Skip to content

AI4CHEMIA/Dimension-Agnostic-and-Interpretable-Modelling-of-Nuclear-Reactor-Steel-Toughness-with-Deep-Learning

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Predictive Framework for DBTT and USE Estimation in Structural Steels

This repository presents a comprehensive workflow for predicting ductile-to-brittle transition temperature (DBTT) and upper shelf energy (USE) in steels using an integrated attention-based deep learning framework. The approach combines physics-informed feature engineering, attention-enhanced multilayer perceptrons (MLP), and Evidential Deep Learning (EDL) to achieve accurate and uncertainty-aware predictions, facilitating reliable evaluation of structural integrity in critical applications such as pipelines, pressure vessels, and cryogenic systems.


✨ Overview

Dataset: 4,105 experimental samples comprising 167 descriptors, including Magpie-derived features, compositional variables, impact energy, and test temperature.

Descriptors: Physics-informed and statistically curated features embedding compositional, microstructural, and thermomechanical information.

Models:

  • Attention-Enhanced MLP: Learns nonlinear feature–property relationships with adaptive weighting through an attention mechanism.
  • Evidential Deep Learning (EDL): Extends the MLP to jointly predict property values and associated aleatoric and epistemic uncertainties.
  • Baselines: Ensemble regressors—Random Forest, Extra Trees, and XGBoost—for comparative evaluation.

Evaluation Metrics: MAE, RMSE, R², and Negative Log-Likelihood (NLL) to assess prediction accuracy and uncertainty calibration.


📂 Repository Structure


├── dataset/ # Dataset files
├── MLP/ # Scripts for attention-based MLP and EDL training
├── mlp_models/ # Saved MLP and EDL model checkpoints
├── Traditional ML/ # Random Forest, XGBoost, and Extra Trees training scripts
└── README.md


⚙️ Methodology Pipeline

1. Data Preprocessing

  • Curated dataset containing 4,105 samples with 167 descriptors.
  • Integration of compositional, impact energy, and temperature data.
  • Standardization and outlier filtering to ensure model robustness.

2. Feature Engineering

  • Physics-informed and Magpie-derived features capturing composition–structure–property relationships.
  • Redundancy reduction via correlation and variance filtering.
  • Domain-guided feature selection to preserve interpretability.

3. Modeling

  • Attention-Enhanced MLP:
    • Layers: Dense + Attention + Dropout
    • Optimizer: Adam with learning rate scheduling
    • Hyperparameter tuning using Optuna
  • Evidential Deep Learning (EDL):
    • Output: Normal–Inverse–Gamma (μ, ν, α, β) parameters
    • Captures both aleatoric and epistemic uncertainty
    • Penalized loss to balance accuracy and uncertainty calibration
  • Baselines:
    • Random Forest, Extra Trees, and XGBoost regressors trained on identical features for fair comparison.

4. Evaluation

  • Metrics: MAE, RMSE, R², and NLL.
  • EDL demonstrated improved uncertainty calibration and overall predictive reliability.

🧩 Key Contributions

  • Integration of attention mechanisms within MLP for selective feature relevance learning.
  • Application of Evidential Deep Learning for simultaneous prediction and uncertainty estimation.
  • Development of a physics-informed, data-driven framework for fracture behavior modeling.
  • Establishes a reliable and interpretable predictive pipeline for materials design and screening.

About

No description, website, or topics provided.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors