Filtern
Dokumenttyp
- Zeitschriftenartikel (2) (entfernen)
Sprache
- Englisch (2)
Referierte Publikation
- ja (2)
Schlagworte
- Metal matrix composite (MMC) (2) (entfernen)
Organisationseinheit der BAM
Microstructural aspects of composites such as reinforcement particle size, shape, and distribution play important roles in deformation behavior. In addition, Fe-rich inclusions and porosity also influence the behavior of these composites, particularly under fatigue loading. Three-dimensional (3-D) visualization of porosity and Fe-rich inclusions in three dimensions is critical to a thorough understanding of fatigue resistance of metal matrix composites (MMCs), because cracks often initiate at these defects. In this article, we have used X-ray synchrotron tomography to visualize and quantify the morphology and size distribution of pores and Fe-rich inclusions in a SiC particle-reinforced 2080 Al alloy composite. The 3-D data sets were also used to predict and understand the influence of defects on the deformation behavior by 3-D finite element modeling.
The greatest challenge when using deep convolutional neural networks (DCNNs) for automatic segmentation of microstructural X-ray computed tomography (XCT) data is the acquisition of sufficient and relevant data to train the working network. Traditionally, these have been attained by manually annotating a few slices for 2D DCNNs. However, complex multiphase microstructures would presumably be better segmented with 3D networks. However, manual segmentation labeling for 3D problems is prohibitive. In this work, we introduce a method for generating synthetic XCT data for a challenging six-phase Al–Si alloy composite reinforced with ceramic fibers and particles. Moreover, we propose certain data augmentations (brightness, contrast, noise, and blur), a special in-house designed deep convolutional neural network (Triple UNet), and a multi-view forwarding strategy to promote generalized learning from synthetic data and therefore achieve successful segmentations. We obtain an overall Dice score of 0.77. Lastly, we prove the detrimental effects of artifacts in the XCT data on achieving accurate segmentations when synthetic data are employed for training the DCNNs. The methods presented in this work are applicable to other materials and imaging techniques as well. Successful segmentation coupled with neural networks trained with synthetic data will accelerate scientific output.