Simulation of inclined dendrites under natural convection by KKS phase field model based on CUDA

In this work, Al-4.5wt.%Cu was selected as the research object, and a phase field-lattice Boltzmann method (PF-LBM) model based on compute unified device architecture (CUDA) was established to solve the problem of low serial computing efficiency of a traditional CPU and achieve significant acceleration. This model was used to explore the evolution of dendrite growth under natural convection. Through the study of the tip velocities, it is found that the growth of the dendrite arms at the bottom is inhibited while the growth of the dendrite arms at the top is promoted by natural convection. In addition, research on the inclined dendrite under natural convection was conducted. It is observed that there is a deviation between the actual growth direction and the preferred angle of the inclined dendrite. With the increase of the preferred angle of the seed, the difference between the actual growth direction and the initial preferred angle of the inclined dendrite shows a trend of increasing at first and then decreasing. In the simulation area, the relative deflection directions of the primary dendrite arms in the top right corner and the bottom left corner of the same dendrite are almost counterclockwise, while the relative deflection directions of the other two primary dendrite arms are clockwise.