Advancing super-resolution of turbulent velocity fields: An artificial intelligence approach

Physics of Fluids, Год журнала: 2025, Номер 37(3)

Опубликована: Март 1, 2025

This study presents a novel deep learning framework aimed at achieving super-resolution of velocity fields within turbulent channel flows across various wall-normal positions. The model excels reconstructing high-resolution flow from low-resolution data, with an emphasis on accurately capturing spatial structures and spectral energy distributions. Input data are generated through fine-grid large eddy simulations, employing data-driven approach. model's efficacy is evaluated using standard image quality metrics, including peak signal-to-noise ratio, structural similarity index measure, root mean square error, absolute good pixel percentage, as well analyses to encapsulate the complex dynamics physics. findings demonstrate substantial correlations between performance location. Specifically, performs superior in regions distal wall but faces challenges recovering small-scale near boundary layer.

Язык: Английский

---

A review of deep learning for super-resolution in fluid flows

Physics of Fluids, Год журнала: 2025, Номер 37(4)

Опубликована: Апрель 1, 2025

Integrating deep learning with fluid dynamics presents a promising path for advancing the comprehension of complex flow phenomena within both theoretical and practical engineering domains. Despite this potential, considerable challenges persist, particularly regarding calibration training models. This paper conducts an extensive review analysis recent developments in architectures that aim to enhance accuracy data interpretation. It investigates various applications, architectural designs, performance evaluation metrics. The covers several models, including convolutional neural networks, generative adversarial physics-informed transformer diffusion reinforcement frameworks, emphasizing components improving reconstruction capabilities. Standard metrics are employed rigorously evaluate models' reliability efficacy producing high-performance results applicable across spatiotemporal data. findings emphasize essential role representing flows address ongoing related systems' high degrees freedom, precision demands, resilience error.

Язык: Английский

---

Real-time inference and extrapolation with Time-Conditioned UNet: Applications in hypersonic flows, incompressible flows, and global temperature forecasting

Computer Methods in Applied Mechanics and Engineering, Год журнала: 2025, Номер 441, С. 117982 - 117982

Опубликована: Апрель 15, 2025

Язык: Английский

---

Refining Flow Structures with Deep Learning and Super Resolution Methods

Опубликована: Сен. 11, 2024

Язык: Английский

---