隨著金屬積層製造技術的發展，選擇性雷射熔融（SLM）技術在工業製造中扮演越來越重要的角色。然而，傳統的磁特性預測方法無法滿足現代製造業對效率和精確度的高要求。本研究提出了一種基於三維卷積視覺轉換模型（3D-CVT）的深度學習方法，用於快速且準確地預測SLM製程成品的磁特性。
3D-CVT模型融合了卷積神經網絡和視覺轉換器的優點，提高了對空間和特徵信息的理解能力，從而在預測磁性能方面達到了比傳統機器學習方法更高的精確度和效率。通過熱力圖技術，我們能夠直觀地展示模型在影像數據中哪些區域的特徵對預測結果影響最大，這有助於我們更好地理解製程參數與成品磁特性之間的關聯。
此外，本研究採用了假設檢定來評估模型預測的統計顯著性，確保了研究結果的可靠性和有效性。透過這些方法的應用，不僅提升了積層製造的品質監控能力，也為製造業提供了一種高效且經濟的品質預測工具。

With the development of metal additive manufacturing technology, Selective Laser Melting (SLM) has become increasingly important in industrial manufacturing. However, traditional methods of predicting magnetic properties cannot meet the high demands for efficiency and accuracy required by modern manufacturing. This study introduces a deep learning approach based on the 3D Convolutional Vision Transformer (3D-CVT) model for the rapid and accurate prediction of magnetic properties in SLM process products.

The 3D-CVT model integrates the advantages of convolutional neural networks and vision transformers, enhancing the understanding of spatial and feature information, thereby achieving higher accuracy and efficiency in predicting magnetic properties compared to traditional machine learning methods. By heat map technology, we can visually display which areas of the image data have the most significant impact on the prediction results, helping us better understand the relationship between process parameters and the magnetic properties of the products.

Additionally, this study employs hypothesis testing to evaluate the statistical significance of the model predictions, ensuring the reliability and validity of the research results. Through the application of these methods, not only is the quality control capability of additive manufacturing enhanced, but also a highly efficient and economical quality prediction tool is provided for the manufacturing industry.

[1] Yap, C. Y., Chua, C. K., Dong, Z. L., Liu, Z. H., Zhang, D. Q., Loh, L. E., & Sing, S. L. (2015). Review of selective laser melting: Materials and applications. Applied Physics Reviews, 2(4), 041101. https://doi.org/10.1063/1.4935926

[2] Tucho, W. M., Lysne, V. H., Austbø, H., Sjolyst-Kverneland, A., & Hansen, V. (2018). Investigation of effects of process parameters on microstructure and hardness of SLM manufactured SS316L. Journal of Alloys and Compounds, 740, 910-925. https://doi.org/10.1016/j.jallcom.2018.01.098

[3] Liu, S., Yang, W., Shi, X., Li, B., Duan, S., Guo, H., & Guo, J. (2019). Influence of laser process parameters on the densification, microstructure, and mechanical properties of a selective laser melted AZ61 magnesium alloy. Journal of Alloys and Compounds, 808, 151160. https://doi.org/10.1016/j.jallcom.2019.06.261

[4] Giganto, S., Zapico, P., Castro-Sastre, M. Á., Martínez-Pellitero, S., Leo, P., & Perulli, P. (2019). Influence of the scanning strategy parameters upon the quality of the SLM parts. Procedia Manufacturing, 41, 698-705. https://doi.org/10.1016/j.promfg.2019.09.060

[5] Aqilah, D. N., Sayuti, A. K. M., Farazila, Y., Suleiman, D. Y., Amirah, M. A. N., & Izzati, W. B. W. N. (2018). Effects of process parameters on the surface roughness of stainless steel 316L parts produced by selective laser melting. Journal of Testing and Evaluation, 46(4), 1673–1683. https://doi.org/10.1520/JTE20170140

[6] Craeghs, T., Clijsters, S., Yasa, E., & Kruth, J.-P. (2011). Online quality control of selective laser melting. In Proceedings of the 22nd Annual International Solid Freeform Fabrication Symposium - An Additive Manufacturing Conference, SFF 2011. http://dx.doi.org/10.26153/tsw/15289

[7] Chaudhary, V., Mantri, S. A., Ramanujan, R. V., & Banerjee, R. (2020). Additive manufacturing of magnetic materials. Progress in Materials Science, 114, 100688. https://doi.org/10.1016/j.pmatsci.2020.100688

[8] Chen, Y., Peng, X., Kong, L., Dong, G., Remani, A., & Leach, R. (2021). Defect inspection technologies for additive manufacturing. International Journal of Extreme Manufacturing, 3(2), 022002. https://doi.org/10.1088/2631-7990/abe0d0

[9] Maurício, J., Domingues, I., & Bernardino, J. (2023). Comparing vision transformers and convolutional neural networks for image classification: A literature review. Applied Sciences, 13, 5521. https://doi.org/10.3390/app13095521

[10] Bakhtiarnia, A., Zhang, Q., & Iosifidis, A. (2022). Single-layer vision transformers for more accurate early exits with less overhead. Neural Networks, 153, 461-473. https://doi.org/10.1016/j.neunet.2022.06.038

[11] Wu, H., Xiao, B., Codella, N., Liu, M., Dai, X., Yuan, L., & Zhang, L. (2021). CvT: Introducing convolutions to vision transformers. In Proceedings of the 2021 IEEE/CVF International Conference on Computer Vision (ICCV) (pp. 22-31). https://doi.org/10.1109/ICCV48922.2021.00009

