Implementation of nucleation in cellular automaton simulation of microstructural evolution during additive manufacturing of Al alloys

Abstract

Abstract In this work, a new approach is developed to integrate a particle-based (seed crystals) nucleation mechanism into the Cellular Automata (CA) simulation of powder-bed laser beam melting. The model aims to reflect two features observed in the microstructures of additively manufactured Al alloys: the fusion boundary nucleation which violates the principle of common bulk nucleation favored under high solidification rates and low thermal gradients, and the epitaxial growth at the melt-pool bottom as observed in the microstructure of AlSi10Mg alloy. Based on the proposed nucleation model, a CA cell will accommodate a nucleus after a few physics-based nomination and activation criteria have been passed. The fusion boundary nucleation can be reproduced by the model with a nomination criterion that restricts nucleation to the CA cells having a maximum experienced temperature below the particle dissolution temperature. However, the second observed feature needs an additional criterion to incorporate the nucleation incubation time. This is interpreted through the competition between the epitaxial growth and the nucleation which results in different behaviors at different melt-pool regions. Accordingly, the model offers high flexibility in calibration based on the growth and nucleation kinetics. The set of parameters used for AlSi10Mg are capable to deliver a similar microstructure and texture compared to the experiments. Moreover, a very high nucleation density along with a lower nucleation incubation time can lead to the bimodal grain structure observed in Sc-modified aluminum alloys.