Parameters Determination of grain Microstructure Prediction for a Single Crystal Casting Simulation and Its Experimental Validation

Abstract

Microstructure prediction is very important in single crystal casting simulation. In order to predict microstructure accurately, both grain nucleation and dendrite growth have to be modeled properly beside the correct setups on geometry, alloy chemistry, thermal/fluid, and processing conditions. For nucleation, the classic normal Gaussian distribution nucleation model was used. An extended KGT model was applied to predict the dendrite growth of multicomponent alloy castings. There are several parameters for both nucleation and growth models which are difficult to measure experimentally. The growth coefficients in the growth model are normally chemistry dependent. All parameters are critical to assure the reliability of the simulation. The current paper shows how to determine those parameters with the comprehensive knowledge of physics and its implementation behind the simulation model. It is intended to shed some light for the readers to better apply this model in their grain structure simulations, particularly on DC and single crystal castings, in the future by applying the presented methodology of the parameter determination in this paper. An even more important one is that a large amount of experiments have been performed. Excellent agreement has been reached statistically between experiment and simulation in terms of grain mis-orientation.