A machine learning approach to predicting the spall strength of metals and alloys

Spall strength is a critical property that characterizes a material's resistance to dynamic failure under impact or shock compression loading. Accurate prediction and control of spall strength are essential for designing materials used in armors, vehicle components, structural barriers, and mining. Traditional methods for measuring the spall strength are slow, destructive, and expensive, while existing models offer limited predictive accuracy. This study introduces a machine learning regression model that rapidly and accurately predicts the spall strength of metals and alloys at strain rates in the range of 105–106 s−1. Trained on a dataset of over 70 metals and alloys and validated with recent data, the model identifies tensile yield strength and bulk modulus as key factors, with higher values of these properties correlating with increased spall strength. The findings offer an efficient method for screening metals and alloys and guiding the selection of high-spall-strength candidates for targeted experimental validation.