Title: DEEP LEARNING AND THERMAL STRAIN VISION FOR RECONSTRUCTION OF FIBER ORIENTATION DISTRIBUTION IN GEOMETRICALLY COMPLEX MOLDED COMPOSITE PARTS

Authors: Richard A. Larson, Mohammad N. Saquib, Jiang Li, Anthony J. Favaloro, Drew E. Sommer, Benjamin R. Denos, R. Byron Pipes, Sergey G. Kravchenko, Oleksandr G. Kravchenko

DOI: 10.33599/nasampe/s.23.0197

Abstract: This study aims to evaluate the ability of artificial intelligence tools to reconstruct local fiber orientation distribution (FOD) in a geometrically complex 3D prepreg platelet molded composite (PPMC) part. A deep convolutional neural network (DCNN) architecture was employed to accurately predict FOD in the entirety of molded pin bracket using thermally induced strain on the surface of the component. The developed DCNN model was trained using thousands of synthetic finite element morphologies of PPMC plates. The training data included PPMC plates with various degrees of alignment. The U-Net was able to accurately and rapidly predict FOD in the simulated PPMC plates and then was deployed to predict FOD in 3D molded component. The proposed methodology can make predictions of the spatially varying FOD in geometrically complex parts and can be used as a part of non-destructive inspection process to detect molded components with erroneous fiber orientation.

