Data-Efficient Indentation Size Effect Mitigation in Steels Using Machine Learning and a Physics-Constrained Neural Network
The study developed a data-efficient workflow to predict high-load reference hardness from a single shallow indentation by using physics-guided feature engineering and a Physics-Constrained Neural Network (PCNN). The PCNN uses physical constraints to correct for the Indentation Size Effect, outperforming conventional models on unseen materials and providing stable predictions.