Aug 2026· Journal of Physics, Conference Series· Vol 3304· 0 citations· 6 references
Physics
TL;DR
This study proposes a model that combines signed distance function (SDF) domain optimization, near-wall adaptive weight, and a U-Net neural network that increases the prediction accuracy of the pressure field and temperature field by 56% and 50%, respectively.
Abstract
Hypersonic vehicles have attracted widespread attention in the fields of national defense and aerospace transportation; therefore, predicting their complex flow fields quickly and accurately is crucial for aerodynamic design and thermal protection analysis. The computational fluid dynamics (CFD) method has the characteristics of high computational cost and long simulation time, which makes it difficult to meet the requirements of real-time flow field prediction. Deep learning has become the main method for real-time prediction of hypersonic flow fields. However, the prediction accuracy of existing models in shocks and near-wall regions remains limited. In this study, we propose a model that combines signed distance function (SDF) domain optimization, near-wall adaptive weight, and a U-Net neural network. Compared with the traditional U-Net model, the proposed model increases the prediction accuracy of the pressure field and temperature field by 56% and 50%, respectively.
The results show that ParaAero-Net performs effectively in small-sample aerodynamic prediction of two-dimensional airfoils and achieves good reconstruction of both flow-field structures and aerodynamic coefficients.
Jun-Qiao Wang, Chen Liu, Xu Zhan et al.· Proceedings of the Instituti...· 0 citations
Accurate velocity prediction of bodies under hypersonic flow is crucial for aerospace engineering and safety evaluation. However, unsteady aerodynamic effects coupled with sparse and noisy experimental data make accurate velocity prediction challenging. Taking conical-end cylinder as example, a prior-guided data fusion...
Aiming at the engineering pain points of insufficient sample size, inadequate working condition coverage, and poor generalization performance in traditional solution methods of vertical hydrodynamic coefficients for Autonomous Underwater Vehicles (AUVs), this paper constructs a high-precision prediction framework for v...
Ding-Feng Yu, Xi Zhang, Xia Yang et al.· Mathematics· 0 citations
Accurate and efficient prediction of airfoil flow fields remains computationally expensive with computational fluid dynamics(CFD) solvers. This paper presents a physics-informed conditional diffusion method that generates flow fields under different inflow conditions. Physical constraints derived from the continuity an...
Shang-Yuan Li, Guo-Cheng Tao, Zhou-Teng Ye et al.· Journal of Physics, Conferen...· 0 citations
Carrier motion under complex sea states continuously changes the relative inflow in the landing glide-path region and produces an unsteady airwake. High-fidelity computational fluid dynamics (CFD) can resolve this coupled wind–wave–carrier problem, but long-sequence and multi-condition calculations remain expensive. He...
Hong-Ze Sun, Yi-Ming Du, Bo-Jia Zeng et al.· The Physics of Fluids· 0 citations
This study investigates spatiotemporal flow-field reconstruction for the unsteady pitching dynamic stall of a NACA0012 airfoil using a sequence-based physics-informed neural network (PINN). During dynamic stall, the formation and shedding of the leading-edge vortex, wake transport, and aerodynamic hysteresis all exhibi...