[12] Gibson, I., Rosen, D., Stucker, B., & Khorasani, M. (2021). Additive manufacturing technologies (3rd ed.). Springer Nature Switzerland AG. https://doi.org/10.1007/978-3-030-56127-7

[13] Zhang, B., Li, Y., & Bai, Q. (2017). Defect formation mechanisms in selective laser melting: A review. Chinese Journal of Mechanical Engineering, 30, 515–527. https://doi.org/10.1007/s10033-017-0121-5

[14] Guo, Q., Zhao, C., Qu, M., Xiong, L., Escano, L. I., Hojjatzadeh, S. M. H., Parab, N. D., Fezzaa, K., Everhart, W., Sun, T., & Chen, L. (2019). In-situ characterization and quantification of melt pool variation under constant input energy density in laser powder bed fusion additive manufacturing process. Additive Manufacturing, 28, 600-609. https://doi.org/10.1016/j.addma.2019.04.021

[15] Campanelli, S. L., Casalino, G., Contuzzi, N., & Ludovico, A. D. (2013). Taguchi optimization of the surface finish obtained by laser ablation on selective laser molten steel parts. Procedia CIRP, 12, 462-467. https://doi.org/10.1016/j.procir.2013.09.079

[16] Tumański, S. (2011). Handbook of magnetic measurements. CRC Press. https://doi.org/10.1201/b10979

[17] Fiorillo, F. (2004). Characterization and measurement of magnetic materials. Academic Press. https://doi.org/10.1016/B978-0-12-257251-7.X5000-X

[18] Jiles, D. C. (2000). Dynamics of domain magnetization and the Barkhausen effect. Czech Journal of Physics, 50, 893–924. https://doi.org/10.1023/A:1022846128461

[19] Szeliski, R. (2010). Computer vision: Algorithms and applications. Springer London. https://doi.org/10.1007/978-1-84882-935-0

[20] Materka, A., & Strzelecki, M. (1998). Texture analysis methods–a review. Technical University of Lodz, Institute of Electronics, COST B11 report, Brussels, 10(1.97), 4968.

[21] Gadelmawla, E. S., Koura, M. M., Maksoud, T. M. A., Elewa, I. M., & Soliman, H. H. (2002). Roughness parameters. Journal of Materials Processing Technology, 123(1), 133-145. https://doi.org/10.1016/S0924-0136(02)00060-2

[22] LeCun, Y., Bengio, Y., & Hinton, G. (2015). Deep learning. Nature, 521, 436–444. https://doi.org/10.1038/nature14539

[23] Azimi, S. M., Britz, D., Engstler, M., Fritz, M., & Mücklich, F. (2018). Advanced steel microstructural classification by deep learning methods. Scientific Report, 8, 2128. https://doi.org/10.1038/s41598-018-20037-5

[24] Holm, E. A., Cohn, R., Gao, N., Kitahara, A. R., Matson, T. P., Lei, B., & Yarasi, S. R. (2020). Overview: Computer Vision and Machine Learning for Microstructural Characterization and Analysis. Metall Mater Trans A 51, 5985–5999. https://doi.org/10.1007/s11661-020-06008-4

[25] Wang, J., Ma, Y., Zhang, L., Gao, R. X., & Wu, D. (2018). Deep learning for smart manufacturing: Methods and applications. Journal of Manufacturing Systems, 48(Part C), 144-156. https://doi.org/10.1016/j.jmsy.2018.01.003

[26] Weimer, D., Scholz-Reiter, B., & Shpitalni, M. (2016). Design of deep convolutional neural network architectures for automated feature extraction in industrial inspection. CIRP Annals, 65(1), 417-420. https://doi.org/10.1016/j.cirp.2016.04.072

[27] Dosovitskiy, A., Beyer, L., Kolesnikov, A., Weissenborn, D., Zhai, X., Unterthiner, T., Dehghani, M., Minderer, M., Heigold, G., Gelly, S., Uszkoreit, J., & Houlsby, N. (2020). An image is worth 16x16 words: Transformers for image recognition at scale. arXiv preprint arXiv:2010.11929. https://doi.org/10.48550/arXiv.2010.11929

[28] Zhang, Y., Hong, G. S., Ye, D., Zhu, K., & Fuh, J. Y. H. (2018). Extraction and evaluation of melt pool, plume, and spatter information for powder-bed fusion AM process monitoring. Materials & Design, 156, 458-469. https://doi.org/10.1016/j.matdes.2018.07.002

[29] Scime, L., & Beuth, J. (2018). Anomaly detection and classification in a laser powder bed additive manufacturing process using a trained computer vision algorithm. Additive Manufacturing, 19, 114-126. https://doi.org/10.1016/j.addma.2017.11.009

[30] 陳日昇（2023）。積層製造成品之三維影像建模與品質預測。﹝碩士論文。國立成功大學機械工程學系﹞。 https://hdl.handle.net/11296/u96xbm。

[31] Kingma, D. P., & Ba, J. (2014). Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. https://doi.org/10.48550/arXiv.1412.6980

[32] Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., & Torralba, A. (2015). Learning Deep Features for Discriminative Localization. arXiv preprint arXiv:1512.04150. https://doi.org/10.48550/arXiv.1512.04150

[33] Selvaraju, R. R., Cogswell, M., Das, A., Vedantam, R., Parikh, D., & Batra, D. (2019). Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization. International Journal of Computer Vision (IJCV). https://doi.org/10.1007/s11263-019-01228-7