# Microstructural Inelastic Fingerprints And Data-Rich Predictions of Plasticity and Damage in Solids

**Authors:** Stefanos Papanikolaou  
**arXiv:** [1905.11289](https://arxiv.org/abs/1905.11289)  
**Date:** 2019-05-27  
**Summary page:** https://www.summarizepaper.com/en/paper/1905.11289v3/

## Key points

- The paper explores modeling of inelastic mechanical responses in solids
- Existing models often rely on linear elasticity to describe microstructural properties
- The authors propose a framework that uses microstructural images to develop fingerprints of inelastic tendencies
- These fingerprints can be used to make data-rich predictions of mechanical responses up to failure
- A deep convolutional neural network is used for image recognition and reconstruction of mechanical responses
- The effectiveness of the approach is demonstrated using a toy model of dislocation plasticity and phase field simulations of model binary alloys under fracture loading
- The framework has potential applications in engineering design and materials science research.

## Summary

The paper titled "Microstructural Inelastic Fingerprints And Data-Rich Predictions of Plasticity and Damage in Solids" by Stefanos Papanikolaou explores the modeling of inelastic mechanical responses in solids. These responses, including plasticity, damage, and crack initiation, are highly dependent on the microstructure and loading conditions. However, existing models often rely on linear elasticity at infinitesimal deformations to describe microstructural properties. To address this limitation, the authors propose a framework that utilizes sequences of microstructural images to develop fingerprints of inelastic tendencies. These fingerprints can then be used to make data-rich predictions of mechanical responses up to failure. Drawing an analogy to common fingerprints, the authors demonstrate that these two-dimensional instability-precursor signatures can be employed to reconstruct the complete mechanical response of unknown sample microstructures. The reconstruction process involves leveraging a deep convolutional neural network that has been fine-tuned for image recognition. By reconstructing appropriate average behaviors, the network assists in predicting the full mechanical response based on the identified fingerprints. The authors showcase the effectiveness of their approach by demonstrating basic aspects of microstructural fingerprinting using a toy model of dislocation plasticity. Furthermore, they illustrate the scalability and robustness of their method through phase field simulations of model binary alloys under mode-I fracture loading. This showcases how their framework can be applied to more complex systems and still yield accurate predictions. Overall, this paper presents a novel approach for capturing and utilizing microstructural fingerprints to predict inelastic tendencies and mechanical responses in solids. The proposed framework has potential applications in various fields where understanding material behavior is crucial, such as engineering design and materials science research.