Investigation of Full-field and Mean-field Models for Pure Grain Growth Simulations

Abstract

Mean-field approaches are commonly used in the simulation of grain growth in metals as they are easier to implement. However, mean-field models only track the evolution of average grain diameter as a function of temperature and time, while they neglect the effect of actual grain size distribution, which limits their applicability. Recently introduced full-field models of grain growth – either by level-set and phase field methods- allow to overcome the barriers of mean-field methods, but they are computationally much more demanding. In this article, the main goal is to investigate the applicability of both approaches for the pure grain growth in a solid with an initial Gaussian grain size distribution. The main idea is to keep the average grain size constant while altering the grain size distribution by modifying the standard deviation. DIGIMU software, which uses the level-set approach - is used for this purpose. The conclusion is that full-field models are beneficial to observe changes during grain growth; alternatively, mean-field models deduce approximately the same results as full-field models for a Gaussian distribution within a shorter time. However, it is found that mean-field models overlook certain important stages of the evolution of the microstructure, while full-field method captures all the details. Therefore, the model to investigate the grain growth mechanism should be selected accordingly.