Article
Engineering, Civil
Ye Qiu, Rui Yu, Bingbing San, Jianhong Li
Summary: This study investigates the possibility of using surrogate models combined with genetic algorithm for aerodynamic shape optimization of large-span coal sheds. Results show that the optimal design presents a reasonable performance improvement, and various compromised designs are provided to seek a balance between wind directionality and construction cost.
ENGINEERING STRUCTURES
(2022)
Article
Construction & Building Technology
Fadi Alkhatib, Narimah Kasim, Wan Inn Goh, Nasir Shafiq, Mugahed Amran, Evgenii Vladimirovich Kotov, Mohammed Abdo Albaom
Summary: Wind-induced loads and motions are crucial in the design of tall buildings. However, contemporary architecture trends make it challenging to evaluate these motions. This paper proposes a computational performance-based aerodynamic optimization method to assist architects and engineers in seeking optimal design decisions.
Article
Automation & Control Systems
Jingjie Xie, Hongyang Dong, Xiaowei Zhao
Summary: This paper proposes a novel reinforcement learning (RL)-based control scheme to reduce structural loads and regulate power generation of floating offshore wind turbines (FOWTs) simultaneously. The proposed method integrates the online-learned FOWT dynamics into the control process and only learns partial system dynamics, which simplifies the design structure and improves learning efficiency.
IEEE TRANSACTIONS ON AUTOMATION SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Aerospace
Jichao Li, Mengqi Zhang
Summary: Geometric filtering based on deep-learning models has been shown to effectively improve the efficiency of aerodynamic shape optimization without excluding innovative aerodynamic shapes. Specific cases validate the application of geometric filtering in aerodynamic shape optimization and showcase its advantages in different scenarios.
AEROSPACE SCIENCE AND TECHNOLOGY
(2021)
Article
Engineering, Mechanical
Giovanni Solari, Patricia Martin
Summary: This paper addressed the dominant topic of wind engineering related to the loading and response of structures due to gust buffeting, focusing on the shape of vibration modes. It clarified the relationship linking aerodynamic admittance with mode shape and provided a closed-form expression for any aerodynamic admittance. The solution is simple for modes with few changes of sign, but becomes laborious as mode complexity increases.
JOURNAL OF ENGINEERING MECHANICS
(2021)
Article
Engineering, Aerospace
Jiaqi Liu, Rongqian Chen, Jinhua Lou, Hao Wu, Yancheng You, Zhengwu Chen
Summary: This article proposes a deep reinforcement learning-based framework for rotor airfoil optimization, which utilizes a trained neural network as a surrogate model to learn and apply optimization strategies for improving the aerodynamic performance of the rotor. It demonstrates the interpretability and generalizability of the optimization strategy learned through deep reinforcement learning.
AEROSPACE SCIENCE AND TECHNOLOGY
(2023)
Article
Engineering, Marine
Ngo Van He, Ngo Van Hien, Ngoc-Tam Bui
Summary: In this study, the wind drag acting on a passenger ship's above water hull was computed using the commercial Computational Fluid Dynamics (CFD) code ANSYS-Fluent. The friction viscous wind drag and pressure viscous wind drag acting on the hull were investigated, and new hull forms with different frontal accommodation shapes were proposed to reduce wind drag and improve aerodynamic performance. A new above water surface hull with reduced wind drag was identified in this study by comparing CFD results.
Article
Acoustics
Kevin Volkmer, Nicholas Kaufmann, Thomas H. Carolus
Summary: This study investigated three modifications of blades for small horizontal axis wind turbines in order to reduce noise emissions, finding that boundary layer tripping and trailing edge serrations can decrease noise but at the expense of turbine shaft power.
JOURNAL OF SOUND AND VIBRATION
(2021)
Article
Engineering, Mechanical
Giovanni Solari, Patricia Martin
Summary: Studies have shown that mode shape plays a significant role in determining the type of solution for gust buffeting and aerodynamic admittance problems. A simple closed-form solution is possible if the mode does not change sign, while a numerical approach is required if the mode changes sign. Proper orthogonal decomposition can provide a full conceptual interpretation and a simple closed-form solution for these problems.
JOURNAL OF ENGINEERING MECHANICS
(2021)
Article
Engineering, Aerospace
Manish Tripathi, M. Manikandan, Pranshul Pandey, Rajkumar S. Pant
Summary: This study compared the aerodynamic performance of trilobed and single-lobed airships, revealing that trilobed airships have higher aerodynamic efficiency but lower stability and higher drag. By analyzing the fluid dynamics, a webbed trilobed hull variant was proposed to reduce drag at the expense of lift.
JOURNAL OF AEROSPACE ENGINEERING
(2023)
Article
Engineering, Marine
Gang Yao, Yuxiao Chen, Yang Yang, Yuanlin Zheng, Hongbo Du, Linjun Wu
Summary: Wind tunnel tests were conducted to investigate the vortex-induced vibration (VIV) characteristics of a large-span double-deck truss girder bridge and measure the VIV suppression effect of aerodynamic mitigation measures. The results showed that the double-deck truss girder exhibited significant VIV at wind attack angles of +3° and +5°, and the aerodynamic mitigation measures had an influence on VIV response. The upper chord fairing and lower chord inverted L-shaped deflector plate played crucial roles in suppressing VIV. Numerical analysis indicated that vortex shedding above the upper deck or in the wake region dominated vertical VIV, while vortex shedding in the wake region of the lower deck dominated torsional VIV. The upper and lower chord structures disrupted the original vortex shedding pattern in both regions, thereby suppressing VIV. This research provides a foundation for bridge design and vibration suppression measures for large-span double-deck truss girder bridges.
JOURNAL OF MARINE SCIENCE AND ENGINEERING
(2023)
Article
Engineering, Multidisciplinary
Xinshuai Zhang, Fangfang Xie, Tingwei Ji, Zaoxu Zhu, Yao Zheng
Summary: The study establishes an effective optimization framework of aerodynamic shape design based on the multi-fidelity deep neural network (MFDNN) model, achieving high accuracy by blending different fidelity information. The proposed framework significantly improves optimization efficiency and outperforms single-fidelity methods.
COMPUTER METHODS IN APPLIED MECHANICS AND ENGINEERING
(2021)
Article
Nanoscience & Nanotechnology
Xinyu Hui, Hui Wang, Wenqiang Li, Junqiang Bai, Fei Qin, Guoqiang He
Summary: The study introduces a method using proximal policy optimization (PPO) for multi-object aerodynamic design optimization, which shows higher efficiency and accuracy compared to traditional algorithms.
Article
Engineering, Civil
Chaorong Zheng, Zhaoyong Wang, Jitong Zhang, Yue Wu, Zhao Jin, Yong Chen
Summary: Experimental investigation on the amplitude characteristics of wind loads on a square cross-sectional tall building under combined aerodynamic control shows that shape modifications have significant control effects on the drag coefficients and lift coefficients. The combined aerodynamic control is more effective in reducing drag than passive or active aerodynamic control individually, with Model 2 demonstrating the best wind-resistant performance.
ENGINEERING STRUCTURES
(2021)
Article
Computer Science, Artificial Intelligence
Yifan Zhang, Peilin Zhao, Qingyao Wu, Bin Li, Junzhou Huang, Mingkui Tan
Summary: Portfolio selection is a challenging task in finance, and this paper proposes a cost-sensitive method using deep reinforcement learning. The proposed method extracts price series patterns and asset correlations, and controls both transaction and risk costs effectively. Empirical results demonstrate its effectiveness and superiority in profitability, cost-sensitivity, and representation abilities.
