New neural operator FG-ConvNO predicts propeller CFD with limited data

Researchers have developed FG-ConvNO, a novel neural operator designed for predicting computational fluid dynamics (CFD) in complex geometries, particularly for propeller designs. This model addresses the challenge of limited training data by incorporating geometry-aware encoding and discrete continuous convolutions to achieve discretization agnosticism. Experiments show FG-ConvNO outperforms existing methods in predicting surface pressure and global thrust, demonstrating improved stability and generalization capabilities in low-data scenarios. AI

IMPACT This model could significantly reduce the computational cost and time required for engineering design simulations, accelerating development cycles.

RANK_REASON The cluster describes a new academic paper detailing a novel neural operator for a specific scientific modeling task. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.LG →

• arXiv
• computational fluid dynamics
• Factorized Geometry Convolutional Neural Operator
• FG-ConvNO
• Hugging Face

• paper
• infra

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →