Sequential Approximate Optimization Based Robust Design of SiC-Si3N4 Nanocomposite Microstructures

Abstract

microstructures of SiC-Si3N4 nanocomposites for desired high temperature toughness is presented. The focus is on finding robust nanocomposite microstructures with maximum toughness at two temperatures: 1500 oC and 1600 oC. Within this context a sequential approximate optimization algorithm under uncertainty is applied to six different test problems addressing different aspects of robust microstructure generation. During optimization, statistical uncertainties inherent to the computational microstructural generation are quantified and introduced in the optimization framework. The results show that the SiC volume fraction, the number of Si3N4 grains, the grain size distribution of the Si3N4 grains, and the grain size of the SiC grains have varied effects on the microstructure toughness at different temperatures. At 1500 oC, the preferred microstructure is the one with higher Si3N4 volume fraction. On the other hand, at 1600 oC, the preferred microstructure is the one with higher SiC volume fraction.