相关参考文献
注意:仅列出部分参考文献,下载原文获取全部文献信息。
Article
Energy & Fuels
Zhiwen Deng et al.
Summary: In recent years, the focus on wind farm wake control has been to maximize wind farm production. This study investigates yaw optimization for wind farm production maximization from the perspective of time-varying wakes. A simplified dynamic wake model is deduced and verified through simulation comparisons. The impact of wake propagation time lag on wind farm production and an optimization method using the proposed dynamic wake model are presented.
Article
Thermodynamics
Zhiwen Deng et al.
Summary: This study presents a decentralized yaw optimization method based on deep reinforcement learning (DRL) to maximize the power production of wind farms. The Multi-Agent Deep Deterministic Policy Gradient (MADDPG) algorithm is applied to design separate agents for each turbine, allowing independent yaw decisions. A novel analytical dynamic wake model is derived to dynamically reflect the wake propagation of the wind turbine, and the proposed method is tested through wind farm simulation. The results demonstrate that the proposed method significantly increases power generation and effectively guides MADDPG towards the optimal strategy.
ENERGY CONVERSION AND MANAGEMENT
(2023)
Article
Automation & Control Systems
Hongyang Dong et al.
Summary: This article proposes a novel data-driven control scheme to maximize the total power output of wind farms subject to strong aerodynamic interactions among wind turbines. The proposed method is model-free and has strong robustness, adaptability, and applicability. The effectiveness, robustness, and scalability of the proposed method are tested by prototypical case studies with a dynamic wind farm simulator.
IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY
(2023)
Article
Mechanics
Aurelien Vadrot et al.
Summary: This paper focuses on the use of reinforcement learning (RL) for near-wall turbulence modeling. A new RL wall model (WM) called VYBA23 is developed, which uses agents dispersed in the flow near the wall. The model is trained on a single Reynolds number and does not rely on high-fidelity data. The results show potential for developing RLWMs that can recover physical laws and for extending this type of ML models to more complex flows in the future.
Article
Physics, Fluids & Plasmas
C. Vignon et al.
Article
Physics, Fluids & Plasmas
Aurelien Vadrot et al.
Summary: This survey investigated the use of data-driven machine learning techniques for wall modeling in large-eddy simulations (LES). Three machine learning wall models were implemented and compared with an equilibrium wall model in LES of half-channel flow at various Reynolds numbers. Results showed promise in data-driven ML wall models, although some models had errors at certain Reynolds numbers.
PHYSICAL REVIEW FLUIDS
(2023)
Article
Engineering, Civil
Navid Zehtabiyan-Rezaie et al.
Summary: Analytical wake models require formulas to replicate the effect of wind turbines on turbulence levels in the wake region. One such formula proposed by A. Crespo, J. Hernandez in 1996 relates added turbulence to turbine induction factor, ambient turbulence intensity, and normalized distance from the rotor using one coefficient and three exponents. However, an incorrect exponent for ambient turbulence intensity has been mistakenly used in the literature. This study implemented both the correct and incorrect formulations of turbine-induced added turbulence in a Gaussian wake model to assess their impact on the Horns Rev 1 wind farm. Results indicate differences of 1.94% and 3.53% in predicted turbulence intensity and normalized power of waked turbines between the correct and incorrect formulations, respectively, at an ambient turbulence intensity of 7.7%. These discrepancies increase to 2.7% and 4.95% at an ambient turbulence intensity of 4%.
JOURNAL OF WIND ENGINEERING AND INDUSTRIAL AERODYNAMICS
(2023)
Article
Green & Sustainable Science & Technology
Yubao Zhang et al.
Summary: "This study proposes a novel communication-based multi-agent deep reinforcement learning approach for controlling power generation in large-scale wind farms. By introducing a multivariate power model to analyze the impact of wake effects, and using a hierarchical communication multi-agent proximal policy optimization algorithm to coordinate continuous controls, the proposed approach significantly increases wind farm power output. Importantly, there is no significant increase in wind turbine blade fatigue damage as the wind farm scale increases."
Article
Green & Sustainable Science & Technology
Venkata Ramakrishna Padullaparthi et al.
Summary: The paper introduces a multi-agent deep reinforcement learning method called FALCON for coordinated control of wind farms, which addresses the trade-off between energy and fatigue damage by jointly controlling the pitch and yaw of all turbines. FALCON achieves scale by using multiple reinforcement learning agents and efficiently capturing the global state-space, leading to better performance compared to baseline PID controllers and learning-based distributed control in a real-world wind farm case study.
Article
Automation & Control Systems
Hongyang Dong et al.
Summary: In this study, a deep reinforcement learning-based control approach with enhanced learning efficiency and effectiveness is proposed to optimize the total power production of wind farms. By introducing a novel composite experience replay strategy and modified importance-sampling weights, the method successfully handles the challenges posed by strong wake effects among wind turbines and the stochastic features of environments, achieving higher rewards with less training costs compared to conventional deep RL-based control approaches.
IEEE TRANSACTIONS ON CONTROL SYSTEMS TECHNOLOGY
(2022)
Article
Automation & Control Systems
Hongyang Dong et al.
