Article
Green & Sustainable Science & Technology
Mohammad Mehdi Hosseini, Masood Parvania
Summary: This paper proposes a model that combines deep reinforcement learning and mathematical optimization to operate distributed energy storage systems in distribution grids. By utilizing the fast response capability of deep reinforcement learning and keeping network constraints in check with mathematical optimization, the paper addresses the issues caused by load and renewable generation uncertainties.
IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
(2023)
Article
Green & Sustainable Science & Technology
Mohammad Mehdi Hosseini, Masood Parvania
Summary: This paper proposes a model that combines deep reinforcement learning (DRL) and mathematical optimization for the operation of energy storage systems (ESS) in distribution grids. The model utilizes the fast response capability of DRL while maintaining network constraints, enabling efficient operation of ESS under uncertain load and renewable generation conditions.
IEEE TRANSACTIONS ON SUSTAINABLE ENERGY
(2023)
Article
Energy & Fuels
Tianjing Wang, Yong Tang
Summary: A transfer-reinforcement-learning-based rescheduling method for differential power grids considering security constraints was proposed to address practical defects, achieve better transfer learning effects, and tackle security issues in different scenarios. It narrowed the action space of deep reinforcement learning, applied domain-adaption transfer learning for different structural changes within the same power grid and policy-based transfer learning for different power grids, showing improved effectiveness and lower control costs compared to other methods.
Article
Engineering, Electrical & Electronic
Tianqiao Zhao, Jianhui Wang, Meng Yue
Summary: This paper proposes a general solution framework for traditional control problems in modern power systems. The framework consists of a model-free controller and a barrier-certification system, which utilize reinforcement learning and control barrier functions. Neural networks are used to represent the barrier function, addressing the challenge of calculating it for complex power systems. The effectiveness of the proposed framework is demonstrated through several case studies.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Huifeng Zhang, Dong Yue, Chunxia Dou, Gerhard P. Hancke
Summary: This article proposes an optimal defensive strategy for microgrids with a distributed DRL approach, which evaluates the impact of FDI attacks and deduces optimal network weight to ensure the economic and security issues of the microgrids system.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2022)
Article
Engineering, Electrical & Electronic
Lipeng Zhu, Yonghong Luo
Summary: This paper proposes an intelligent data-driven predictive UVLS scheme, using a deep feedback learning machine (DFLM) to accurately predict future voltage violations, and introduces two strategies to improve the scheme's reliability and adaptability.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2021)
Article
Multidisciplinary Sciences
Ferenc Molnar, Takashi Nishikawa, Adilson E. Motter
Summary: This study demonstrates that improving power grid stability through heterogeneity among generators is more effective than homogeneity, showcasing the phenomenon of converse symmetry breaking. The findings present a new method for identifying this counterintuitive effect in other networks relying on behavioral homogeneity.
NATURE COMMUNICATIONS
(2021)
Review
Automation & Control Systems
Tao Liu, Yue Song, Lipeng Zhu, David J. Hill
Summary: This article surveys classic and novel results on the stability and control of power grids, providing a perspective on the ongoing transition and development of new stability and control paradigms.
ANNUAL REVIEW OF CONTROL ROBOTICS AND AUTONOMOUS SYSTEMS
(2022)
Article
Automation & Control Systems
Jian Sun, Guanqiu Qi, Neal Mazur, Zhiqin Zhu
Summary: With the rapid increase in data measurement from power grids, machine learning research in transient control has gained significant attention. This article proposes a sparse neural network based reinforcement learning scheme for optimizing the transient stability enhancement of power grids with energy storage systems. The simulation results confirm the feasibility, advantages, and adaptability of the proposed method.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Energy & Fuels
Jizhe Liu, Yuchen Zhang, Ke Meng, Zhao Yang Dong, Yan Xu, Siming Han
Summary: This paper proposes a risk-averse deep learning method for real-time emergency load shedding. By training deep neural networks, this method is more inclined to avoid load undercutting events, thereby reducing the significant costs incurred by control failure.
