한국전산구조공학회논문집 (Journal of the Computational Structural Engineering Institute of Korea)

  • 제36권5호
  • /
  • Pages.339-346
  • /
  • 2023
  • /
  • 1229-3059(pISSN)
  • /
  • 2287-2302(eISSN)
한국전산구조공학회 (Computational Structural Engineering Institute of Korea)
권호 표지
DOI QR코드

DOI QR Code

시멘트풀의 공극분포특성에 기반한 인장강도 예측 CNN 모델

CNN Model for Prediction of Tensile Strength based on Pore Distribution Characteristics in Cement Paste

  • 홍성욱 (연세대학교 건설환경공학과) ;
  • 한동석 (연세대학교 건설환경공학과)
  • Sung-Wook Hong (Department of Civil and Environmental Engineering, Yonsei University) ;
  • Tong-Seok Han (Department of Civil and Environmental Engineering, Yonsei University)
  • 투고 : 2023.08.30
  • 심사 : 2023.09.12
  • 발행 : 2023.10.31
PDF 다운로드
⟨ 이전 논문 다음 논문 ⟩

초록

미세구조 특성의 불확실성은 재료 특성에 많은 영향을 준다. 시멘트 기반 재료의 공극 분포 특성은 재료의 역학적 특성에 큰 영향을 미치며, 재료에 랜덤하게 분포되어 있는 많은 공극은 재료의 물성 예측을 어렵게 한다. 공극의 특성 분석과 재료 응답 간의 상관관계 규명에 대한 기존 연구는 통계적 관계 분석에 국한되어 있으며, 그 상관관계가 아직 명확히 규명되어 있지 않다. 본 연구에서는 합성곱 신경망(CNN, convolutional neural network)을 활용한 이미지 기반 데이터 접근법을 통해 시멘트 기반 재료의 역학적 응답을 예측하고, 공극분포와 재료 응답의 상관관계를 분석하였다. 머신러닝을 위한 데이터는 고해상도 마이크로-CT 이미지와 시멘트 기반 재료의 물성(인장강도)로 구성하였다. 재료의 메시 구조 특성을 분석하였으며, 재료의 응답은 상장균열모델(phase-field fracture model)에 기반을 둔 2D 직접 인장(direct tension) 유한요소해석 시뮬레이션을 활용하여 평가하였다. 입력 이미지 영역의 기여도를 분석하여 시편에서 재료 응답 예측에 가장 큰 영향을 미치는 영역을 CNN을 통하여 식별하였다. CNN 과정 중 활성 영역과 공극분포를 비교 분석하여 공극분포특성과 재료 응답의 상관관계를 분석하여 제시하였다.

The uncertainties of microstructural features affect the properties of materials. Numerous pores that are randomly distributed in materials make it difficult to predict the properties of the materials. The distribution of pores in cementitious materials has a great influence on their mechanical properties. Existing studies focus on analyzing the statistical relationship between pore distribution and material responses, and the correlation between them is not yet fully determined. In this study, the mechanical response of cementitious materials is predicted through an image-based data approach using a convolutional neural network (CNN), and the correlation between pore distribution and material response is analyzed. The dataset for machine learning consists of high-resolution micro-CT images and the properties (tensile strength) of cementitious materials. The microstructures are characterized, and the mechanical properties are evaluated through 2D direct tension simulations using the phase-field fracture model. The attributes of input images are analyzed to identify the spot with the greatest influence on the prediction of material response through CNN. The correlation between pore distribution characteristics and material response is analyzed by comparing the active regions during the CNN process and the pore distribution.

키워드

과제정보

이 연구는 2023년도 산업통상자원부 및 산업기술평가관리원(KEIT) 연구비 지원에 의한 연구임(RS-2023-00264747)

