Filtern
Dokumenttyp
Sprache
- Englisch (5)
Schlagworte
- 3D imaging (5) (entfernen)
Organisationseinheit der BAM
We propose a novel iterative segmentation algorithm (i.S.Sy.Da.T.A: Iterative Segmentation Synthetic Data Training Algorithm) employing Deep Convolutional Neural Networks and synthetic training data for X-ray tomographic reconstructions of complex microstructures. In our method, we reinforce the synthetic training data with experimental XCT datasets that were automatically segmented in the previous iteration. This strategy produces better segmentations in successive iterations. We test our algorithm with experimental XCT re constructions of a 6-phase Al-Si Matrix Composite reinforced with ceramic fibers and particles. We perform the analysis in 3D with a special network architecture that demonstrates good generalization with synthetic training data. We show that our iterative algorithm returns better segmentations compared to the standard single training approach. More specifically, phases possessing similar attenuation coefficients can be better segmented: for Al2O3 fibers, SiC particles, and Intermetallics, we see an increase of the Dice score with respect to the classic approach: from 0.49 to 0.54, from 0.66 to 0.72, and from 0.55 to 0.66 respectively. Furthermore, the overall Dice score increases from 0.77 to 0.79. The methods presented in this work are also applicable to other materials and imaging techniques.
Understanding the complicated interplay of the continuously evolving electrode materials in their inherent 3D states during the battery operating condition is of great importance for advancing rechargeable battery research.
In this regard, the synchrotron X-ray tomography technique, which enables non-destructive, multi-scale, and 3D imaging of a variety of electrode components before/during/after battery operation, becomes an essential tool to deepen this understanding. The past few years have witnessed an increasingly growing interest in applying this technique in battery research. Hence, it is time to not only summarize the already obtained battery-related Knowledge by using this technique, but also to present a fundamental elucidation of this technique to boost future studies in battery research. To this end, this review firstly introduces the fundamental principles and experimental Setups of the synchrotron X-ray tomography technique. After that, a user guide to ist application in battery research and examples of its applications in Research of various types of batteries are presented. The current review ends with a discussion of the future opportunities of this technique for next-generation rechargeable batteries research. It is expected that this review can enhance the reader’s understanding of the synchrotron X-ray tomography technique and stimulate new ideas and opportunities in battery research.
Automatic segmentation of cracks using crack detection methods fails when applied on concrete with embedded reinforcements; it results in artifacts called false-detected cracks. We present a two-step approach to trace the cracks which reduces significantly detecting the false ones. In the first step, the reinforcements were found and removed from the 3D data set, and in the second step, the cracks were detected in the 3D data set free from reinforcements. To visualize and determine the structure and the degree of damages nondestructively, planar tomography as a special case of laminography was used.
We examine the behavior of reinforced concrete components subjected to impact induced loading conditions which might be caused by vessels collisions such as aircraft fuel tanks The concrete plates were impact damaged at TU Dresden and shipped to BAM At BAM laminar tomography as the imaging method is used to determine and quantify the damage state An automatic crack detection method based on template matching is applied to find the cracks and we aim to develop a new method using machine learning Algorithms In addition numerical models are developed to understand the experiment and to predict the damage structures based on failure mechanisms.
Within the perspective of increasing reliability of AM processes, real-time monitoring allows part inspection while it is built and simultaneous defect detection. Further developments of real-time monitoring can also bring to self-regulating process controls. Key points to reach such a goal are the extensive research and knowledge of correlations between sensor signals and their causes in the process.