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
Mechanics
Sangseung Lee et al.
Summary: This study proposes a transfer learning framework for modeling drag on irregular rough surfaces with a limited dataset. The transfer learning of empirical correlations can improve the performance of neural networks for drag prediction.
JOURNAL OF FLUID MECHANICS
(2022)
Article
Mechanics
Mustafa Z. Yousif et al.
Summary: In this paper, an efficient method based on deep neural networks for generating turbulent inflow conditions is proposed. The method combines a multiscale convolutional auto-encoder with a subpixel convolution layer (MSCSP-AE) and a long short-term memory (LSTM) model. The model is able to generate realistic turbulent inflow conditions with accurate statistics and spectra, as compared with the ground truth data. The performance of the model is evaluated using direct numerical simulation (DNS) data of turbulent channel flow, and it demonstrates good ability in predicting flow fields and producing turbulence statistics and spectral content similar to those obtained from the DNS. The effects of changing various components in the model and the impact of transfer learning (TL) using different amounts of training data are thoroughly investigated. The results show that the method can significantly reduce the training time without affecting the performance of the model, and using physics-guided deep-learning-based models can be efficient in predicting turbulent flows with relatively low computational cost.
JOURNAL OF FLUID MECHANICS
(2022)
Article
Physics, Multidisciplinary
Martin Oberlack et al.
Summary: The calculation of turbulence statistics is a key problem in fluid mechanics, and recent research has used symmetry theory to derive scaling laws for turbulent moments and validate them through experiments, leading to important results.
PHYSICAL REVIEW LETTERS
(2022)
Article
Mechanics
Mustafa Z. Yousif et al.
Summary: This study presents a model for turbulent wake control using artificial neural networks and proper orthogonal decomposition. The model can predict the dynamics of the flow field with low computational cost. Long short-term memory neural networks and bidirectional long short-term memory neural networks are used for predicting the temporal evolution of time coefficients at different heights. Transfer learning is utilized to improve the prediction capability of the networks. The results show that the bidirectional long short-term memory neural network performs better in terms of training and prediction error.
Article
Mechanics
Mustafa Z. Yousif et al.
Summary: This study presents a deep learning-based framework that utilizes the concept of generative adversarial networks to recover high-resolution turbulent velocity fields from extremely low-resolution data. The model, a multiscale enhanced super-resolution generative adversarial network, accurately reconstructs high-resolution velocity fields, as demonstrated by evaluating its performance using direct numerical simulation data. The results show that the model is capable of reconstructing high-resolution velocity fields at different down-sampling factors and within the range of the training Reynolds numbers.
Article
Physics, Fluids & Plasmas
Sergio Hoyas et al.
Summary: A new direct numerical simulation of a Poiseuille channel flow with a friction Reynolds number of 10000 has been conducted. Results show a longer logarithmic layer of the mean streamwise velocity than previously thought. The maximum intensity of the streamwise velocity increases with the Reynolds number, but the elusive second maximum has not yet appeared. The scaling of the turbulent budgets in the center of the channel is almost perfect above 1000 wall units, while the peak of the pressure intensity grows with the Reynolds number and does not scale in wall units.
PHYSICAL REVIEW FLUIDS
(2022)
Article
Physics, Multidisciplinary
D. Schmekel et al.
Summary: Turbulent flow is chaotic in nature and impossible to predict far into the future. Coherent structures are regions exhibiting specific physical behaviors in turbulent flow, with this study focusing on structures connected with Reynolds stresses. Deep-learning techniques have shown promising results in modeling turbulence, with a convolutional neural network successfully predicting the number and volume of coherent structures in a turbulent channel flow.
FRONTIERS IN PHYSICS
(2022)
Article
Engineering, Mechanical
Kai Fukami et al.
Summary: This paper presents a machine learning technique that accurately estimates the state of turbulent flow in engineering systems using limited sensor measurements. The technique can reconstruct turbulent vortical structures in a pump sump from sparse surface pressure measurements and accurately estimate flow with only a few sensor measurements, identifying the presence of adverse vortices.
