Developing automated characterization techniques to quantify 3D datasets for ceramic matrix composite materials

被引:0
作者
Hilmas, Ashley M. [1 ]
Przybyla, Craig [1 ]
Schey, Mathew [1 ,2 ]
机构
[1] Air Force Res Lab, Mat & Mfg Directorate, Dayton, OH 45433 USA
[2] BlueHalo Co, UES, Dayton, OH 45432 USA
基金
英国科研创新办公室;
关键词
X-ray tomography; Microstructure; Composite; Machine learning; SYNCHROTRON X-RAY; FIBER; MICROSTRUCTURE;
D O I
10.1557/s43579-024-00637-1
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
The ability to obtain 3D microstructural information from techniques such as x-ray computed tomography (XCT) has become an imperative means for understanding the relationship between processing, structure, and performance of materials. Due to the large datasets that are a product of XCT, the development of automated characterization techniques are required to efficiently and accurately quantify microstructural features. Segmentation techniques such as thresholding are common for simplistic datasets, while machine learning tools such as deep learning used more complex datasets. For example, composites, specifically ceramic matrix composites, are often difficult to analyze due to the low contrast between constituents. Deep learning has proven to be a promising tool for segmenting multi-phase microstructures. For composite materials, fiber classification and fiber tracking are also growing areas of interest to understand how fiber variation affects the local microstructure. Faster R-CNNs have been developed for feature identification and have been proven to be able to identify features, such as fibers in extremely noisy XCT datasets. Fiber tracking is performed using association algorithms and Kalman filtering to provide linkages between composite structure and performance.
引用
收藏
页码:876 / 887
页数:12
相关论文
共 44 条
  • [1] Least-squares orthogonal distances fitting of circle, sphere, ellipse, hyperbola, and parabola
    Ahn, SJ
    Rauh, W
    Warnecke, HJ
    [J]. PATTERN RECOGNITION, 2001, 34 (12) : 2283 - 2303
  • [2] Deep learning based semantic segmentation of μCT images for creating digital material twins of fibrous reinforcements
    Ali, Muhammad A.
    Guan, Qiangshun
    Umer, Rehan
    Cantwell, Wesley J.
    Zhang, TieJun
    [J]. COMPOSITES PART A-APPLIED SCIENCE AND MANUFACTURING, 2020, 139
  • [3] Computationally efficient method of tracking fibres in composite materials using digital image correlation
    Amjad, K.
    Christian, W. J. R.
    Dvurecenska, K.
    Chapman, M. G.
    Uchic, M. D.
    Przybyla, C. P.
    Patterson, E. A.
    [J]. COMPOSITES PART A-APPLIED SCIENCE AND MANUFACTURING, 2020, 129 (129)
  • [4] Validation of Deep Learning Segmentation of CT Images of Fiber-Reinforced Composites
    Badran, Aly
    Parkinson, Dula
    Ushizima, Daniela
    Marshall, David
    Maillet, Emmanuel
    [J]. JOURNAL OF COMPOSITES SCIENCE, 2022, 6 (02):
  • [5] Automated segmentation of computed tomography images of fiber-reinforced composites by deep learning
    Badran, Aly
    Marshall, David
    Legault, Zacharie
    Makovetsky, Ruslana
    Provencher, Benjamin
    Piche, Nicolas
    Marsh, Mike
    [J]. JOURNAL OF MATERIALS SCIENCE, 2020, 55 (34) : 16273 - 16289
  • [6] GENERALIZING THE HOUGH TRANSFORM TO DETECT ARBITRARY SHAPES
    BALLARD, DH
    [J]. PATTERN RECOGNITION, 1981, 13 (02) : 111 - 122
  • [7] Fiber orientation measurements in composite materials
    Blanc, R
    Germain, C
    Da Costa, JP
    Baylou, P
    Cataldi, M
    [J]. COMPOSITES PART A-APPLIED SCIENCE AND MANUFACTURING, 2006, 37 (02) : 197 - 206
  • [8] Anomaly Detection of Microstructural Defects in Continuous Fiber Reinforced Composites
    Bricker, Stephen
    Simmons, J. P.
    Przybyla, Craig
    Hardie, Russell
    [J]. COMPUTATIONAL IMAGING XIII, 2015, 9401
  • [9] In situ X-ray microtomography characterization of damage in SiCf/SiC minicomposites
    Chateau, C.
    Gelebart, L.
    Bornert, M.
    Crepin, J.
    Boller, E.
    Sauder, C.
    Ludwig, W.
    [J]. COMPOSITES SCIENCE AND TECHNOLOGY, 2011, 71 (06) : 916 - 924
  • [10] Material extrusion of highly-loaded silicon nitride aqueous inks for solid infilled structures
    Costakis Jr, William J.
    Wyckoff, Connor
    Schlup, Andrew
    Wallace, Margeaux
    Craigs, Tyriek
    Malek, Elizabeth
    Hilmas, Ashley
    Rueschhoff, Lisa
    [J]. ADDITIVE MANUFACTURING, 2023, 64