Article
Engineering, Electrical & Electronic
Heling Yuan, Yan Xu, Cuo Zhang
Summary: This paper proposes a robust optimization method to address the impact of wind power generators on transient stability of a power system. By considering uncertain wind power output, coordinating generation dispatch and emergency load shedding, this method offers an effective solution.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2022)
Article
Engineering, Electrical & Electronic
Keyu Wu, Min Wu, Zhenghua Chen, Ruibing Jin, Wei Cui, Zhiguang Cao, Xiaoli Li
Summary: Domain adaptation transfers knowledge from label-rich source domains to label-scarce target domains for generalized learning in new environments. Partial domain adaptation (PDA) extends this concept by considering scenarios where the target label space is a subset of the source label space. This paper proposes a Reinforced Adaptation Network (RAN) that combines deep reinforcement learning with domain adaptation techniques to address the challenging PDA problem. Experimental results show that RAN significantly outperforms existing state-of-the-art methods on three benchmark datasets.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS FOR VIDEO TECHNOLOGY
(2023)
Article
Energy & Fuels
Tianjing Wang, Yong Tang
Summary: This paper proposes an automatic voltage control method based on transfer learning and deep reinforcement learning for differential power grids. The method takes into account the magnitude and number of voltage deviations in the reward function. Constrained multi-agent deep reinforcement learning is used to develop the AVC method. Distribution adaptation transfer learning and parameter-based transfer learning are introduced for different transfer circumstances. The efficacy of the method is tested using two IEEE systems and two real-world power grids.
CSEE JOURNAL OF POWER AND ENERGY SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Zhongtuo Shi, Wei Yao, Yong Tang, Xiaomeng Ai, Jinyu Wen, Shijie Cheng
Summary: This paper proposes a bidirectional active transfer learning (Bi-ATL) framework for more adaptive power system stability assessment (PSSA) and dominant instability mode (DIM) identification. The framework integrates forward active learning and backward active learning to achieve more efficient adaptation.
IEEE TRANSACTIONS ON POWER SYSTEMS
(2023)
Article
Computer Science, Artificial Intelligence
Jingyi Zhang, Yonghong Luo, Boya Wang, Chao Lu, Jennie Si, Jie Song
Summary: In this study, an innovative solution approach to the challenging dynamic load-shedding problem in large power grids is introduced. The proposed approach, called DQN-LS, takes into account spatial and temporal information of the power system and uses a ConvLSTM network to capture dynamic features. It provides real-time, fast, and accurate load-shedding decisions to increase the quality and probability of voltage recovery.
IEEE TRANSACTIONS ON NEURAL NETWORKS AND LEARNING SYSTEMS
(2023)
Article
Engineering, Electrical & Electronic
Yuxin Liu, Chunhua Liu, Rundong Huang, Zaixin Song
Summary: One-to-multiple wireless power transfer (O2M-WPT) is convenient and flexible but has drawbacks, such as burning out of the primary coil and inability to achieve constant-current or constant voltage output. This study proposes a primary multi-frequency constant-current compensation (MFCC) network to solve these problems. The MFCC allows the primary side to act as a current source at multiple frequencies, enabling the receivers to achieve CC or CV output. Experimental results validate the effectiveness of the MFCC network for O2M-WPT systems.
IEEE TRANSACTIONS ON CIRCUITS AND SYSTEMS II-EXPRESS BRIEFS
(2023)
Article
Engineering, Electrical & Electronic
Wei Gao, Rui Fan, Renke Huang, Qiuhua Huang, Wenzhong Gao, Liang Du
Summary: This paper proposes a novel reinforcement learning based HVDC oscillation damping controller using the augmented random search (ARS) algorithm, which leverages wide-area information to effectively damp inter-area oscillations.
ELECTRIC POWER SYSTEMS RESEARCH
(2023)
Article
Automation & Control Systems
Yong Chen, Chunhua Liu, Senyi Liu, Zaixin Song
Summary: In this article, a novel cascaded adaptive deadbeat (CADB) control method is proposed for permanent magnet synchronous motor drives. An adaptive deadbeat (DB) based current controller is first proposed with a simplified first-order current loop dynamic model. The motor parameters are compressed into few identifiable coefficients, and an improved gradient method is used to identify these time-varying coefficients. Then, a similar model of speed loop is presented and employed for the design of an adaptive speed controller. The stability of the proposed adaptive controller is confirmed by using the Lyapunov theorem. Finally, the proposed CADB control system is experimentally carried out to validate its performance.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Automation & Control Systems
Zaixin Song, Zhiping Dong, Wusen Wang, Senyi Liu, Chunhua Liu
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Automation & Control Systems
Wusen Wang, Chunhua Liu, Zaixin Song, Zhiping Dong
Summary: Harmonic currents in dual three-phase permanent magnet synchronous machines are effectively suppressed using a deadbeat predictive harmonic current control (DPHCC) method. The DPHCC method calculates the harmonic control voltage based on sampled harmonic currents and the machine model, thereby improving both steady-state and transient-state performance. Additionally, a disturbance observer is implemented to handle inaccurate machine parameters and unmodeled disturbances, utilizing discrete-time sliding mode theory and an improved sliding mode reaching law. Experimental results confirm the effectiveness of the proposed control system in suppressing harmonic currents.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Automation & Control Systems
Wusen Wang, Chunhua Liu, Hang Zhao, Zaixin Song
Summary: This article presents two simplified DB-DTFC algorithms and designs improved observers to enhance the performance and parameter robustness of deadbeat-direct torque and flux control in permanent magnet synchronous machines.