JOURNAL OF FLUIDS ENGINEERING-TRANSACTIONS OF THE ASME
(2022)
Article
Computer Science, Artificial Intelligence
Hamidreza Eivazi et al.
Summary: Modal-decomposition techniques are important for capturing dominant flow features. This research proposes a deep probabilistic-neural-network architecture for learning non-linear modes from turbulent-flow data. The method extracts non-linear modes and encourages the learning of independent latent variables. By constraining the shape of the latent space, a set of parsimonious modes can be extracted for high-quality flow reconstruction.
EXPERT SYSTEMS WITH APPLICATIONS
(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
Physics, Fluids & Plasmas
Ricardo Vinuesa et al.
Summary: This review summarizes current trends in flow control techniques aimed at improving the aerodynamic efficiency of wings, covering active methods for controlling turbulence and separation, various levels of modeling relying on turbulence simulation, as well as data-driven approaches with a focus on deep reinforcement learning. It concludes that this methodology has the potential to discover novel control strategies in complex turbulent flows relevant to aerodynamics.
Review
Computer Science, Artificial Intelligence
Syed Muhammad Arsalan Bashir et al.
Summary: This article provides an overview of recent advancements in single-image super-resolution (SR) research, particularly focusing on the application of deep learning methods. It categorizes image SR methods into four groups and introduces image quality metrics, reference datasets, and challenges in SR. State-of-the-art image SR methods such as EDSR, CinCGAN, MSRN are evaluated and discussed.
PEERJ COMPUTER SCIENCE
(2021)
Article
Mechanics
Hyojin Kim et al.
Summary: This study presents an unsupervised learning model using CycleGAN technology for super-resolution reconstruction of turbulent flows with unpaired data. The model demonstrates high performance in recovering original flow fields, reconstructing full-resolution fields, and generating DNS-resolution flow fields. The research also shows the feasibility of unsupervised learning in turbulence data, showcasing potential for wide application of super-resolution reconstruction techniques.
JOURNAL OF FLUID MECHANICS
(2021)
Article
Mechanics
Luca Guastoni et al.
Summary: Two convolutional neural network models were trained to predict velocity-fluctuation fields in turbulent open-channel flow, with one directly predicting fluctuations and the other reconstructing flow fields using a linear combination of orthonormal basis functions. Both models outperformed traditional methods in predicting nonlinear interactions in the flow.
JOURNAL OF FLUID MECHANICS
(2021)
Article
Mechanics
Mustafa Z. Yousif et al.
Summary: In this study, a deep learning-based approach is used to reconstruct high-resolution turbulent flow fields using a multi-scale enhanced super-resolution generative adversarial network. The model demonstrates good performance in reconstructing high-resolution laminar flows and wall-bounded turbulent flow fields. Additionally, the potential for transfer learning in turbulent channel flow is thoroughly examined, showing the effectiveness of reducing training data and time while maintaining model performance.
Article
Mechanics
Taichi Nakamura et al.
Summary: The applicability of machine learning based reduced order model (ML-ROM) to three-dimensional complex flows is investigated. The combination of CNN-AE and LSTM in the current ML-ROM successfully reproduces turbulent flow fields and conducts statistical analysis.
Article
Mechanics
Francisco Alcantara-Avila et al.
Summary: A direct numerical simulation of turbulent heat transfer in a channel flow was conducted at a Reynolds number of and a Prandtl number of air, . The study obtained mean values and intensities of temperature, calculated parameters such as the von Karman constant and the Nusselt number, and proposed correlations. It was observed that an asymptotic behavior of the von Karman constant was present as the Reynolds number increased.
JOURNAL OF FLUID MECHANICS
(2021)
Article
Physics, Fluids & Plasmas
M. Buzzicotti et al.
Summary: The study investigates the use of computer vision tools for data reconstruction, specifically examining the application of convolutional neural networks and deep generative adversarial models in turbulent data. Two methods based on context encoders are proposed to improve the quality of reconstructed fields. A comparison with a different data assimilation tool, based on Nudging, is presented, offering new research directions in the study of turbulence.
