AT-PINN-HC: A refined time-sequential method incorporating hard-constraint strategies for predicting structural behavior under dynamic loads

Computer Methods in Applied Mechanics and Engineering, Journal Year: 2025, Volume and Issue: 436, P. 117691 - 117691

Published: Jan. 10, 2025

Language: Английский

---

Solving parametric high-Reynolds-number wall-bounded turbulence around airfoils governed by Reynolds-averaged Navier–Stokes equations using time-stepping-oriented neural network

Physics of Fluids, Journal Year: 2025, Volume and Issue: 37(1)

Published: Jan. 1, 2025

Physics-informed neural networks (PINNs) have recently emerged as popular methods for solving forward and inverse problems governed by partial differential equations. However, PINNs still face significant challenges when high-Reynolds-number flows with multi-scale phenomena. In our previous work, we proposed time-stepping-oriented network (TSONN), which transforms the ill-conditioned optimization problem of into a series well-conditioned sub-problems, successfully three-dimensional laminar flow around wing at Reynolds number 5000. this paper, extend TSONN to wall-bounded turbulence airfoils Reynolds-Averaged Navier–Stokes (RANS) equations Spalart–Allmaras (SA) model. Specifically, propose semi-coupled strategy address convergence issues caused This updates certain terms in model only during outer iterations while freezing these inner iterations, thereby avoiding excessive gradients that could jeopardize optimization. Using strategy, solve airfoils. Furthermore, parametric respect angle attack. Our experimental results demonstrate computational cost using is comparable single problem, highlighting its efficiency problems. To best knowledge, first time PINN-like method has been used RANS coupled complex model, paving way fluid-related engineering

Language: Английский

---

An analysis and solution of ill-conditioning in physics-informed neural networks

Journal of Computational Physics, Journal Year: 2024, Volume and Issue: 520, P. 113494 - 113494

Published: Oct. 9, 2024

Language: Английский

---

Adopting Computational Fluid Dynamics Concepts for Physics-Informed Neural Networks

AIAA SCITECH 2022 Forum, Journal Year: 2025, Volume and Issue: unknown

Published: Jan. 3, 2025

Language: Английский

---

Parallel spatiotemporal order-reduced Gaussian process for dynamic full-field multi-physics prediction of hypervelocity collisions in real-time with limited data

Computer Methods in Applied Mechanics and Engineering, Journal Year: 2025, Volume and Issue: 438, P. 117810 - 117810

Published: Feb. 20, 2025

Language: Английский

---

A pseudo-time stepping and parameterized physics-informed neural network framework for Navier–Stokes equations

Physics of Fluids, Journal Year: 2025, Volume and Issue: 37(3)

Published: March 1, 2025

Physics-informed neural networks (PINNs) have emerged as a popular approach in scientific machine learning for solving both forward and inverse problems of partial differential equations (PDEs). However, complex physical systems are often characterized by parameters, such viscosity Reynolds number fluid dynamics, which pose significant challenges parameterized PDE solutions. The inherent limitations PINNs include the need repeated time-consuming training under varying parameter conditions, minimization residuals with PDE-based soft constraints, makes “ill-conditioned” problem. To address these issues, this paper proposes an innovative framework: pseudo-time stepping physics-informed network (P2PINN). P2PINN leverages explicit encoding only two parameters' latent representations to enable efficient interpolation extrapolation across wide range parameters. By integrating method deep learning, framework significantly alleviates ill-conditioned We validated our context Navier–Stokes equations, experimental results demonstrate that achieves solution speedups up 2–4 orders magnitude compared baseline their variants, while also surpassing them accuracy.

Language: Английский

---

Solving ultra-high-load, low-pressure turbine cascade flow using physics-informed neural networks with boundary identification strategy

Physics of Fluids, Journal Year: 2025, Volume and Issue: 37(4)

Published: April 1, 2025

This paper constructs a physics-informed neural network with boundary identification strategy (BI-PINN) to reconstruct the flow field of an ultra-high-load, low-pressure turbine cascade using sparse data. Since layer position is unknown before training, iterative method required determine range. The BI-PINN trains based on dual convergence criterion: achieving reconstruction in both and mainstream regions extract current range then ensuring To improve while maintaining high training speed, this study assigns different weights governing equation terms loss function for regions. demonstrates computational accuracy comparable fluid dynamics inverse problems incomplete conditions. finds that high-load, cascades, selecting pressure observations only blade surface sufficient high-accuracy reconstruction. However, proposes multiple normal-direction along achieve small observation set. balances conflict between low-gradient region high-gradient By leveraging physical equations data, it enables precise full cascades dataset limited number parameters.

Language: Английский

---

Solution and applications of parameterized convective heat transfer equations using physics-informed neural networks

International Journal of Heat and Mass Transfer, Journal Year: 2025, Volume and Issue: 247, P. 127040 - 127040

Published: April 26, 2025

Language: Английский

---

CF-DeepONet: Deep operator neural networks for solving compressible flows

Aerospace Science and Technology, Journal Year: 2025, Volume and Issue: unknown, P. 110329 - 110329

Published: May 1, 2025

Language: Английский

---