PSI - Issue 42

Satyajit Dey et al. / Procedia Structural Integrity 42 (2022) 943–951 Satyajit Dey et al / Structural Integrity Procedia 00 (2019) 000–000

950

8

Table 2. IoU values for the di ff erent defects as predicted by the segmentation model Defect Type IoU

All defects

0.61 0.82 0.68 0.43

Lack of fusion

Pores

Micro-cracks

5. Conclusion and Future Work

It can be concluded that ANN using segmentation models can be used to detect manufacturing defects in additive manufactured components. The predictions for micro-cracks were not satisfactory in this study but it can be attributed to the tiny size of micro-cracks. It could be observed that for low resolution images (320x320), even employing a strong encoder like senet154 did not result in prediction of micro-cracks whereas a smaller encoder could predict micro-cracks for a high resolution image (960x960). The requirement of high resolution images can significantly impact training and prediction e ffi ciency and therefore a reduction of image features followed by principal component analysis (PCA) will be employed in future to obtain an optimised manufacturing defect detection tool for components made by additive manufacturing.

Acknowledgements

GM and AAM are funded through the Seˆr Cymru II 80761-BU-103 project by Welsh European Funding O ffi ce (WEFO) under the European Development Fund (ERDF).

References

[1] Mohsen Attaran. The rise of 3-d printing: The advantages of additive manufacturing over traditional manufacturing. Business horizons , 60(5):677– 688, 2017. [2] Matthew Hiser, Alyssa Schneider, Margaret Audrain, and Amy Hull. Regulatory research perspective on additive manufacturing for nuclear component applications. Journal of Nuclear Materials , 546:152726, 2021. [3] C Hensley, K Sisco, S Beauchamp, A Godfrey, H Rezayat, T McFalls, D Galicki, F List III, K Carver, C Stover, et al. Qualification pathways for additively manufactured components for nuclear applications. Journal of Nuclear Materials , 548:152846, 2021. [4] Jack Sklansky. Image segmentation and feature extraction. IEEE Transactions on Systems, Man, and Cybernetics , 8(4):237–247, 1978. [5] Salem Saleh Al-Amri, Namdeo V Kalyankar, et al. Image segmentation by using threshold techniques. arXiv preprint arXiv:1005.4020 , 2010. [6] Aly A Farag. Edge-based image segmentation. Remote sensing reviews , 6(1):95–121, 1992. [7] Yanni Zou and Bo Liu. Survey on clustering-based image segmentation techniques. In 2016 IEEE 20th International Conference on Computer Supported Cooperative Work in Design (CSCWD) , pages 106–110. IEEE, 2016. [8] Catalin Amza. A review on neural network-based image segmentation techniques. De Montfort University, Mechanical and Manufacturing Engg., The Gateway Leicester, LE1 9BH, United Kingdom , 1:23, 2012. [9] Neeraj Sharma, Lalit M Aggarwal, et al. Automated medical image segmentation techniques. Journal of medical physics , 35(1):3, 2010. [10] Yongbin Chen, Jingran Wang, and Guitang Wang. Intelligent welding defect detection model on improved r-cnn. IETE Journal of Research , pages 1–10, 2022. [11] Christian Nowroth, Tiansheng Gu, Jan Grajczak, Sarah Nothdurft, Jens Twiefel, Jo¨ rg Hermsdorf, Stefan Kaierle, and Jo¨rg Wallaschek. Deep learning-based weld contour and defect detection from micrographs of laser beam welded semi-finished products. Applied Sciences , 12(9):4645, 2022. [12] Yu Cheng, HongGui Deng, YuXin Feng, and JunJiang Xiang. Weld defect detection and image defect recognition using deep learning technology. 2021. [13] Yanzhou Fu, Austin RJ Downey, Lang Yuan, Tianyu Zhang, Avery Pratt, and Yunusa Balogun. Machine learning algorithms for defect detection in metal laser-based additive manufacturing: a review. Journal of Manufacturing Processes , 75:693–710, 2022. [14] Rui Li, Mingzhou Jin, and Vincent C Paquit. Geometrical defect detection for additive manufacturing with machine learning models. Materials & Design , 206:109726, 2021. [15] Lequn Chen, Xiling Yao, Peng Xu, Seung Ki Moon, and Guijun Bi. Rapid surface defect identification for additive manufacturing with in-situ point cloud processing and machine learning. Virtual and Physical Prototyping , 16(1):50–67, 2021.