Quantifying Uncertainty in Materials Properties from Microstructure Variability

Most materials are inherently inhomogeneous. This means that their internal structure varies from point to point on the microscale; from region to region on the macroscale; from part to part; and throughout time as they age. Because a material's microstructure often controls its properties, the variability in structure leads to uncertainty about the material's properties and performance. We will discuss the concept of a statistics-based treatment of the process-structure-properties-performance relationships in engineering materials, and describe both experiments and computational simulations designed to quantify the statistics underlying structure-properties relationships in poly-silicon MEMS devices, stainless steel welds, and alpha-brass with engineered defects.