3D 프린팅된 인장 시편의 인장강도 향상을 위한 머신러닝
Machine Learning to Improve Tensile Strength of 3D-Printed Tensile Specimens
0
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
As the range of 3D printed applications expands, there is an increasing demand for the production of outputs with excellent durability and reliability. In this study, the highest tensile strength printing condition was identified by printing a tensile test specimen using PLA (Poly Lactic Acid) resin, considering various printing conditions. To determine the optimal combination of printing conditions, various machine learning algorithms were compared, and Stochastic Gradient Descent(SGD) demonstrated the best performance in predicting tensile strength. Using SGD, 3,000 sets of printing conditions were generated by combining various parameters, and the best printing condition set was selected. A tensile test specimen was then produced according to the selected printing conditions, and the subsequent tensile test yielded a measured tensile strength value of 41.86 N/mm². The predicted tensile strength value by the SGD algorithm was 43.34 N/mm², resulting in a prediction accuracy of 96.23%.
Abstract
As the range of 3D printed applications expands, there is an increasing demand for the production of outputs with excellent durability and reliability. In this study, the highest tensile strength printing condition was identified by printing a tensile test specimen using PLA (Poly Lactic Acid) resin, considering various printing conditions. To determine the optimal combination of printing conditions, various machine learning algorithms were compared, and Stochastic Gradient Descent(SGD) demonstrated the best performance in predicting tensile strength. Using SGD, 3,000 sets of printing conditions were generated by combining various parameters, and the best printing condition set was selected. A tensile test specimen was then produced according to the selected printing conditions, and the subsequent tensile test yielded a measured tensile strength value of 41.86 N/mm². The predicted tensile strength value by the SGD algorithm was 43.34 N/mm², resulting in a prediction accuracy of 96.23%.
참고문헌 (12)
- Quanjin, M., "The energy-absorbing characteristics of filament wound hybrid carbon fiber-reinforced plastic", polylactic acid tubes with different infill pattern structures, 2019, pp.38(23–24), 1067~1088
- Ma, Q., "Potentiality of MWCNT on 3D printed bio-inspired spherical-roof cubic core under quasi-static loading", J. Mech. Behav. Biomed. Mater, 136, pp.105514~ (2022)
- Sun, B., "Additive Manufacturing in Medical Applications: a Brief Review", IOP Conference Series: Materials Science and Engineering (2021), IOP Publishing
- Na, D. H., "A Study on Tensile Strength Considering Weight and Printing Time of 3D Infill Patterns using 3D Printing", Transactions of Materials Processing, 32(5), pp.255~267 (2023)
- Kim, K. A., "Evaluation of Mechanical Properties by Process Factorss of ABS Specimens for 3D Printing", J. Korean Soc. Mech. Technol, 23(2), pp.300~305 (2021)
- Woo, I. Y., "Improvement of Tensile Strength through Asymmetric Tool Path in Material Extrusion-Type 3D Printing", Polym. Korea, 45(4), pp.649~653 (2021)
- Na, D. H., "A Study on Tensile Strength Dependent on Variation of Infill Pattern and Density of PLA+ Material Using 3D Printing", Transactions of Materials Processing, 31(5), pp.281~289 (2022)
- Patro, P.K, "Investigation of Mechanical Properties of 3D Printed Sandwich Structures Uisng PLA and ABS", Materials Today: Proceedings[DOI]
- Durga Rajesh, K.V, "Experimental Research on the Mechanical Characteristics of Fused Deposition Modelled ABS, PLA and PETG Specimens Printed in 3D", Materials Today: Proceedings[DOI]
- Kim, B. J., "A study of mechanical properties with FDM 3D printing layer conditions", The Korea Society Die & Mold Engineering, 12(3), pp.19~24 (2018)
- Hardt, M., "Train faster, generalize better : Stability of stochastic gradient descent", International conference on machine learning, pp.1225~1234 (2016), PMLR
- Vijayalakshmi, K., "Prediction of Virtual Generge Storage Capacity of the Air-Conditioner using a Stochastic Gradient Descent Based Artificial Neural Network", Electric Power Systems Research, 208, pp.107879~ (2022)