References: [1] P. Feraboli, E. Peisto, T. Cleveland, and P. Stickler, “Characterization of Prepreg-Based Discontinuous Carbon Fiber/Epoxy Systems,” J. Reinf. Plast. Compos., vol. 28, no. 10, 2008, doi: https://doi.org/10.1177/0731684408088883. [2] D. E. Sommer, S. G. Kravchenko, B. R. Denos, A. J. Favaloro, and R. B. Pipes, “Integrative analysis for prediction of process-induced, orientation-dependent tensile properties in a stochastic prepreg platelet molded composite,” Compos. Part Appl. Sci. Manuf., vol. 130, p. 105759, Mar. 2020, doi: 10.1016/j.compositesa.2019.105759. [3] S. G. Kravchenko et al., “Tensile properties of a stochastic prepreg platelet molded composite,” Compos. Part Appl. Sci. Manuf., vol. 124, p. 105507, Sep. 2019, doi: 10.1016/j.compositesa.2019.105507. [4] I. Taketa, T. Okabe, H. Matsutani, and A. Kitano, “Flowability of unidirectionally arrayed chopped strands in compression molding,” Compos. Part B Eng., vol. 42, no. 6, pp. 1764–1769, Sep. 2011, doi: 10.1016/j.compositesb.2011.01.021. [5] D. LeBlanc, B. Landry, A. Levy, P. Hubert, S. Roy, and A. Yousefpour, “Compression moulding of complex parts using randomly-oriented strands thermoplastic composites,” Proc. SAMPE Tech. Conf., 2014. [6] I. NORDAM Group, “Boeing 787 features composite window frames,” Reinf. Plast. Appl. News, vol. 5, no. 3, p. 4, 2007. [7] “Hexcel ready to fly on the A350 XWB,” Reinf. Plast., vol. 57, no. 3, pp. 25–26, May 2013, doi: 10.1016/S0034-3617(13)70089-4. [8] “NORDAM adds new product type to Airbus portfolio with supply contract to build cargo-floor ‘crutch’ for A350-1000 | Airframer.” https://www.airframer.com/news_story.html?release=34310 (accessed Jun. 13, 2021). [9] B. Denos and S. G. Kravchenko, “Simulations of mechanical properties for discontinuous prepreg composites,” SPE Automot. Compos. Conf. Exhib., 2018. [10] T. A. Oswald, Discontinuous Fiber Composites. 2019. [Online]. Available: www.mdpi.com/books, 2019 [11] S. Sattar, B. Beltran Laredo, D. Pedrazzoli, M. Zhang, S. G. Kravchenko, and O. G. Kravchenko, “Mechanical behavior of long discontinuous glass fiber nylon composite produced by in-situ polymerization,” Compos. Part Appl. Sci. Manuf., vol. 154, p. 106779, Mar. 2022, doi: 10.1016/j.compositesa.2021.106779. [12] S. Sattar, B. Laredo, S. Kravchenko, and O. Kravchenko, “Effect of Platelet Length and Stochastic Morphology on Flexural Behavior or Prepreg Platelet Molded Composites,” Polym. Compos., doi: 10.1002/pc.27230. [13] J. Teuwsen, S. K. Hohn, and T. A. Osswald, “Direct Fiber Simulation of a Compression Molded Ribbed Structure Made of a Sheet Molding Compound with Randomly Oriented Carbon/Epoxy Prepreg Strands—A Comparison of Predicted Fiber Orientations with Computed Tomography Analyses,” J. Compos. Sci., vol. 4, no. 4, Art. no. 4, Dec. 2020, doi: 10.3390/jcs4040164. [14] S. G. Kravchenko and R. B. Pipes, “Progressive Failure Analysis in Discontinuous Composite System of Prepreg Platelets with Stochastic Meso-Morphology,” p. 14. [15] D. E. Sommer, S. G. Kravchenko, and R. B. Pipes, “A numerical study of the meso-structure variability in the compaction process of prepreg platelet molded composites,” Compos. Part Appl. Sci. Manuf., vol. 138, p. 106010, Nov. 2020, doi: 10.1016/j.compositesa.2020.106010. [16] P. Feraboli, T. Cleveland, M. Ciccu, P. Stickler, and L. DeOto, “Defect and damage analysis of advanced discontinuous carbon/epoxy composite materials,” Compos. Part Appl. Sci. Manuf., vol. 41, no. 7, pp. 888–901, Jul. 2010, doi: 10.1016/j.compositesa.2010.03.002. [17] B. Denos, S. G. Kravchenko, and B. Pipes, “Progressive failure analysis in platelet based composites using CT-measured local microstructure,” SAMPE May 22-25 2017 Soc. Adv. Mater. Process Eng. - N. Am., 2017. [18] S. G. Kravchenko, “Failure Analysis in Platelet Molded Composite Systems,” Theses Diss. Available ProQuest, pp. 1–340, Jan. 2017. [19] S. Zambal, C. Heindl, C. Eitzinger, and J. Scharinger, “End-to-End Defect Detection in Automated Fiber Placement Based on Artificially Generated Data,” Fourteenth Int. Conf. Qual. Control Artif. Vis., p. 68, Jul. 2019, doi: 10.1117/12.2521739. [20] K. Sreeshan, R. Dinesh, and K. Renji, “Nondestructive inspection of aerospace composite laminate using thermal image processing,” SN Appl. Sci., vol. 2, no. 11, p. 1830, Oct. 2020, doi: 10.1007/s42452-020-03619-9. [21] R. A. Larson, R. Hoque, V. Jamora, J. Li, S. Kravchenko, and O. Kravchenko, “Hyperparameters Effect in Deep Convolutional Neural Network Model on Prediction of Fiber Orientation Distribution in Prepreg Platelet Molded Composites,” in AIAA SCITECH 2022 Forum, American Institute of Aeronautics and Astronautics. [22] A. J. Favaloro, D. E. Sommer, B. R. Denos, and R. B. Pipes, “Simulation of prepreg platelet compression molding: Method and orientation validation,” J. Rheol., vol. 62, no. 6, Art. no. 6, Oct. 2018 [23] O. Ronneberger, P. Fischer, and T. Brox, “U-Net: Convolutional Networks for Biomedical Image Segmentation,” ArXiv150504597 Cs, May 2015, Accessed: Dec. 15, 2020. [Online]. Available: http://arxiv.org/abs/1505.04597 [24] “Digimat - The Material Modeling Platform.” https://www.mscsoftware.com/product/digimat (accessed Jun. 01, 2021). [25] “Abaqus 2018 Online Documentation.” Dassault Systemes, Johnston, RI. [26] S. G. Advani and C. L. Tucker, “The Use of Tensors to Describe and Predict Fiber Orientation in Short Fiber Composites,” J. Rheol., vol. 31, no. 8, pp. 751–784, Nov. 1987, doi: 10.1122/1.549945. [27] R. Larson, R. Hoque, V. Jamora, J. Li, S. Kravchenko, and O. Kravchenko, “Reconstruction of Local Fiber Orientation State in Prepreg Platelet Molded Composites via Deep Learning,” Submitt. Eng. Appl. Artif. Intell.. [28] M. Saquib, R. Larson, S. Sattar, J. Li, S. Kravchenko, and O. Kravchenko, “Fiber Orientation Distribution Reconstruction in Discontinuous Molded Composites Using Deep Learning,” Submitt. Compos. Part A.

Conference: SAMPE 2023

Publication Date: 2023/04/17

SKU: TP23-0000000197

Pages: 15

Price: $30.00