How esanghaneh/multi-fidelity-data-fusion-for-in-elastic-woven-composites works
This research repository provides a notebook and datasets for surrogate modeling of woven-composite homogenization. The documented workflow combines low-fidelity mean-field data with high-fidelity FFT simulations, uses GRUs to model path-dependent behavior, and applies transfer learning, interpolation-based augmentation, and training/evaluation across loading scenarios. No notebook source excerpt was supplied, so internal implementation details and dataset-to-notebook code paths cannot be verified; the graph shows only README-documented relationships.
Main entry points: RNN_paper_GRu_transferLearning_CST.ipynb Multi-fidelity fusion · 27355_Random_MF_Plastic_Simulations_Woven_original_values_only_mat_scaled.zip Mean-field data · 3rd_WovenFullField_FFT_cyclic_shear12_only_mat_scaled.mat FFT shear 12 · 3rd_WovenFullField_FFT_cyclic_shear23_only_mat_scaled.mat FFT shear 23
Components 12
Modeling workflow
- GRU workflow
RNN_paper_GRu_transferLearning_CST.ipynb - Train and evaluate
RNN_paper_GRu_transferLearning_CST.ipynb
Simulation data
- Mean-field data
27355_Random_MF_Plastic_Simulations_Woven_original_values_only_mat_scaled.zip - FFT shear 12
3rd_WovenFullField_FFT_cyclic_shear12_only_mat_scaled.mat - FFT shear 23
3rd_WovenFullField_FFT_cyclic_shear23_only_mat_scaled.mat - FFT material data
10e3_wovenFFT_only_materials_scaled.mat
Learning capabilities
- Multi-fidelity fusion
RNN_paper_GRu_transferLearning_CST.ipynb - Temporal GRU model
RNN_paper_GRu_transferLearning_CST.ipynb - Transfer learning
RNN_paper_GRu_transferLearning_CST.ipynb - Interpolation augmentation
RNN_paper_GRu_transferLearning_CST.ipynb - Loading scenarios
RNN_paper_GRu_transferLearning_CST.ipynb
Other components
- Researcher
Connections 11
- Researcher → GRU workflow: runs workflow
README.md - Mean-field data → Multi-fidelity fusion: low-fidelity data
README.md - FFT shear 12 → Multi-fidelity fusion: high-fidelity data
README.md - FFT shear 23 → Multi-fidelity fusion: high-fidelity data
README.md - FFT material data → Multi-fidelity fusion: high-fidelity data
README.md - Multi-fidelity fusion → Temporal GRU model: feeds model
README.md - Temporal GRU model → Transfer learning: supports adaptation
README.md - Interpolation augmentation → Train and evaluate: augments data
README.md - Loading scenarios → Train and evaluate: sets scenarios
README.md - GRU workflow → Train and evaluate: trains and evaluates
README.md - Transfer learning → Train and evaluate: improves prediction
README.md