|About this Abstract
||2018 TMS Annual Meeting & Exhibition
||Coupling Advanced Characterization and Modeling Tools for Understanding Fundamental Phase Transformation Mechanisms: An MPMD Symposium in Honor of Hamish Fraser
||Probabilistic Methodology for Analyzing and Reconstructing Parent Microstructures from EBSD Maps of Transformation Products
||Stephen Niezgoda, Eric Payton, Alex Brust, Vikas Sinha
|On-Site Speaker (Planned)
The properties and performance of transformation microstructures are often dependent on features of the prior microstructure, such as texture or grain size, which have been obscured by the transformation. Reconstruction of the parent is typically mathematically ill-posed, as the forward transformation is often one-to-many; exhibiting multiple orientation variants or transformation pathways. Point-to-point reconstruction techniques, which rely on a pre-supposed orientation relationship, may not be robust to noise or deviations from ideal conditions. Here we present a probabilistic approach to quantify the uncertainty in the orientation relationship and noise due to measurement resolution, variation in parent orientation, and sample preparation. The reconstruction is formulated as a global optimization where the target is to minimize the probability of misindexed points or equivalently to find the parent microstructure which was most likely to generate the observed transformed dataset. The approach is material agnostic and will be demonstrated on steels and titanium alloys.
||Planned: Supplemental Proceedings volume