Data-Efficient Indentation Size Effect Correction in Steels Using Machine Learning and Physics-Guided Augmentation

2026-04-30 · Source: Takara TLDR - Daily AI Papers · Field: Science & Research — Engineering & Applied Sciences, Artificial Intelligence & Machine Learning · Depth: Expert, medium

Summary

This study introduces a machine learning workflow for correcting the indentation size effect (ISE) in steels, a common issue in shallow nanoindentation measurements that inflates hardness readings. Researchers collected approximately 700 experimental indentations from three steel specimens with hardness ranging from 2-6.5 GPa. This dataset was augmented with physically motivated variations simulating instrumental noise, session drift, and multiphase boundary blending. The input features included Oliver-Pharr values and mechanics descriptors like indentation work partitioning ($H/E_{r}$) and the area-invariant compliance proxy ($P_{\max}/S^{2}$). Various models, including Ridge Regression, Random Forest, XGBoost, and Neural Networks, were evaluated on a quarantined fourth steel specimen. Nonlinear models achieved high internal accuracy ($R^2 > 0.98$), with a constrained Neural Network (64-8-64) performing best, yielding an RMSE of 0.470 GPa and MAPE of 5.4% on the quarantined steel. This method provides stable hardness estimates in the shallow regime, unlike traditional Nix-Gao analysis.

Key takeaway

For materials scientists and engineers performing nanoindentation, this research demonstrates a robust machine learning approach to overcome the limitations of traditional ISE correction methods, especially in shallow regimes. You should consider integrating physics-guided data augmentation and advanced feature engineering into your characterization workflows to achieve more reliable hardness measurements from limited experimental data. This can significantly improve the accuracy of mechanical property assessment for thin films and volume-constrained materials.

Key insights

Physics-guided data augmentation enables accurate ISE correction in steels with small experimental datasets.

Principles

• Nonlinear models are crucial for hardness mapping.
• Area-invariant descriptors improve model reliability.

Method

The method involves collecting shallow nanoindentation data, augmenting it with physics-guided variations, engineering features from Oliver-Pharr values and mechanics descriptors, and training a constrained Neural Network for ISE correction.

In practice

• Apply physics-guided augmentation for small datasets.
• Use $H/E_{r}$ and $P_{\max}/S^{2}$ as key features.
• Favor Neural Networks for nonlinear material responses.

Topics

• Indentation Size Effect
• Nanoindentation
• Machine Learning
• Data Augmentation
• Steels Characterization

Best for: Research Scientist, AI Scientist, Machine Learning Engineer

Related on AIssential

• Machine learning enables roughness-driven inverse design of milling processes — An ML-based framework integrates DNN/RF models with Bayesian optimization for inverse design of milling processes, optimizing surface roughness with <5% error.
• Complex Layout Classification in the Wild: A Low-Resource Approach with Layout-Preserving Augmentations — Domain-aware augmentations significantly improve low-resource complex layout classification by focusing on global geometric arrangements.
• Leveraging systems' non-linearity to tackle the scarcity of data in the design of Intelligent Fault Diagnosis Systems — Leveraging system non-linearity and novel visualization creates data for DTL-based Intelligent Fault Diagnosis Systems despite scarcity.
• A Large-Scale Study on the Accuracy vs Cost Trade-offs of Training and Evaluation Settings in Fine-Grained Image Recognition — Data-aware augmentations during training can significantly reduce FGIR inference costs while maintaining high accuracy.
• Physics-Guided Dimension Reduction for Simulation-Free Operator Learning of Stiff Differential--Algebraic Systems — An extended Newton implicit layer enables simulation-free, accurate operator learning for stiff DAEs by enforcing constraints and reducing dimensions.
• Realistic noise synthesis reduces bias and improves tissue microstructure estimation with supervised machine learning — Realistic noise synthesis in simulated data is crucial for unbiased dMRI microstructure estimation, especially at low SNR.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Takara TLDR - Daily AI Papers.