A hybrid physics-informed graph neural network for tornado wind-field modeling

Keyphrases

• Advanced Engineering 33%
• Analytical Model 100%
• Baseline Test 33%
• Computational Efficiency 66%
• Critical Infrastructure 33%
• Engineering Analysis 33%
• Engineering Application 33%
• Engineering Design 33%
• Flow Structure 33%
• Global Test 33%
• Graph Neural Network 100%
• Graph Neural Network Model 33%
• Hybrid Framework 33%
• Hybrid Graph Neural Network 33%
• Large Eddy Simulation 100%
• Mean Flow 33%
• Neural Network Architecture 33%
• Nonlinear Residuals 33%
• Novel Hybrids 33%
• Penalty Function 33%
• Physical Approximation 33%
• Physical Consistency 33%
• Physics-based 66%
• Physics-informed 100%
• Physics-informed Deep Learning 33%
• Prediction Accuracy 33%
• Root Mean Square Error 66%
• Simulation Data 33%
• Spatial Dependence 33%
• Tornado 33%
• Tornado Wind Field 100%
• Turbulent Characteristics 33%
• Velocity Field 33%
• Vortex 66%
• Vortex Field 33%
• Wind Field Modelling 100%
• Wind-resistant Design 33%

Engineering

• Accurate Prediction 33%
• Analytical Model 100%
• Critical Infrastructure 33%
• Deep Learning Method 33%
• Engineering Analysis 33%
• Engineering Application 33%
• Flow Structure 33%
• Interdependency 33%
• Large Eddy Simulation 100%
• Network Model 33%
• Neural Network Architecture 33%
• Numerical Analysis 33%
• Numerical Study 33%
• Penalty Function 33%
• Resistant Design 33%
• Root Mean Square Error 66%
• Velocity Field 33%