IEEE TRANSACTIONS ON KNOWLEDGE AND DATA ENGINEERING
(2022)
Article
Engineering, Civil
Reda Snaiki, Teng Wu, Andrew S. Whittaker, Joseph F. Atkinson
TRANSPORTATION RESEARCH RECORD
(2020)
Article
Engineering, Civil
Reda Snaiki, Teng Wu
JOURNAL OF WIND ENGINEERING AND INDUSTRIAL AERODYNAMICS
(2020)
Article
Engineering, Mechanical
Shaopeng Li, Reda Snaiki, Teng Wu
Summary: The optimized deep reinforcement learning control scheme can effectively simulate complex wind fields in a multiple-fan wind tunnel, without the need for expensive fluid dynamics modeling and time-consuming manual adjustment of control parameters.
JOURNAL OF ENGINEERING MECHANICS
(2021)
Article
Environmental Sciences
Reda Snaiki, Teng Wu
Summary: A knowledge-enhanced deep learning (KEDL) method is developed in this study to estimate the boundary-layer winds associated with extratropical cyclones (ETCs) over eastern North America. The KEDL integrates physics-based equations and semi-empirical formulas as part of the system loss function to enhance the deep neural network, providing accurate and efficient estimation of ETC wind fields.
Article
Engineering, Civil
Shaopeng Li, Teng Wu
Summary: The uncertain weather and traffic conditions caused by hurricanes make managing transportation infrastructure a stochastic decision problem. This study proposes a deep reinforcement learning-based decision support system to minimize network-level losses caused by hurricanes.
Article
Environmental Sciences
Minglong Lu, Shaopeng Li, Teng Wu
Summary: This study investigates the impacts of solitary waves on box girders and develops a fast prediction model based on computational fluid dynamics (CFD) simulations and artificial neural network (ANN). The maximum wave forces show a non-linear relationship with the deck aspect ratio (W/h) for relatively large relative wave heights (H/h) and small submergence coefficients (C-s). The trained ANN-based model has a high prediction accuracy of 98.6% for vertical wave forces and 98.1% for horizontal wave forces.
Article
Chemistry, Multidisciplinary
Xingyu An, Shaopeng Li, Teng Wu
Summary: With the increase in long-span bridges and the threat of intense hurricanes under climate change, mitigating flutter risk in bridge design is crucial. This study proposes an efficient reduced-order model based on LSTM network to simulate nonlinear aeroelastic forces on bridge decks and predict their post-flutter behavior.
APPLIED SCIENCES-BASEL
(2023)
Article
Engineering, Civil
Shaopeng Li, Teng Wu
Summary: To reduce threats and losses caused by hurricanes and traffic conditions, stakeholders need to make decisions considering uncertainties. A risk-informed decision-support framework is proposed for optimal operation of hurricane-impacted transportation networks, incorporating modules for hurricane hazard, transportation infrastructure, and traffic flow. The framework utilizes low-dimensional modeling approaches and a reinforcement learning-based decision-making tool to optimize traffic safety and mobility.
NATURAL HAZARDS REVIEW
(2023)
Article
Construction & Building Technology
Teng Wu, Jiachen He, Shaopeng Li
Summary: This study proposes a nonlinear model-free controller based on deep reinforcement learning for active flutter control of long-span bridges. A deep neural network is used to approximate the nonlinear functions and map the system state to the control command. The performance of the proposed scheme is demonstrated through numerical examples.
WIND AND STRUCTURES
(2023)
Review
Construction & Building Technology
Teng Wu, Reda Snaiki
Summary: This review examines the state of research and practice of machine learning (ML) applications in wind engineering, providing a comprehensive summary of its utilization in various topic areas and identifying critical challenges and prospects for future research efforts.
FRONTIERS IN BUILT ENVIRONMENT
(2022)
Article
Construction & Building Technology
Reda Snaiki, Teng Wu
FRONTIERS IN BUILT ENVIRONMENT
(2020)
Article
Engineering, Multidisciplinary
Lin Zhao, Xi Xie, Teng Wu, Shao-peng Li, Zhi-peng Li, Yao-jun Ge, Ahsan Kareem
JOURNAL OF ZHEJIANG UNIVERSITY-SCIENCE A
(2020)