enable temporal decoupling so that the multistage problem admits an equivalent, tractable two-stage formulation. A parallel Nested-C&CG algorithm is employed to compute the optimal robust OR (OR<inline-formula><tex-math>$^{mathrm{2}}$</tex-math></inline-formula>) efficiently. In addition, an enhanced look-ahead intra-day dispatch is introduced that dynamically updates a window-specific robust OR using ultra–short-term VRE forecasts and their uncertainty, while anchoring the window end to the day-ahead OR<inline-formula><tex-math>$^{mathrm{2}}$</tex-math></inline-formula> to preserve recursive feasibility; this mechanism reduces day-ahead conservatism and maintains robustness across the rolling horizon. Numerical experiments conducted on modified IEEE 6-bus and IEEE 118-bus systems demonstrate the model's advantages in robustness, economic efficiency, and tractability.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"3019-3032"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148443230","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Differentiated Adaptive Operation of R-SOP- Integrated Distribution Systems With Joint Learning","authors":"Xiaodong Yang;Zhixiang Hao;Hui Fang;Da Xu;Wei Sun;Qiuwei Wu;Lijian Ding;Jinyu Wen","doi":"10.1109/TSTE.2025.3638704","DOIUrl":"https://doi.org/10.1109/TSTE.2025.3638704","url":null,"abstract":"Traditional fixed and single control mode suffers from poor adaptability and low decision-making precision, making it difficult to meet the differentiated regulation requirements of distribution systems (DSs). This paper proposes a differentiated adaptive operation (DAO) framework for DSs to adapt to complex operational requirements. Firstly, a new topology of reconfigurable soft open point (R-SOP) is introduced into the DAO framework to enhance the flexibility and adaptability of DSs. However, the action space of the R-SOP includes both discrete and continuous variables with inherent interdependencies, which poses challenges for implementing rapid decision-making in time-varying environments. To this end, a joint learning agent is developed based on soft actor-critic and double deep Q networks to achieve real-time accurate scheduling for R-SOP. On this basis, a R-SOP assisted differentiated adaptive operation method is proposed for the first time, which integrates an adaptive switching mechanism that enables flexible switching among multiple control modes based on real-time measurement data, thereby accommodating the diverse operational requirements of DSs. Finally, numerical simulations on modified IEEE 33-node, IEEE 123-node and an actual 220-node test systems validate the effectiveness of the proposed method.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"1967-1982"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148518569","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Novel Federated Learning Framework With Adaptive Parameter Updater for Wind Power Forecasting","authors":"Wang Han;Zhang Ziyun;Yan Jie;Zhang Ning;Zhang Jiawei;Liu Yongqian","doi":"10.1109/TSTE.2025.3639633","DOIUrl":"https://doi.org/10.1109/TSTE.2025.3639633","url":null,"abstract":"Current research indicates that the integrated power forecasting for multiple wind farms can effectively improve prediction accuracy, but this approach requires data open sharing and strict temporal synchronization, posing challenges for practical applications. Federated learning serves as a promising solution to these issues. However, it exhibits limitations in handling Non-Independent and Identically Distributed (non-IID) data and demonstrates poor local adaptability for participating wind farms, while being susceptible to interference from external anomalous data. To overcome the above challenges, this paper proposes a wind power forecasting method with Adaptive Parameter Updater, by integrating and enhancing federated learning framework and fine-tuning strategies of meta-learning. On the one hand, for each participating wind farm, it leverages local data to perform personalized fine-tuning of the returned global model. On the other hand, it enables simultaneous pre-training of global model and fine-tuning of local model parameters, thereby resolving convergence difficulties caused by non-IID data distribution across wind farms and preventing information loss. This method achieves high-precision wind power forecasting while preserving data privacy, particularly benefiting data-scarce wind farms. The proposed method is applicable to any temporal scale and forecasting model. This paper takes day-ahead forecasting as an example and constructs power forecasting models using CNN, GRU, and BERT, respectively. Two datasets are used to validate the effectiveness and robustness of the proposed method. The results show that the proposed method achieves an average error reduction of 2.12 percentage points under data-rich scenarios, while attaining an average decrease of 4.18 percentage points under datascarce conditions.