Title: Machine Learning for Reduced-Order Modeling of Composites Processing

Authors: Min Kim, Navid Zobeiry

DOI: 10.33599/nasampe/s.21.0535

Abstract: Machine learning concepts have made their way into industry practice and can offer quick and accurate solutions for various design and in-service applications including process simulation of composites. High-fidelity finite element simulation tools are currently used for thermo-chemical analysis of parts during processing. This research investigates reduced-order modeling of complex composite parts using machine learning methods to speed-up the simulation. At first, process simulations of representative 2-D stringer geometries were conducted using finite element analysis to extract leading and lagging responses. Equivalency of exothermic responses of 2-D simulations to 1-D finite element analysis was established using an in-house developed machine learning framework at the University of Washington (CompML). It is demonstrated that for a given 2-D geometry, a trained NN can identify an equivalent 1-D thermal stack to yield similar exothermic responses. The results and methods developed in this study can significantly reduce the computational cost of finite element simulation by successfully establishing reduced-order models of complex geometries.

References: [1] Fabris J, Lussier D, Zobeiry N, Mobuchon C, Poursartip A. Development of standardized approaches to thermal management in composites manufacturing. Int. SAMPE Tech. Conf., 2014. [2] Fabris J, Mobuchon C, Zobeiry N, Lussier D, Fernlund G, Poursartip A. Introducing thermal history producibility assessment at conceptual design. Int. SAMPE Tech. Conf., vol. 2015- Janua, 2015. [3] Zobeiry N, Park J, Poursartip A. An infrared thermography-based method for the evaluation of the thermal response of tooling for composites manufacturing. J Compos Mater 2019;53:1277–90. https://doi.org/10.1177/0021998318798444. [4] Park J, Zobeiry N, Poursartip A. Tooling materials and their effect on surface thermal gradients. Int. SAMPE Tech. Conf., 2017. [5] Fernlund G, Mobuchon C, Zobeiry N. 2.3 Autoclave Processing. In: Carl Zweben PB, editor. Compr. Compos. Mater. II, vol. 2, 2018, p. 42–62. https://doi.org/10.1016/b978-0-12-803581-8.09899-4. [6] Fernlund G, Poursartip A, Nelson K, Wilenski M, Swanstrom F. Process modeling for dimensional control-sensitivity analysis of a composite spar process. Int. SAMPE Symp. Exhib., 1999. [7] Nelson KM, Poursartip A, Fernlund G. Cure kinetics and the dimensional control of composite structure. Int SAMPE Symp Exhib 2000. [8] Zobeiry N, Forghani A, Li C, Gordnian K, Thorpe R, Vaziri R, et al. Multiscale characterization and representation of composite materials during processing. Philos Trans R Soc A Math Phys Eng Sci 2016;374:20150278. https://doi.org/10.1098/rsta.2015.0278. [9] Zobeiry N, Stewart A, Poursartip A. Applications of Machine Learning for Process Modeling of Composites. SAMPE Virtual Conf., Seattle, WA: 2020. [10] Zobeiry N, Humfeld KD. An iterative scientific machine learning approach for discovery of theories underlying physical phenomena. ArXiv 2019;arXiv:1909. [11] Zobeiry N, Reiner J, Vaziri R. Theory-guided machine learning for damage characterization of composites. Compos Struct 2020;246:112407. https://doi.org/10.1016/j.compstruct.2020.112407. [12] Zobeiry N, Humfeld KD. A Physics-Informed Machine Learning Approach for Solving Heat Transfer Equation in Advanced Manufacturing and Engineering Applications. ArXiv 2020. [13] Wang J-XX, Wu J-LL, Xiao H. Physics-informed machine learning approach for reconstructing Reynolds stress modeling discrepancies based on DNS data. Phys Rev Fluids 2017;2:034603. https://doi.org/10.1103/PhysRevFluids.2.034603. [14] Wang J, Wu J, Xiao H. A Physics-Informed Machine Learning Approach of Improving RANS Predicted Reynolds Stresses. 55th AIAA Aerosp. Sci. Meet., Reston, Virginia: American Institute of Aeronautics and Astronautics; 2017. https://doi.org/10.2514/6.2017- 1712. [15] Www.convergent.ca/products/compro-simulation-software. COMPRO simulation software 2014. [16] Humfeld KD, Zobeiry N. Machine learning-based process simulation approach for real-time optimization and active control of composites autoclave processing. SAMPE Virtual Conf., vol. accepted, Long Beach, CA: 2021. [17] HEXCEL. HexPly 8552 Epoxy matrix (180°C/365°F curing matrix) Product Data 2016. https://www.hexcel.com/user_area/content_media/raw/HexPly_8552_eu_DataSheet.pdf

Conference: SAMPE NEXUS 2021

Publication Date: 2021/06/29

SKU: TP21-0000000535

Pages: 12

Price: FREE