Development and Application of Phase-Field Models to Study the Solidification of Steels in a Weld Pool

The quality of welded products is largely determined by solidified structures and dendrite is the predominant pattern. However, studying the evolution of morphology during solidification at micrometre scale using experimental techniques is of high cost. Therefore, in this PhD project phase-field (PF) models were developed and applied to simulate the dendritic growth of industrially important steels in a weld pool. The problem of Ohno and Matsuuras proposal to include solid diffusivity in PF simulations is that a priori unknown term associated with the solute flux needs to be input. In this work, mathematical and numerical analyses indicate that credible results can be obtained by setting that term as zero. The effect of solid diffusivity on the steady-state dendrite under free growth was next investigated, showing that the solid diffusivity of C should be included for Fe-C alloys. The evolution of side branches under the imposed transient conditions by decreasing the velocity was examined because dendrites do not grow under steady state in a weld pool. It was observed that the remelting of the smaller side branches started from the side branch tips and the necklace width of the surviving side branches increased during the ripening process whether under steady state or transient conditions. However, the final necklace widths of the surviving side branches are finer under the transient conditions, implying that the detachment of side branches is more likely to occur under the transient condition, in agreement with the published experimental reports. Finally a new quantitative thin-interface PF solidification model for ternary alloys was extended and validated. The proposed model has the advantages: (1) the schemes of thin-interface limit analysis and anti-trapping current are adopted to simultaneously ensure the calculation efficiency and accuracy, and (2) the solid diffusivity can be included for steels.