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2016-2027"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148519162","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Degradation-Aware Real-Time Power Allocation Approach for Heterogeneous BESS","authors":"Jiaxiong Xiang;Ming Zhou;Yi Ren;Chang Li;Zhi Zhang;Zhaoyuan Wu","doi":"10.1109/TSTE.2026.3655105","DOIUrl":"https://doi.org/10.1109/TSTE.2026.3655105","url":null,"abstract":"With growing hybrid configurations and phased deployments, rising heterogeneity in battery energy storage system (BESS) challenges the cost-minimization power allocation due to the difficulty in accurately quantifying real-time degradation cost. To address this, we propose a dual energy distribution model which stores the cycle information related to SoC trajectories within the energy distribution, tackling the problem to seek the full cycle for current action with past operation. A real-time degradation cost calculation formulation is presented to analytically express nonlinear degradation over a continuous action space, without requiring historical SoC backtracking or the rainflow simulation. A real-time power allocation algorithm is designed based on dual energy distribution ranking. Unlike traditional rainflow-based online methods that rely on estimating the marginal degradation cost of infinitesimal actions, our approach allocates power based on the degradation cost evaluated across the entire power range of each battery. Numerical tests and comparison are done for heterogeneous BESS response to frequency regulation service, the results show our method reduces daily degradation costs by 12% to 31% compared to conventional strategies, and with fast solving speed of each response within milliseconds.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2587-2601"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148519394","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Optimal Operation of Grid-Connected Renewable Power to Ammonia System Considering the Dynamic Time Interval Regulation Strategy","authors":"Chunyan Li;Kaiwen Hou;Tao Wu;Kaigui Xie;Ruoxuan Wang","doi":"10.1109/TSTE.2025.3647108","DOIUrl":"https://doi.org/10.1109/TSTE.2025.3647108","url":null,"abstract":"Renewable power to ammonia (RePtA) is a promising pathway to facilitate large-scale integration of green energy and serves as a key technological route for deep decarbonization of the chemical industry. However, the inherent volatility of renewable energy (RE) generation challenges the secure operation of ammonia synthesis (AS), demonstrating the urgent need to develop flexible operational strategies for RePtA systems. This paper proposes an optimal scheduling model for the grid-connected RePtA system. A dynamic time interval regulation (DTIR) strategy of AS is first proposed to enhance operational flexibility. This strategy decomposes each regulation interval into a fixed-duration transition stage and a variable-duration steady stage. The variable durations are jointly optimized with the quasi-steady-state ammonia loads. Then, an equivalent transformation and structural simplification approach is developed to derive a tractable MILP reformulation. Finally, the entropy weighted confidence level-based distributionally information gap decision theory method is introduced to address multiple RE uncertainties. Numerical simulations on a grid-connected RePtA project are conducted to verify the effectiveness of the proposed framework.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2333-2349"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148517183","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Preference-Oriented Evolutionary Multi-Objective Energy-Noise Optimization for Fast Wind Farm Layout Design","authors":"Ya-Zhi Zhang;Sheng Huang;Juan Wei;Bing-Chuan Wang;Yun Feng;Yaonan Wang","doi":"10.1109/TSTE.2026.3661110","DOIUrl":"https://doi.org/10.1109/TSTE.2026.3661110","url":null,"abstract":"Mitigating wind farm noise through layout optimization while maintaining power production under varying noise limits remains challenging. To reduce workload and decision time, and engage DMs more in optimization, this study proposes a preference-oriented evolutionary multi-objective (EMO) framework for wind farm layout optimization that incorporates decision-makers’ (DMs’) preferences through interactively updated reference points. The framework guides the search toward DMs’ region of interest effectively while preserving population diversity. Case studies on a real-world wind farm show that the proposed algorithm outperforms four state-of-the-art EMO algorithms and a traditional layout method. Compared with the real-world layouts, the resulting solutions reduce the noise level by 10.45% –23.04% and increase the annual energy production by 3.03% –4.09%, while aligning with DMs’ preferences.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2898-2910"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148517824","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Active Structural Load Mitigation in Wind Turbines Using Linear Unique Model Predictive Control","authors":"Rafael Barcena;Ainhoa Etxebarria","doi":"10.1109/TSTE.2025.3641374","DOIUrl":"https://doi.org/10.1109/TSTE.2025.3641374","url":null,"abstract":"This paper describes a new model-based predictive controller, specially designed to enable its practical implementation. In that sense, it uses a unique linear model for the entire turbine operating range. To achieve this, external processing blocks are used to handle the nonlinearities present in the rotor aerodynamics. The objective of this active vibration control system is to reduce the structural load on the rotor and tower of wind turbines, without the need to use specific structural control actuators (e.g., tuned mass dampers). In fact, special effort is put in analyzing the additional effort to be made by the collective pitch actuators, to evaluate the circumstances in which the addition of specific control devices is advisable. Also, the proposed system works in parallel to the baseline controller in charge of the rotor speed control, thus avoiding the substitution of the latter.