Engineering Papers⌕ Search

Engineering topics

Stan, Tiberiu

Publications and source records attributed to Stan, Tiberiu.

Segmentation of tomography datasets using 3D convolutional neural networks

Dendritic microstructures are ubiquitous in nature and are the primary solidification morphologies in metallic materials. Techniques such as X-ray computed tomography (XCT) have provided new insights into dendritic phase transformation phenomena. However, manual identification of dendritic morphologies in microscopy data can be both labor intensive and potentially ambiguous. The analysis of 3D datasets is particularly challenging due to their large sizes (terabytes) and the presence of artifacts scattered within the imaged volumes. Here, in this study, we trained 3D convolutional neural networks (CNNs) to segment 3D datasets. Three CNN architectures were investigated, including a new version of FCDenseNet which we extended to 3D. We show that using hyperparameter optimization (HPO) and fine-tuning techniques, both 2D and 3D CNN architectures outperform the previous state of the art. The 3D U-Net architecture trained in this study produced the best segmentations according to quantitative metrics (intersection-over-union of 95.56% and a boundary displacement error of 0.58 pixels), while 3D FCDense produced the smoothest boundaries and best segmentations according to visual inspection. The trained 3D CNNs are able to segment entire 852 × 852 × 250 voxel 3D volumes in only ~60 s, thus hastening the progress towards a deeper understanding of phase transformation phenomena such as dendritic solidification.

36 MATERIALS SCIENCE↗

Microstructural characterization of dendritic evolution using two-point statistics

Characterizing large four-dimensional materials datasets is difficult due to the presence of complex mi- crostructures and time-varying length scales. We showcase the use of two-point statistics as an efficient and un-biased way of extracting materials parameters from an Al-Cu alloy during solidification. The evo- lution of dendrite primary arm thickness, average secondary arm spacing, and average tip-to-tip spacing were tracked using two-point Pearson auto-correlations of scaled mean curvatures. Insights into compet- itive side-branching are also reported. We show both visually and quantitatively that most length scales change rapidly during early stages of dendritic growth, but slow as diffusion fields of dendrites overlap.

Aluminum alloys↗

X-ray absorption spectroscopy characterization of embedded and extracted nano-oxides

Here, the chemistries and structures of both embedded and extracted Ysingle bondTisingle bondO nanometer-scale oxides in a nanostructured ferritic alloy (NFA) were probed by x-ray absorption spectroscopy (XAS). Y 2 Ti 2 O 7 is the primary embedded phase, while the slightly larger extracted oxides are primarily Y 2 TiO 5 . Analysis of the embedded nano-oxides is difficult partly due to the multiple Ti environments associated with different oxides and those still residing in matrix lattice sites. Thus, bulk extraction followed by selective filtration was used to isolate the larger Y 2 TiO 5 oxides for XAS, while the smaller predominant embedded phase Y 2 Ti 2 O 7 oxides passed through the filters and were analyzed using the log-ratio method.

73 NUCLEAR PHYSICS AND RADIATION PHYSICS↗