Summary: This article presents a novel preview-based robust deep reinforcement learning method for wind-farm power tracking problem, which can handle tasks subject to uncertain environmental conditions and strong aerodynamic interactions among wind turbines. The control problem is transformed into a zero-sum game to quantify the influence of unknown wind conditions and future reference signals.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Article
Automation & Control Systems
Jingjie Xie et al.
Summary: This article proposes a model-free deep reinforcement learning (DRL) method for wind farm control, aiming to maximize the total power generation. By combining induction control and yaw control, a novel double-network (DN)-based DRL approach is designed. The simulation results show that this method can significantly increase power generation for wind farms with different layouts.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2022)
Review
Green & Sustainable Science & Technology
Navid Zehtabiyan-Rezaie et al.
Summary: With the increasing number of wind farms, research in wind-farm flow modeling is shifting towards data-driven techniques. However, the complexity of fluid flows in real wind farms poses unique challenges for data-driven modeling, requiring models to be interpretable and have some degree of generalizability.
JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY
(2022)
Review
Energy & Fuels
Daniel R. Houck
Summary: The advancement of wind turbine technology has led to optimized machines, but the issue of wake interference in wind farms hinders their efficiency and longevity. Research on wake management techniques is crucial for maximizing performance and reducing costs in the wind energy industry.
Article
Energy & Fuels
Marion Coquelet et al.
Summary: Individual pitch control can reduce the oscillating loads experienced by wind turbine blades. This study presents a novel controller structure that uses a neural network trained with reinforcement learning to achieve load reduction and handle turbulent flows.
Article
Multidisciplinary Sciences
H. Jane Bae et al.
Summary: Researchers propose a multi-agent reinforcement learning approach to discover wall models for large-eddy simulations, solving the challenge of capturing near-wall dynamics in turbulent flow simulations.
NATURE COMMUNICATIONS
(2022)
Article
Engineering, Electrical & Electronic
Binghao He et al.
Summary: The wake effect is a major obstacle in wind farm power generation, and choosing a suitable wake model that balances computational cost and accuracy is difficult. This study proposes an ensemble-based DRL wind farm control framework, introducing the Actor Bagging Deep Deterministic Policy Gradient algorithm to address the high cost issue of DRL. Experimental results show that this method can learn the optimal control policy with lower learning cost and a more robust learning process.
INTERNATIONAL JOURNAL OF ELECTRICAL POWER & ENERGY SYSTEMS
(2022)
Article
Mechanics
Ali Eidi et al.
Summary: This study quantifies the model-form uncertainties in RANS simulations using a data-driven machine-learning technique. By applying a two-step feature-selection method and the extreme gradient boosting algorithm, more accurate representations of the Reynolds stress anisotropy are obtained. The proposed framework provides optimal estimation of uncertainty bounds for the RANS-predicted quantities of interest.
Article
Energy & Fuels
Michael F. Howland et al.
Summary: In this study, a validated model for implementing collective operation of wind turbines is reported, which increases the energy production of wind farms. By designing a control protocol, the energy production of wind farms can be improved at different wind speeds.
Article
Green & Sustainable Science & Technology
Johan Meyers et al.
Summary: Control of wind farms is a crucial research topic with challenges including complex physics, uncertainties in load prediction, and multidisciplinary design optimization. Key research areas for enabling commercial success include flow physics understanding, algorithms and AI, validation and industry implementation, and integrating control with system design.
WIND ENERGY SCIENCE
(2022)
Article
Computer Science, Interdisciplinary Applications
Ricardo Vinuesa et al.
Summary: Machine learning is rapidly integrating into scientific computing, offering significant opportunities for advancing computational fluid dynamics. Key areas of impact include accelerating numerical simulations, enhancing turbulence modeling, and developing simplified models, while potential limitations should also be taken into consideration.
NATURE COMPUTATIONAL SCIENCE
(2022)
Review
Energy & Fuels
Hongyang Dong et al.
Summary: Wind power plays a vital role in global efforts towards achieving net zero emissions, and wind farm control technologies are crucial for enhancing the efficiency of wind energy utilization. This paper provides a comprehensive review of the development and recent advances in wind farm control technologies, covering system modeling, main challenges, and control objectives. Different control methods for various purposes are investigated, and the differences and similarities between model-based, model-free, and data-driven wind farm approaches are discussed. The paper also highlights the latest wind farm control technologies based on reinforcement learning, a rapidly developing machine learning technique that has garnered global attention. Furthermore, future challenges and research directions in wind farm control are analyzed.
PROGRESS IN ENERGY
(2022)
Review
Automation & Control Systems
Carl R. Shapiro et al.
Summary: This paper discusses the importance of the dynamic changes in the turbulent atmospheric boundary layer in wind farm energy production and the optimization of control approaches. Studying the dynamics of the turbulent flow field is beneficial for improving wind farm control efficiency and plays a crucial role in the transition of wind farms into major electricity suppliers.
ANNUAL REVIEW OF CONTROL ROBOTICS AND AUTONOMOUS SYSTEMS
(2022)
Review
Thermodynamics
Ryan Nash et al.