IEEE TRANSACTIONS ON INDUSTRIAL ELECTRONICS
(2023)
Article
Engineering, Electrical & Electronic
Zaixin Song, Rundong Huang, Wusen Wang, Zhiping Dong, Chunhua Liu
Summary: This research aims to provide a current harmonic control strategy for low-reactance open-winding permanent magnet synchronous motor (LR OW-PMSM) by combining nonlinear lumped disturbance observer (NLDO) and improved space vector modulation (SVM) scheme. Experimental results show that the proposed NLDO, compared with ordinary PI and deadbeat controllers, can reduce current harmonics on both the motor and inverter sides and has high robustness, making it practical for LR motors.
IEEE TRANSACTIONS ON VEHICULAR TECHNOLOGY
(2023)
Article
Engineering, Electrical & Electronic
Zhiping Dong, Yuxin Liu, Bowen Zhang, Chunhua Liu
Summary: This article introduces a half open-end winding topology based nine-leg voltage source inverter (HONL-VSI) for driving an asymmetrical six-phase permanent magnet synchronous motor (ASP-PMSM). Three winding connection modes (WCMs) are proposed for different output requirements. The features of HONL-VSIs with three WCMs are investigated and compared with existing topologies for ASP-PMSM drives. The feasibility of the HONL-VSI with three WCMs is verified through simulation and experiment results.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2023)
Article
Engineering, Electrical & Electronic
Zhiping Dong, Hao Wen, Zaixin Song, Chunhua Liu
Summary: In this article, a three-dimensional space vector modulation (3-D SVM) strategy for three-phase open-end winding drives with common dc bus is proposed, which can be conducted in both a-b-c and alpha-beta-z coordinates, leading to full utilization of the dc bus voltage and convenient over-modulation adjustment. The voltage vector distributions in the 3-D spaces with a-b-c and alpha-beta-z coordinates are revealed, and three adjacent voltage vectors are selected for reference voltage vector synthesis based on the relationship between leg-voltage potentials. Furthermore, the overmodulation in different coordinates is also investigated. Experimental results validate the feasibility of the proposed 3-D SVM for OWDs with common dc bus.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2023)
Article
Engineering, Electrical & Electronic
Yang Xiao, Yong Yang, Chunhua Liu, Jose Rodriguez
Summary: In this study, a novel wireless motor system called the wireless ultrasonic motor (WUM) system is proposed and analyzed. By combining wireless power transfer and ultrasonic motor technology, the system addresses the issues of complex controllers and converters, regular maintenance, and output ripple. Matching inductors and optimized coil structures are designed to improve power quality and filter harmonic voltages.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2023)
Article
Engineering, Electrical & Electronic
Yong Chen, Chunhua Liu, Hao Wen, Zhiping Dong
Summary: This article considers a multimotor system supplied by the reduced-switch-count voltage source inverter (VSI). An improved multiple vectors model predictive control (MV-MPC) scheme is proposed for driving n three-phase permanent magnet synchronous motors (PMSMs) using (2n + 1)-leg VSI. The proposed scheme allows independent control of each motor by partitioning the control period into multi-intervals. The experiment results demonstrate the effectiveness of the MV-MPC method in reducing current ripple and achieving a faster dynamic response, as well as the overcurrent elimination method in minimizing overcurrent regardless of motor load conditions.
IEEE TRANSACTIONS ON POWER ELECTRONICS
(2023)
Article
Energy & Fuels
Cong Wang, Suangshuang Jin, Renke Huang, Qiuhua Huang, Yousu Chen
Summary: Dynamic contingency analysis is crucial for modern power systems to anticipate potential issues and enhance system stabilities. This research presents a two-level hierarchical computing architecture implemented on GPUs to accelerate the intensive computations of massive DCA. The experimental results demonstrate up to 2.8x and 4.2x speedup using one and two GPUs, respectively, compared to a CPU-based parallel approach. The proposed architecture significantly improves the overall computational performance of massive DCAs while maintaining strong scaling capability under various resource configurations.
IEEE OPEN ACCESS JOURNAL OF POWER AND ENERGY
(2023)