Study of the motion of the grain boundaries ensemble in pure aluminum at high temperatures by cellular automata and machine learning methods

The motion of grain boundary (GB) ensemble in pure aluminum under annealing conditions at temperature 630 K is investigated by cellular automata (CA) method. Here, the CA method predetermines the computational grid following the grain boundary motion and grain growth. The motion of the GB is defined through mobility, energy and curvature of the boundary. The curvature of the GBs is measured by the Height Function method. The kinetics of the grain boundary ensemble is shown using the conventional Reed — Shockley function and neural network surrogate functions to determine the energy of grain boundaries.