AI and the World's Hard Problems · Chemistry & materials · arXiv · 2026-10-20
Data-Efficient Indentation Size Effect Mitigation in Steels Using Machine Learning and a Physics-Constrained Neural Network
Shallow nanoindentation overestimates hardness because of the indentation size effect, which is a problem for thin layers and small volumes. The authors trained seven models on indentations from three steels to predict a high-load reference hardness from a single shallow indentation. On an unseen steel, ridge regression and physics-constrained neural networks reached 3.7–4.8% MAPE, while Random Forest and XGBoost transferred worst, at 10.2% and 8.5%.