Search

Caching-enabled multi-access edge computing (MEC) has attracted wide attention to support future intelligent vehicular networks, especially for delivering high-definition videos in the internet of vehicles with limited backhaul capacity. However, factors such as the constrained storage capacity of MEC servers and the mobility of vehicles pose challenges to caching reliability, particularly for supporting multiple bitrate video streaming caching while achieving considerable quality of experience (QoE). Motivated by the above challenges, in this paper, we propose an intelligent caching strategy that takes into account vehicle mobility, time-varying content popularity, and backhaul capability to improve the QoE of vehicle users effectively. First, based on the mobile video mean opinion score (MV-MOS), we designed an average download percentage (ADP) weighted QoE evaluation model. Then, the video content caching problem is formulated as a Markov decision process (MDP) to maximize the ADP weighted MV-MOS. Owing to the prior knowledge of video content popularity and channel state information that may not be available at the road side unit in practical scenarios, we propose a deep reinforcement learning (DRL)-based caching strategy to solve the problem while achieving a maximum ADP weighted MV-MOS. To accelerate its convergence speed, we further integrate the prioritized experience replay, dueling, and double deep Q-network technologies, which improve the performance of DRL algorithm. Numerical results demonstrate that the proposed DRL-based caching strategy significantly improves QoE, and achieves better video delivery reliability compared to existing non-learning approaches. [ABSTRACT FROM AUTHOR]

Hydrofoils play a crucial role in enhancing the efficiency of fluid machinery designed for ocean environments, reducing lift-induced drag and contributing to improved overall performance. To optimize hydrofoil design, a profound comprehension of the complex fluid flows around the hydrofoil structure is essential. In fluid mechanics, a precise and continuous representation of flow dynamics is essential for analysis and control purposes. Thus, obtaining a complete flow field, either through computational fluid dynamics (CFD) or experimental testing is of great significance. Nevertheless, the rigorous requirements of flow field tests render it impractical to directly measure complete flows around the airfoil using current instrumentation, especially for those with complex physical geometries. To tackle this issue, a novel deep learning framework is proposed to reconstruct the complete flow field by leveraging incomplete complementary flow data. As the representative benchmark problems, the flows around the Clark-Y hydrofoil at Re = 7 × 105 and the experimental NACA0012 2D hydrofoil at Re = 1800 have been investigated under different missing-flow scenarios of varying proportions, locations and orientations. Results demonstrate a remarkable agreement between the reconstructed flow field and the ground truth data, indicating the excellent performance of the proposed deep-learning model for missing flow reconstruction. A sensitivity analysis assesses the impact of the snapshot number and the latent space, revealing the method's robustness in selecting these parameters and simplifying its implementation in practical applications. This deep learning method offers the advantage of being implemented using paired incomplete flow fields, without the need for pre-known ground truth results as labels, holding the potential for addressing more complex full-field reconstruction problems in the future. [Display omitted] • Develop novel MS-AE framework for missing flow field reconstruction, surpassing fused POD and AE methods. • Conduct sensitivity analysis of AE-SE method to assess crucial parameters, ensuring reliability. • Apply unsupervised machine learning and multi-scale approach for innovative fluid mechanics reconstruction. [ABSTRACT FROM AUTHOR]

A precise and cost-effective prediction tool for fluid-structure interaction (FSI) analysis is crucial for optimizing the structural design of tidal turbine blades. However, the high computational costs associated with fluid dynamic analysis pose a significant challenge, as the current lack of efficient FSI prediction methods hinders the advancement of cutting-edge tidal turbine designs. To address this issue, this paper proposes a novel consolidated framework that integrates deep learning convolutional neural networks (CNN) with blade element momentum (BEM) theory and finite element method (FEM) to perform deformation analysis of turbine blade structures. The proposed CNN-BEM-FEM integrated framework efficiently identifies the geometric features and predicts the hydrodynamic parameters of turbine blades and thus, achieving accurate assessments of the structural behavior of tidal turbines. The study applies two-step verification procedures to validate the prediction accuracy of the CNN-BEM-FEM framework and the result demonstrates excellent agreement with experimental tests for hydrodynamic performance and blade deformation. When compared with the static one-way FSI calculated by Ansys Workbench software, the computational efficiency of CNN-BEM-FEM framework increases by more than 18 times, with discrepancies in blade deformation and equivalent stress calculations generally less than 5 %. By applying the proposed method to predict the FSI performance of tidal turbine blades with various shear web structures, the practical applicability for composite turbine blade design is successfully demonstrated. The results underscore the potential of the CNN-BEM-FEM framework as an efficient and accurate prediction tool for optimizing the structural design of tidal turbine blades. • A high-efficient framework that combines CNN, BEM and FEM for tidal turbine rotor structural analysis is proposed. • Convolutional neural network (CNN) is used for identifying hydrodynamic performance of turbine blade section. • Blade element momentum (BEM) is for turbine prediction, while finite element method (FEM) is for structural analysis. • It shows great agreement with experiments, ensuring precise prediction of turbine blade deformations and hydrodynamics. • It outperforms benchmark software by more than 18 times, demonstrating exceptional computational efficiency. [ABSTRACT FROM AUTHOR]

Real-time acquisition of dynamic wake field information has garnered substantial attention in the wind farm industry, as it provides a crucial data source for intelligent wind farm monitoring and control. However, the existing wind measurement technologies, such as light detection and ranging (LiDAR), are limited to sparse data point measurements. This paper explores the utilization of a Physics-Informed Neural Network (PINN) to reconstruct wind turbine wake dynamics, specifically focusing on the influence of active wind turbine yaw operation on wake evolution. The methodology involves creating a tailored loss function that combines sparse wake measurement data with the Navier-Stokes (NS) equations. More precisely, the neural network incorporates the NS equations as constraints to guide the prediction of physical quantities in the output, including downwind velocity, crosswind velocity, and pressure. Taking the dynamic wake during yawing as a case study, the proposed method showcases remarkable universality and robustness across diverse scenarios involving varying scanning angle intervals, measurement point spacings, frequencies, and noise levels. It successfully captures the dynamic trends in wake evolution during yawing and accurately forecasts the wake trajectory and deflection. Even when tested with actual wake measurement data, the method can still effectively reconstruct the flow field, indicating significant potential for the real wind farm yaw control. • Dynamic spatiotemporal wake reconstruction of a utility-scale yawed wind turbine is carried out in this paper. • A novel approach combining physics-informed neural network with sparse LiDAR measurement is proposed. • The neural network maps spatiotemporal coordinates [ t , x , y ] to flow physical quantities [ u , v , p ] integrated with NS equations. • Various cases with different LiDAR measurement parameters are investigated for sensitivity and robustness analysis. • The new approach shows great effectiveness in reconstructing dynamic wake behavior of trajectory and deflection changes. [ABSTRACT FROM AUTHOR]

Complete and clear global wind turbine wake data is very important for the study of wind turbine wake characteristics in increasingly large offshore wind farms. Existing wake measurement techniques can only obtain local high-resolution (HR) wake flow field, or sacrifice accuracy to obtain larger measurement area, which is insufficient for accurate modeling of wake effect. To overcome this challenge, this paper proposes a novel super-resolution (SR) reconstruction approach that can reconstruct the global HR wake flow field from low-resolution (LR) wake flow field measurement data effectively. The proposed approach utilizes a deep learning framework called down-sampled skip-connection and multi-scale network. The performance of the SR approach is evaluated by enhancing the resolution of the wake flow field at different scale factors, and its potential application is demonstrated by assessing the prediction accuracy of three typical wake models. The results indicate that the resolution of the global wind turbine wake can be improved by 16 times using the SR model, and the reconstructed global SR wake flow fields are consistent with the ground truth in terms of both the spatial distribution and the temporal variation. By comparing the prediction results of three different wake models with the LR or SR wake data, it is shown that the SR flow reconstruction method can be applied to more accurately evaluate the wake model prediction performance, which has the potential to improve wake models. Overall, this study presents an innovative solution to the problem of incomplete and inaccurate wake flow measurement in the wind energy industry, which could reduce the workload of experimental measurements and the cost burden of accurate measuring equipment for engineering applications. • A novel deep learning method is proposed to facilitate high resolution wake flow measurement of offshore wind turbine. • The deep learning framework adopts down-sampled skip-connection and multi-scale network. • Super-resolution flow field is achieved by reconstructing global high-resolution wake from local low-resolution wake. • Resolution of reconstructed wake flow attains 16 times improvement with great accuracy in spatiotemporal distribution. • Super-resolution reconstruction framework has been successfully applied for better assessing the prediction accuracy of traditional wake models. [ABSTRACT FROM AUTHOR]

A tidal turbine can benefit from exquisitely designed morphing blades with a flexible trailing edge by mitigating up to 90% of the load fluctuation in harsh ocean environments, which reduces the overall cost of tidal energy. However, existing fluid–structure interaction (FSI) methods of resolving flow-induced deformation of the blades is computationally expensive, which poses an important challenge to effective morphing blade design. This paper presents a novel static FSI tool based on deep learning to cost-efficiently analyze the fluid–structure coupling of a hydrofoil. Specifically, adopting a convolutional neural network (CNN) to predict the fluid force and finite element method (FEM) to solve the solid structure response, a new CNN-FEM framework with an iterative scheme for solving the FSI problem is developed to achieve equilibrium between the fluid and structural forces. The new framework is used to predict the elastic deformation of the flexible blade section of the hydrofoil to demonstrate its effectiveness in the FSI evaluation. Comparison of the results to those produced by commercially developed software (i.e., Ansys Workbench) shows that this method yields extremely close prediction results of average equivalent stress and an accuracy of more than 92%. Moreover, it is 100 times more computationally efficient than the commercial Ansys Workbench software, requiring less than 3s for one-way FSI calculation. Taking advantage of this cost-effectiveness, the CNN-FEM can be used to achieve the accurate prediction of the deformation characteristics of the flexible hydrofoil under various flow scenarios that lay a foundation for advanced morphing blade design in the future. [ABSTRACT FROM AUTHOR]