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2081-2094"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148518392","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Bridging Market Efficiency and System Resilience: Unified Multi-Service Operation of Long-Duration Energy Storage in Decarbonized Power Systems","authors":"Airin Rahman;Wei Sun","doi":"10.1109/TSTE.2026.3659001","DOIUrl":"https://doi.org/10.1109/TSTE.2026.3659001","url":null,"abstract":"The rapid growth of variable renewable generation and the increasing frequency of extreme events demand power systems that are both economical and resilient. Long-duration energy storage (LDES) stands out as a unique solution, capable of shifting energy across multiple days while providing critical support during large-scale outages. However, our review of existing approaches highlights critical gaps: market operations and system restoration are typically optimized in isolation, and the technology-specific characteristics of LDES make their coordinated use especially challenging. We first investigate these challenges, including efficiency-capacity trade-offs, duration mismatches, and restoration-sequence dependencies, and underscore the need for an integrated framework. To address this gap, this paper introduces a unified multi-service optimization model that co-optimizes LDES for both economic and resilience objectives within a single mixed-integer linear program. The formulation combines energy arbitrage, capacity adequacy, and renewable integration with resilience metrics based on avoided unserved energy and a priority-based restoration mechanism capturing spatiotemporal re-energization and progressive load pickup. Multi-objective Pareto analysis reveals cost-resilience trade-offs, providing actionable insights for system planners. To demonstrate technology-agnostic applicability, the model is applied to two contrasting LDES technologies, hydrogen (H<inline-formula><tex-math>$_{2}$</tex-math></inline-formula>) and flow battery (FB), on a benchmark test system with seasonal PV and load profiles under diverse blackout scenarios. Simulation results show that the proposed framework reduces operating costs in normal conditions while accelerating critical-load recovery and system-wide restoration after disruptions, advancing the state-of-the-art in storage-integrated power system operation.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2759-2772"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148515581","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A Hybrid Data-Model Driven Method for Generating Typical Operating Conditions of UHVDC Sending-End Power Grids With High-Penetration Renewable Energy","authors":"Lei Wu;Pupu Chao;Weixing Li;Zhimin Li","doi":"10.1109/TSTE.2025.3640144","DOIUrl":"https://doi.org/10.1109/TSTE.2025.3640144","url":null,"abstract":"With the integration of high-penetration renewable energy, the short-term voltage stabilities of modern power systems have fundamentally changed, making traditional typical operating conditions inadequate to represent the running boundaries, especially in ultra-high voltage direct current (UHVDC) sending-end power grids. In response to this issue, this paper proposes a hybrid data-model driven method for generating new typical operating conditions (composed of extreme and central ones). Considering various voltage stability patterns in sending-end power grids with high-penetration renewable energy, an extremity index system to describe the stability degree of the operating conditions is proposed with a new definition of typical operating conditions. Furthermore, an optimal dispatch model is developed, accounting for the seasonal and volatile nature of renewable energy scenarios. The annual operating conditions for future grid are generated through chronological operation simulation. The local outlier factor (LOF) algorithm is used to detect the initial set of extreme operating conditions, and a hybrid data-model driven approach is employed to extract the final extreme operating conditions from initial set. The LOF and K-medoids algorithm are then applied to generate central operating conditions efficiently by eliminating the influence of extreme operating points. Case studies demonstrate that the proposed method effectively enhances the accuracy of identifying stability boundaries and operating centers in power grids compared with traditional ones.","PeriodicalId":452,"journal":{"name":"IEEE Transactions on Sustainable Energy","volume":"17 3","pages":"2028-2049"},"PeriodicalIF":10.3,"publicationDate":"2026-07-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"148519163","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}