PHYSICAL REVIEW FLUIDS
(2021)
Article
Mathematics
Federico Lluesma-Rodriguez et al.
Summary: This numerical method uses the vorticity-velocity gradient formulation and excludes pressure, successfully applied to analyze passive thermal flow in turbulent thermal channel. Multiple passive scalar quantities can be studied easily.
Article
Thermodynamics
Hamidreza Eivazi et al.
Summary: The study evaluated the capabilities of recurrent neural networks and Koopman-based frameworks in predicting the temporal dynamics of the low-order model of near-wall turbulence. The results showed that properly trained LSTM networks could achieve excellent reproductions of long-term statistics and dynamic behavior of the chaotic system, while the KNF framework offered the same level of accuracy in statistics at a lower computational cost.
INTERNATIONAL JOURNAL OF HEAT AND FLUID FLOW
(2021)
Article
Mechanics
A. Guemes et al.
Summary: This study evaluates the applicability of super-resolution generative adversarial networks (SRGANs) for reconstructing turbulent flow quantities, demonstrating that SRGAN can enhance the resolution of coarse wall measurements and provide better instant reconstructions. Despite challenges with lower resolution inputs, SRGAN yields very good reconstruction results, capturing large-scale flow structures effectively.
Article
Mechanics
Chao Jiang et al.
Summary: A new machine learning framework for turbulence modeling is proposed in this study, which achieves more accurate predictions and good generalization ability. The framework consists of two parallel modules for structural and parametric representations, combining data-driven and knowledge-driven approaches.
Review
Mechanics
Steven L. Brunton et al.
ANNUAL REVIEW OF FLUID MECHANICS, VOL 52
(2020)
Article
Mechanics
Bo Liu et al.
Article
Mechanics
Bing-Zheng Han et al.
Article
Computer Science, Theory & Methods
Samira Pouyanfar et al.
ACM COMPUTING SURVEYS
(2019)
Review
Mechanics
Karthik Duraisamy et al.
ANNUAL REVIEW OF FLUID MECHANICS, VOL 51
(2019)
Article
Mechanics
Kai Fukami et al.
JOURNAL OF FLUID MECHANICS
(2019)
Article
Physics, Fluids & Plasmas
P. A. Srinivasan et al.
PHYSICAL REVIEW FLUIDS
(2019)
Article
Mechanics
Sangseung Lee et al.
JOURNAL OF FLUID MECHANICS
(2019)
Review
Oncology
Khushboo Munir et al.
Article
Mechanics
Zhiwen Deng et al.
Article
Chemistry, Analytical
Xiaofeng Du et al.
Article
Mechanics
J. Nathan Kutz
JOURNAL OF FLUID MECHANICS
(2017)
Proceedings Paper
Computer Science, Artificial Intelligence
Christian Ledig et al.
30TH IEEE CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2017)
(2017)
Article
Engineering, Mechanical
N. J. Neeteson et al.
EXPERIMENTS IN FLUIDS
(2016)
Article
Engineering, Mechanical
Daniel Schanz et al.
EXPERIMENTS IN FLUIDS
(2016)
Article
Mechanics
Adrian Lozano-Duran et al.
JOURNAL OF FLUID MECHANICS
(2014)
Review
Engineering, Multidisciplinary
F. Scarano
MEASUREMENT SCIENCE AND TECHNOLOGY
(2013)
Article
Engineering, Electrical & Electronic
Li Deng et al.
FOUNDATIONS AND TRENDS IN SIGNAL PROCESSING
(2013)
Article
Computer Science, Hardware & Architecture
Michael Garland et al.
Article
Mechanics
S Hoyas et al.
Article
Engineering, Multidisciplinary
F Lekien et al.
INTERNATIONAL JOURNAL FOR NUMERICAL METHODS IN ENGINEERING
(2005)