Predicting high-fidelity multiphysics data from low-fidelity fluid flow and transport solvers using physics-informed neural networks

Engineering & Materials Science

• Flow of fluids 92%
• Physics 90%
• Computational fluid dynamics 65%
• Neural networks 63%
• Transfer learning 20%
• Partial differential equations 17%
• Porous materials 17%
• Computational efficiency 16%
• Heat transfer 11%
• Temperature 6%

Physics & Astronomy

• fluid flow 100%
• physics 65%
• computational fluid dynamics 33%
• learning 6%
• education 5%
• partial differential equations 5%
• heat transfer 4%
• approximation 3%
• temperature 1%

Chemical Compounds

• Heat Transfer 22%
• Velocity 17%