Horizontal axis tidal turbines (HATT) conversion of ocean tidal waves into electricity represents a promising source of clean and sustainable energy. However, the widespread adoption of these turbines has been hindered by persistent challenges, primarily stemming from the high costs associated with their construction and maintenance; and efficient conversion of tidal energy. Addressing these challenges is paramount for propelling tidal turbine technology and ensuring its economic viability. This study focuses on the efficient conversion of tidal energy into electricity by optimization of composite material turbine blades which is a complex problem that spans multiple physical domains, including hydrodynamics (the study of water flow) and structural mechanics (the study of material behavior under loads). To tackle these multifaceted challenges, we introduce an innovative DLFSI (Deep learning fluid-structure interaction) model which represents a groundbreaking approach to predict and optimize the hydrodynamic and structural performance of tidal turbine blades. DLFSI leverages the power of convolutional neural networks (CNN) to recognize intricate geometric features of turbine blades rapidly and accurately. By seamlessly integrating the blade element momentum (BEM) theory and finite element method (FEM), the DLFSI model facilitates comprehensive predictions of how composite blades will perform in real-world conditions. With this approach, we have achieved substantial improvements in critical performance metrics such as the power coefficient (a measure of energy conversion efficiency) and the maximum equivalent stress (a key indicator of structural integrity). The innovative DLFSI model presented in this study holds the potential for practical application within the realm of tidal turbine design and is poised to catalyze the sustainable progression of renewable energy technologies. [Display omitted] [ABSTRACT FROM AUTHOR]

The paper proposes a novel cost-effective framework that combines deep learning convolutional neural network (CNN) and blade element momentum (BEM) models for optimizing the performance of three-dimensional (3D) horizontal axis tidal turbines (HATTs). The framework employs signed distance function (SDF) to reconstruct the three-dimensional blade geometry, utilizes CNN to identify the hydrodynamic performance of each blade section, and ultimately predicts the performance of HATT using BEM. On top of the new CNN-BEM model, the rotor blade geometrical optimization by multi-objective non-dominated sorting genetic algorithm (NSGA-II) is carried out to obtain a better trade-off solution with the maximal power coefficient of turbine and minimal hydrodynamic load exerted on the blades. The results show that the CNN-BEM model has good agreement with experimental data and reduces prediction time by 46.7% compared to the conventional Xfoil-BEM model, while reducing general optimization time by 20.1%. The new model's cost-efficiency allows for a better trade-off solution with reduced hydrodynamic load while maintaining the power coefficient. Thus, the proposed model has the capability to deliver both accurate and fast prediction and optimization of HATT performance, making it a valuable tool for guiding the design of tidal turbines. • A cost-effective framework coupling convolutional neural network (CNN) and blade element momentum (BEM). • The new coupled CNN-BEM model is applied for tidal turbine blade optimization. • Trade-off with maximal power coefficient and minimal hydrodynamic load is achieved by multi-objective optimization. • The new model reduces prediction time by 46.7% and general optimization time by 20.1%. • The new model's cost-efficiency allows for better trade-off solution than conventional Xfoil-BEM model. [ABSTRACT FROM AUTHOR]

Convolutional Neural Network (CNN) is a commonly used deep learning algorithm due to its excellent capability in identification of structural features and parameter predictions in many domains. In addition, it has incomparable advantages of high analysis efficiency and generalization performance. However, it has been questioned in the research community on whether CNN method can be applied to effectively predict hydrofoil performance for hydraulic machinery design. To this end, this paper demonstrates a novel optimization platform using CNN for hydrofoil performance prediction, which can effectively and accurately obtain the optimized hydrofoils results in aid of the structural design of tidal turbine. The prediction model uses signed distance function (SDF) to graphically represent the shape of the hydrofoil which is subsequently imported into CNN as the network input. Three different hydrofoil performance properties including the lift coefficient, drag coefficient and pressure coefficient of surface are used as output to train the neural network. In order to guarantee the accuracy of the forecasting model, Computational Fluid Dynamics (CFD) method characterized by high precision is applied to generate the dataset for neural network training. The results show that it can accurately predict the hydrodynamic parameters at a lower angle of attack with extremely short period of time. On top of the established hydrofoil performance prediction model, the Pareto curve of the optimized hydrofoils is obtained and applied to the design of 3D horizontal axis tidal turbine (HATT) blades. It proves that the optimization platform is effective and versatile in a manner that achieves both accurate and rapid prediction/optimization of the hydrofoil, which greatly facilitates to apply it for the tidal turbine rotor design. • A novel platform for hydrofoil shape optimization based on CNN and CFD is established. • Through CNN, properties of hydrofoil including lift, drag, and pressure coefficients are accurately predicted. • By incorporating NSGA-Ⅱ, Pareto front of optimized hydrofoil shapes from UIUC database is achieved. • Optimized HATT yields the maximum 45% power coefficient and a wider high-efficiency operational range. • Best hydrofoil selection for HATT rotor design is concluded. [ABSTRACT FROM AUTHOR]