참고문헌

  1. Butler, K.T., Davies, D.W., Cartwright, H., Isayev, O., Walsh, A. (2018) Machine Learning for Molecular and Materials Science, Nat., 559(7715), pp.547~555. https://doi.org/10.1038/s41586-018-0337-2
  2. Cetinic, E., Lipic, T., Grgic, S. (2018) Fine-Tuning Convolutional Neural Networks for Fine Art Classification, Expert Syst. Appl., 114, pp.107~118. https://doi.org/10.1016/j.eswa.2018.07.026
  3. Chung, S.-Y., Lehmann, C., Elrahman, M., Stephan, D. (2018) Microstructural Characterization of Foamed Concrete with Different Densities using Microscopic Techniques, Cem. Wapno Beton, 3, pp.216~225.
  4. Coker, D.A., Torquato, S. (1995) Extraction of Morphological, Quantities from a Digitized Medium, J. Appl. Phys., 77(12), pp.6087~6099. https://doi.org/10.1063/1.359134
  5. Ghiringhelli, L.M., Vybiral, J., Levchenko, S.V., Draxl, C., Scheffler, M. (2015) Big Data of Materials Science: Critical Role of the Descriptor, Phys. Rev. Lett., 114(10), p.105503.
  6. Goodfellow, I., Bengio, Y., Courville, A. (2016) Deep Learning, MIT Press, p.800.
  7. Han, T.-S., Zhang, X., Kim, J.-S., Chung, S.-Y., Lim, J.-H., Linder, C. (2018) Area of Linal-Path Function for Describing the Pore Microstructures of Cement Paste and Their Relations to the Mechanical Properties Simulated from µ-CT Microstructures, Cem. Concr. Compos., 89, pp.1~17. https://doi.org/10.1016/j.cemconcomp.2018.02.008
  8. He, K., Zhang, X., Ren, S., Sun, J. (2016) Deep Residual Learning for Image Recognition, Proc. IEEE Conf. Computer Vision and Pattern Recognition, pp.770~778.
  9. Kim, J.-S., Chung, S.-Y., Stephan, D., Han, T.-S. (2019a) Issues on Characterization of Cement Paste Microstructures from µ-CT and Virtual Experiment Framework for Evaluating Mechanical Properties, Constr. & Build. Mater., 202, pp.82~102. https://doi.org/10.1016/j.conbuildmat.2019.01.030
  10. Kim, J.-S., Kim, J.-H., Han, T.-S. (2019b) Microstructure Characterization of Cement Paste from Micro-CT and Correlations with Mechanical Properties Evaluated from Virtual and Real Experiments, Mater. Charact., 155, p.109807.
  11. Kim, J.-S., Lim, J.-H., Stephan, D., Park, K., Han, T.-S. (2022) Mechanical behavior Comparison of Single and Multiple Phase Models for Cement Paste using Micro-CT Images and Nanoindentation, Constr. & Build. Mater., 342, p.127938.
  12. Krizhevsky, A., Sutskever, I., Hinton, G.E. (2012) Imagenet Classification with Deep Convolutional Neural Networks, In Advances in Neural Information Processing Systems, pp. 1097~1105.
  13. Ma, H., Xu, B., Liu, J., Pei, H., Li, Z. (2014) Effect of Water Content, Magnesia-to-Phosphate Molar Ratio and Age on Pore Structure, Strength and Permeability of Magnesium Potassium Phosphate Cement Paste, Mater. Des., 64, pp. 497~502. https://doi.org/10.1016/j.matdes.2014.07.073
  14. Maruyama, I., Nishioka, Y., Igarashi, G., Matsui, K. (2014) Microstructural and Bulk Property Changes in Hardened Cement Paste During the First Drying Process, Cem. Concr. Res., 58, pp.20~34. https://doi.org/10.1016/j.cemconres.2014.01.007
  15. Miehe, C., Schanzel, L.-M., Ulmer, H. (2015) Phase Field Modeling of Fracture in Multi-Physics Problems, Part I: Balance of Crack Surface and Failure Criteria for Brittle Crack Propagation in Thermo-Elastic Solids, Comput. Methods Appl. Mech. Eng., 294, pp.449~485. https://doi.org/10.1016/j.cma.2014.11.016
  16. Miehe, C., Hofacker, M., Welschinger, F. (2010) A Phase Field Model for Rate-Independent Crack Propagation: Robust Algorithmic Implementation based on Operator Splits, Comput. Methods Appl. Mech. & Eng., 199(45-48), pp.2765~2778. https://doi.org/10.1016/j.cma.2010.04.011
  17. Pichler, B., Hellmich, C., Eberhardsteiner, J., Wasserbauer, J., Termkhajornkit, P., Barbarulo, R., Chanvillard, G. (2014) Effect of Gelspace Ratio and Microstructure on Strength of Hydrating Cementitious Materials: An Engineering Micromechanics Approach, Cem & Concr. Res., 45, pp.55~68.
  18. Seko, A., Togo, A., Tanaka, I. (2018) Descriptors for Machine Learning of Materials Data, Nanoinformatics; Tanaka, I., Ed.; Springer: Singapore.
  19. Selvaraju, R.R., Das, A., Vedantam, R., Cogswell, M., Parikh, D., Batra, D. (2016) Grad-cam: Why Did You Say that? Visual Explanations from Deep Networks via Gradient-based Localization, arXiv preprint arXiv:1610.02391.
  20. Simonyan, K., Zisserman, A. (2014) Very Deep Convolutional Networks for Large-Scale Image Recognition, arXiv preprint arXiv:1409.1556.
  21. Singh, H., Gokhale, A., Lieberman, S., Tamirisakandala, S. (2008) Image based Computations of Lineal Path Probability Distributions for Microstructure Representation, Mater. Sci. Eng., A, 474, pp.104~111. https://doi.org/10.1016/j.msea.2007.03.099
  22. Swann, E., Sun, B., Cleland, D.M. (2018) Barnard, A.S. Representing Molecular and Materials Data for Unsupervised Machine Learning, Mol. Simul., 44(11), pp.905~920. https://doi.org/10.1080/08927022.2018.1450982
  23. Szegedy, C., Liu, W., Jia, Y., Sermanet, P., Reed, S., Anguelov, D., Erhan, D., Vanhoucke, V., Rabinovich, A. (2015) Going Deeper with Convolutions, Proc. IEEE Conf. Computer Vision and Pattern Recognition, pp.1~9.
  24. Takahashi, K., Tanaka, Y. (2016) Materials Informatics: A Journey Towards Material Design and Synthesis, Dalton Trans., 45(26), pp.1497~1499.
  25. Thanapol, P., Lavangnananda, K., Bouvry, P., Pinel, F., Leprevost, F. (2020) Reducing Overfitting and Improving Generalization in Training Convolutional Neural Network (CNN) under Limited Sample Sizes in Image Recognition, In 2020-5th International Conference on Information Technology, pp.300~305.
  26. Wu, J.-Y. (2019) X-ray Computed Tomography Images based Phase-Field Modeling of Mesoscopic Failure in Concrete, Eng. Fract. Mech., 208, pp.151~170. https://doi.org/10.1016/j.engfracmech.2019.01.005
  27. Zhou, B., Khosla, A., Lapedriza, A., Oliva, A., Torralba, A. (2016) Learning Deep Features for Discriminative Localization, Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition, pp.2921~2929.
  28. Ziletti, A., Kumar, D., Scheffler, M., Ghiringhelli, L.M. (2018) Insightful Classification of Crystal Structures using Deep Learning, Nat. Commun, 9(1), p.2775.