PSI - Issue 83

Miloslav Kepka et al. / Procedia Structural Integrity 83 (2026) 138–145

145

Therefore, a more advanced CNN-based system capable of full defect segmentation was developed. This will provide granular data on the type, magnitude, and precise coordinates of each powder bed defect. By integrating this with the known spatial coordinates of the printed parts, we aim to create a robust, fully automated diagnostic system. The final stage of this research will involve training a predictive model to correlate these in-situ observations with potential structural failures within the printed parts.

Fig. 11. Showcase of defect segmentation.

4. Conclusions The contribution aimed to show how research conducted in international cooperation with experts from the RTI research center clarified some aspects of additive manufacturing, in particular, the influence of surface properties on the integrity of metal materials and components printed using SLM technology. Another area of research that is open for international collaboration was presented. It concerns the use of knowledge from in situ monitoring of the printing process to predict the mechanical properties of printed parts using machine learning, i.e., a contribution to quality management in additive manufacturing. The solution has already been launched within the framework of the project with the acronym RTIdigi. Acknowledgements This conference contribution was supported by the project CZ.02.01.01/00/23_021/0009165 “Development of digital twins of structural components with support of on-line monitoring of their operational loading and simulations in laboratory conditions” funded by the Johannes Amos Comenius Programme. The project is co-funded by the European Union. References Simson, T., Koch, J., Rosenthal, J., Kepka, M., Zetek, M., 2019. Mechanical Properties of 18Ni-300 maraging steel manufactured by LPBF, Procedia Structural Integrity 17 (2019) 843–849, https://doi.org/10.1016/j.prostr.2019.08.112 Stepanek, M., Zetek, M., Zetkova, I., Kepka jr., M., Margetin, M., Chmelko, V., Kepka, M., 2024. Parametric predictions of fatigue life of printed samples from MS1 and 316L materials, Procedia Structural Integrity 53 (2024) 58–64, https://doi.org/10.1016/j.prostr.2024.01.008 Lee, S., Shao, S., Wells, D.N., Zetek, M., Kepka, M., Shamsaei, N., 2022. Fatigue behavior and modelling of additively manufactured IN718: The effect of surface treatments and surface measurement techniques. Journal of Materials Processing Technology, 2022, 302, 117475. https://doi.org/10.1016/j.jmatprotec.2021.117475 Böhm, M., Nies ł ony, A., Derda, S., Owsi ń ski, R., Kepka jr., M., Zetkova, I., Zetek, M., Houdkova, S., Prazmowski, M., 2023. General Reference and Design S–N Curves Obtained for 1.2709 Tool Steel. Materials 2023, 16, 1823, https://doi.org/10.3390/ma16051823 Owsi ń ski, R., Böhm, M., Nies ł ony, A., Kepka, M., Kepka jr., M., Zetkova, I., Vostrak, R., Houdkova, S., 2025. Fatigue resistance of thermal ‑ sprayed additively manufactured MS1 steel under varied mechanical loading conditions. Archives of Civil and Mechanical Engineering (2025) 25:206, https://link.springer.com/article/10.1007/s43452-025-01261-z