Summary: The overall power production of a wind farm is often lower than its nominal power due to the aerodynamic wake effects between wind turbines. Researchers have been working on layout optimization and control strategies to minimize this loss. Several active wake control strategies have been proposed and studied to decrease power loss of downstream turbines by manipulating or weakening the upstream wakes.
ENERGY CONVERSION AND MANAGEMENT
(2021)
Article
Computer Science, Interdisciplinary Applications
Jonathan Viquerat et al.
Summary: Deep Reinforcement Learning (DRL) has achieved remarkable achievements in various domains within physics and engineering, but there is still much to be explored before the capabilities of these methods are well understood. This paper presents the first application of DRL to direct shape optimization, demonstrating that an artificial neural network trained through DRL can generate optimal shapes autonomously, paving the way to new generic shape optimization strategies in fluid mechanics and other domains where relevant reward functions can be defined.
JOURNAL OF COMPUTATIONAL PHYSICS
(2021)
Article
Energy & Fuels
Gustav A. Speakman et al.
Summary: This study applies wake steering to multirotor turbines and uses large-eddy simulations to analyze the impact of rotor yaw on wake velocity deficit distribution and magnitude. The results show that rotor yaw can significantly improve power production for downstream turbines by expanding, channeling, or redirecting wakes. Lower-fidelity models are compared with LES data and show reasonable agreement in capturing wake trends over a large streamwise range.
Article
Energy & Fuels
Hongyang Dong et al.
Summary: This study proposes a wind farm control scheme based on deep reinforcement learning, which uses a reward regularization module and composite learning controller to optimize power generation and yaw tracking. The scheme demonstrates robustness and adaptability in handling uncertain wind conditions.
Article
Green & Sustainable Science & Technology
P. Stanfel et al.
Summary: This paper presents a proof-of-concept distributed reinforcement learning framework for maximizing wind farm energy capture, utilizing Q-learning in a wake-delayed wind farm environment with time-varying wind conditions. The proposed algorithm modifications create the GARLIC framework for optimizing wind farm energy capture in time-varying conditions, which is compared to the static lookup table wind farm controller baseline FLORIS.
JOURNAL OF RENEWABLE AND SUSTAINABLE ENERGY
(2021)
Proceedings Paper
Energy & Fuels
Henry Korb et al.
Summary: In this study, Reinforcement Learning is applied for wind farm control for the first time, optimizing the total power production by tuning existing control parameters and extending to multiple turbines. New control strategies based on generator torque control were attempted, with the tuned helix approach showing an average increase in total power output.
WAKE CONFERENCE 2021
(2021)
Review
Mechanics
Steven L. Brunton et al.
ANNUAL REVIEW OF FLUID MECHANICS, VOL 52
(2020)
Review
Meteorology & Atmospheric Sciences
Fernando Porte-Agel et al.
BOUNDARY-LAYER METEOROLOGY
(2020)
Proceedings Paper
Automation & Control Systems
Zhiwei Xu et al.
Article
Automation & Control Systems
Huan Zhao et al.
IEEE TRANSACTIONS ON INDUSTRIAL INFORMATICS
(2020)
Article
Green & Sustainable Science & Technology
Eric Simley et al.
WIND ENERGY SCIENCE
(2020)
Review
Mechanics
Karthik Duraisamy et al.
ANNUAL REVIEW OF FLUID MECHANICS, VOL 51
(2019)
Article
Green & Sustainable Science & Technology
Nicolo Gionfra et al.
Review
Engineering, Civil
Ali C. Kheirabadi et al.
JOURNAL OF WIND ENGINEERING AND INDUSTRIAL AERODYNAMICS
(2019)
Review
Multidisciplinary Sciences
Paul Veers et al.
Review
Energy & Fuels
Cristina L. Archer et al.
Review
Green & Sustainable Science & Technology
Tuhfe Gocmen et al.
RENEWABLE & SUSTAINABLE ENERGY REVIEWS
(2016)
Article
Multidisciplinary Sciences
Volodymyr Mnih et al.
Review
Engineering, Civil
D. Mehta et al.
JOURNAL OF WIND ENGINEERING AND INDUSTRIAL AERODYNAMICS
(2014)
Article
Green & Sustainable Science & Technology
Majid Bastankhah et al.
Article
Energy & Fuels
H. Aa Madsen et al.
Review
Mechanics
Tongguang Wang
THEORETICAL AND APPLIED MECHANICS LETTERS
(2012)
Article
Engineering, Mechanical
Martin O. L. Hansen et al.
JOURNAL OF FLUIDS ENGINEERING-TRANSACTIONS OF THE ASME
(2011)
Review
Energy & Fuels
B. Sanderse et al.
Review
Neurosciences
Joseph O'Neill et al.
TRENDS IN NEUROSCIENCES
(2010)
Article
Computer Science, Artificial Intelligence
Jan Peters et al.
Review
Engineering, Aerospace
LJ Vermeer et al.
PROGRESS IN AEROSPACE SCIENCES
(2003)