Abstract

ABSTRACT The solidification of metals and alloys is a typical example of multiphysics and multiscale engineering systems. The phenomenon of different time and spatial scales should be taken into consideration in the modeling of a microstructure formation: heat diffusion, the components diffusion in the liquid and solid phases, the thermodynamics of phase transformation under a condition of inhomogeneous chemical composition of growing and vanishing phases, phase interface kinetics, and grains nucleation. The results of a two-dimensional modeling of the microstructure formation in a ductile cast iron are presented. The cellular automaton model (CA) was used for the simulation. The model takes into account the nucleation of two kinds of grains that appear inside of the liquid during solidification: austenite and graphite. The six states of CA cells correspond to the above-mentioned three phases (liquid, austenite and graphite) and to the three two-phase interfaces. A numerical solution was used for the modeling of concentration and temperature fields. The parabolic nonlinear differential equations with a source function were solved by using the finite element method and explicit scheme. In the mono-phase cells the source function is equal to zero. In the interface cells the value of the source function varies depending on the local undercooling. The undercooling value depends on the front curvature, the local temperature and the local chemical composition of the phases. Overlapping lattices with the same spatial step were used for concentration field modeling and for the CA. The time scale of the temperature field for this lattice is about 104 times shorter. Due to the above reasons, another lattice was used with a multiple spatial step and the same time step. The new grain nucleation of solid phases from a liquid is a phenomenon which must be taken into account for correct simulation of a polycrystalline structure formation. The cumulative distribution curve approach was used to calculate the number of substrates on which nucleation takes place as a function of under-cooling below the equilibrium temperature. An algorithm of continuous nucleation modeling during solidification is presented. The undercooling of solid phase grain nucleation was calculated on the basis of the inverse function of the above-mentioned cumulative distribution curve (fractile) with the argument equal to the random number generated in the interval 0…1 with uniform density. The domain of correct usage of this algorithm was analyzed. INTRODUCTION The structure of alloys formed during solidification affects the utilization properties of materials and ready products, ascast, heat-treated, or fabricated by plastic working. This is the reason why in modern metallurgy and related technological