←Back to feed
🧠 AI🟢 Bullish
Data-Free PINNs for Compressible Flows: Mitigating Spectral Bias and Gradient Pathologies via Mach-Guided Scaling and Hybrid Convolutions
🤖AI Summary
Researchers developed a data-free Physics-Informed Neural Network (PINN) that can solve compressible flows around circular cylinders at extreme speeds up to Mach 15. The system uses hybrid convolutions and Mach-guided scaling to overcome traditional limitations and successfully captures shock waves without requiring training data.
Key Takeaways
- →Novel PINN architecture combines radial 1D and azimuthal 2D convolutions to embed directional biases for fluid dynamics modeling.
- →Mach-number-guided dynamic scaling strategy enables stable optimization across supersonic to hypersonic flow regimes.
- →System successfully captures detached bow shock waves without any reference training data.
- →Upstream Fixing boundary loss and Total Variation loss suppress non-physical phenomena in extreme aerodynamics.
- →Framework demonstrates unprecedented stability for data-free neural networks in computational fluid dynamics applications.
#physics-informed-neural-networks#computational-fluid-dynamics#machine-learning#aerodynamics#shock-waves#data-free#deep-learning#scientific-computing
Read Original →via arXiv – CS AI
Act on this with AI
Stay ahead of the market.
Connect your wallet to an AI agent. It reads balances, proposes swaps and bridges across 15 chains — you keep full control of your keys.
Related Articles