A three-dimensional computer model to simulate the sintering of a randomly packed green compact in order to investigate the evolution of the microstructure

The objective of this research project was to develop a three-dimensional computer model to simulate the sintering of a randomly packed green compact in order to investigate the evolution of the microstructure as densification progresses. The model randomly generates a population of spheres to simulate the particle size distributions present in commercial alumina powders. The spheres were packed into a cubic volume to construct a Representative Volume Element (RVE) with properties statistically comparable to the properties measured in real sintered materials. Sintering was modeled as uniform flattening of the contact between spheres. The sintering process was assumed to be isotropic and homogenous. The simulation of the sintering process was later expanded to include Local Particle Rearrangement (LPR). The model was compared to experimental data obtained from measurements on sintered alumina samples using two techniques: the stereological parameters for the RVE determined by the simulation, and additional measurable quantities obtained from applying tessellation to cross sections taken from the RVE to further describe the characteristics of the microstructure. Based on the measured stereological parameters and the properties of the tessellated cells, the simulation reproduced behavioral trends in the microstructure similar to those observed in the measurement of the experimental samples thus demonstrating the feasibility of employing this type of model to explore the microstructure. The results also indicate that the generation of agglomerated particle packing arrangements may be useful in simulating the microstructural evolution of sintering since the presence of the large voids used to simulate agglomeration had a significant effect on the results. The use of LPR in the simulation did not appear to affect the results for a particle arrangement with no agglomerations, but it did influence the results for the particle arrangement with agglomeration. The results for the LPR simulations suggest that the ability of the particles to slide past each other, modeled in the simulation by the tangential viscosity, may influence the rate at which small pores are formed and/or eliminated in the RVE. However, the LPR simulations indicate that more particles are needed to effectively model the influence of agglomeration in the particle arrangement.