Concurrent model for sharp and progressive columnar to equiaxed transitions validated by directional solidification experiments processed in microgravity conditions

Abstract

Columnar and equiaxed structures, which occur during solidification of metallic alloys, influence the texture and properties of castings, welded joints and additively manufactured components. During transient solidification, where grain refiner particles provide the predominant nucleation mechanism, a Columnar to Equiaxed Transition (CET) occurs when conditions that had originally favoured directional columnar growth change to those favouring equiaxed. Constitutional undercooling ahead of the columnar front can permit equiaxed nucleation and growth. By carrying out experiments in microgravity conditions, liquid flows due to thermal and solutal buoyancy effects are suppressed. In these diffusion-controlled conditions, we have observed examples of both sharp (clear) and progressive (gradual) CET. The experimental outcomes, especially the observation of a progressive CET, has highlighted the need for a continuum model that allows for competitive columnar and equiaxed structure development; hence, the Concurrent Columnar to Equiaxed Transition (C2ET) model is proposed. The C2ET is thermally transient and relies on the well-known concept of extended growth for impingement mechanics; thereby, greatly reducing numerical complexity. Importantly, the proposed approach removes the need for a specific equiaxed-blocking criterion, which is often proposed as an essential requirement in other CET models. The C2ET model is validated by four experimental solidification scenarios: two velocity jumps and two thermal-gradient decreases. The velocity jumps induced sharp CETs; whereas, thermal-gradient decreases gave progressive CETs. The C2ET model gave good agreement for the columnar and equiaxed transition zones for both sharp and progressive CET. Results are compared with the classic Hunt model. Unlike Hunt’s model, the C2ET model predicted all macrostructure transitions faithfully using a single (or consistent) set of nucleation input parameters across all four scenarios. Since, the same level of grain refinement was used in each experiment, a consistent set of nucleation parameters was expected. The validated approach can enable effective simulation at lower computational cost for industrial processes that rely on a solidification processing step.