Uncertainty-Aware Machine-Learning Framework for Predicting Dislocation Plasticity and Stress-Strain Response in FCC Alloys

2025-06-25

AI Summary

Accurate prediction of dislocation-mediated plasticity and stress–strain responses in face-centered cubic (FCC) alloys remains challenging due to low predictive fidelity and difficulty in quantifying epistemic and aleatoric uncertainties. Method: This study proposes a physics-informed machine learning framework integrating latent-variable modeling and uncertainty quantification. It introduces mixture density networks (MDNs) for the first time to jointly predict the probabilistic distributions of dislocation density and local stress at the grain scale, embedding statistical parameters directly into a dislocation-based constitutive model. The framework is trained and validated on multi-source experimental data from the literature. Contribution/Results: The approach achieves high-accuracy, physically consistent predictions of FCC alloy hardening behavior while explicitly quantifying prediction uncertainty. It establishes a new computational paradigm for accelerated alloy design—offering enhanced accuracy, robustness, and interpretability without sacrificing physical fidelity.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Reasoning under Uncertainty: Relational Probabilistic Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Applied ML and AI for Web-based mobile applications
• Semantics and Knowledge: Methods, algorithms and applications for the development of semantic models, knowledge graphs and other forms of structured data models with machine-interpretable semantics

Abstract

Machine learning has significantly advanced the understanding and application of structural materials, with an increasing emphasis on integrating existing data and quantifying uncertainties in predictive modeling. This study presents a comprehensive methodology utilizing a mixed density network (MDN) model, trained on extensive experimental data from literature. This approach uniquely predicts the distribution of dislocation density, inferred as a latent variable, and the resulting stress distribution at the grain level. The incorporation of statistical parameters of those predicted distributions into a dislocation-mediated plasticity model allows for accurate stress-strain predictions with explicit uncertainty quantification. This strategy not only improves the accuracy and reliability of mechanical property predictions but also plays a vital role in optimizing alloy design, thereby facilitating the development of new materials in a rapidly evolving industry.

Problem

• Predict dislocation plasticity in FCC alloys with uncertainty awareness
• Model stress-strain response using grain-level dislocation density distributions
• Enhance alloy design accuracy via uncertainty-quantified machine learning

Innovation

• Uses mixed density network for dislocation prediction
• Quantifies uncertainty in stress-strain modeling
• Integrates statistical parameters into plasticity model