Browse Teach Curate

Parametric Neural-Network Surrogate Solver for Laminar Airfoil Flows

A physics-informed solver that computes all laminar flows around airfoils at once by combining the Time-Stepping-Oriented Neural Network reformulation with a body-fitted mesh transformation. Trains a single label-free surrogate model spanning a high-dimensional space of airfoil shapes, Reynolds numbers, and angles of attack, overcoming the ill-conditioning that limits standard physics-informed neural networks on viscous flows.

Data Text

Solving all laminar flows around airfoils all-at-once using a parametric neural network solver

Physics-informed neural networks (PINNs) solve the partial differential equations of fluid flow by training a network to minimize the PDE, boundary, and initial-condition residuals, needing no labele…