Novel DeepONet architecture to predict stresses in elastoplastic structures with variable complex geometries and loads

Keyphrases

• Complex Geometry 100%
• Network Architecture 100%
• Complex Load 100%
• Deep Operator Network 100%
• Elastoplastic Structure 100%
• U-Net 50%
• Trunk 20%
• Fully Connected Network 20%
• Trunk Network 20%
• Finite Element Simulation 10%
• Order of Magnitude 10%
• Stress Distribution 10%
• Branched Network 10%
• Prediction Error 10%
• Stress Response 10%
• Highly Nonlinear 10%
• Stress Field 10%
• Design Sensitivity Analysis 10%
• Topology Optimization 10%
• Design Optimization 10%
• Uncertainty Quantification 10%
• Memory Efficiency 10%
• Comparable Accuracy 10%
• Prediction Accuracy 10%
• Stress Concentration 10%
• Concentration Point 10%
• Varying Geometries 10%
• Latent Space 10%
• Information Fusion 10%
• Nonlinear Analysis 10%
• Variable Geometry 10%
• Network Use 10%
• Variable Load 10%
• Fully Connected 10%
• Framework Architecture 10%
• Elastoplastic Material Behaviour 10%

Engineering

• Network Operator 100%
• Complex Variable 100%
• Finite Element Simulation 10%
• Stress Field 10%
• Prediction Error 10%
• Material Behavior 10%
• Uncertainty Quantification 10%
• Nonlinear Analysis 10%
• Design Sensitivity Analysis 10%
• Stress Concentration Point 10%