Molecular dynamic characteristic temperatures for predicting metallic glass forming ability

We explore the use of characteristic temperatures derived from molecular dynamics to predict aspects metallic Glass Forming Ability (GFA). Temperatures cooling curves self-diffusion, viscosity, and energy were used as features for machine learning models GFA. Multiple target model combinations with these explored. First, we logarithm critical casting thickness, log10(Dmax), trained regression on 21 compositions. Application 3-fold cross-validation log10(Dmax) alloys showed only weak correlation between predictions values. Second, GFA quantified by melt-spinning or suction amorphization behavior, that crystalline phases after synthesis classified Poor those pure amorphous Good Binary classification was then modeled using decision tree-based methods (random forest gradient boosting models) assessed nested-cross validation. The maximum F1 score precision–recall positive class 0.82±0.01 best type. also compared simple functions in place themselves found no statistically significant difference predictive abilities. Although ability developed here are modest, this work demonstrates clearly one can simulations metal glass forming ability.