StressNet - Deep learning to predict stress with fracture propagation in brittle materials
Abstract
Abstract Catastrophic failure in brittle materials is often due to the rapid growth and coalescence of cracks aided by high internal stresses. Hence, accurate prediction of maximum internal stress is critical to predicting time to failure and improving the fracture resistance and reliability of materials. Existing high-fidelity methods, such as the Finite-Discrete Element Model (FDEM), are limited by their high computational cost. Therefore, to reduce computational cost while preserving accuracy, a deep learning model, StressNet, is proposed to predict the entire sequence of maximum internal stress based on fracture propagation and the initial stress data. More specifically, the Temporal Independent Convolutional Neural Network (TI-CNN) is designed to capture the spatial features of fractures like fracture path and spall regions, and the Bidirectional Long Short-term Memory (Bi-LSTM) Network is adapted to capture the temporal features. By fusing these features, the evolution in time of the maximum internal stress can be accurately predicted. Moreover, an adaptive loss function is designed by dynamically integrating the Mean Squared Error (MSE) and the Mean Absolute Percentage Error (MAPE), to reflect the fluctuations in maximum internal stress. After training, the proposed model is able to compute accurate multi-step predictions of maximum internal stress in approximately 20 seconds,more »
- Authors:
- Publication Date:
- Research Org.:
- Los Alamos National Lab. (LANL), Los Alamos, NM (United States)
- Sponsoring Org.:
- USDOE Laboratory Directed Research and Development (LDRD) Program
- OSTI Identifier:
- 1765066
- Alternate Identifier(s):
- OSTI ID: 1822735
- Report Number(s):
- LA-UR-19-31067
Journal ID: ISSN 2397-2106; 6; PII: 151
- Grant/Contract Number:
- 89233218CNA000001
- Resource Type:
- Published Article
- Journal Name:
- npj Materials Degradation
- Additional Journal Information:
- Journal Name: npj Materials Degradation Journal Volume: 5 Journal Issue: 1; Journal ID: ISSN 2397-2106
- Publisher:
- Nature Publishing Group
- Country of Publication:
- United Kingdom
- Language:
- English
- Subject:
- 36 MATERIALS SCIENCE; Computational methods; mechanical engineering
Citation Formats
Wang, Yinan, Oyen, Diane, Guo, Weihong, Mehta, Anishi, Scott, Cory Braker, Panda, Nishant, Fernández-Godino, M. Giselle, Srinivasan, Gowri, and Yue, Xiaowei. StressNet - Deep learning to predict stress with fracture propagation in brittle materials. United Kingdom: N. p., 2021.
Web. doi:10.1038/s41529-021-00151-y.
Wang, Yinan, Oyen, Diane, Guo, Weihong, Mehta, Anishi, Scott, Cory Braker, Panda, Nishant, Fernández-Godino, M. Giselle, Srinivasan, Gowri, & Yue, Xiaowei. StressNet - Deep learning to predict stress with fracture propagation in brittle materials. United Kingdom. https://doi.org/10.1038/s41529-021-00151-y
Wang, Yinan, Oyen, Diane, Guo, Weihong, Mehta, Anishi, Scott, Cory Braker, Panda, Nishant, Fernández-Godino, M. Giselle, Srinivasan, Gowri, and Yue, Xiaowei. Wed .
"StressNet - Deep learning to predict stress with fracture propagation in brittle materials". United Kingdom. https://doi.org/10.1038/s41529-021-00151-y.
@article{osti_1765066,
title = {StressNet - Deep learning to predict stress with fracture propagation in brittle materials},
author = {Wang, Yinan and Oyen, Diane and Guo, Weihong and Mehta, Anishi and Scott, Cory Braker and Panda, Nishant and Fernández-Godino, M. Giselle and Srinivasan, Gowri and Yue, Xiaowei},
abstractNote = {Abstract Catastrophic failure in brittle materials is often due to the rapid growth and coalescence of cracks aided by high internal stresses. Hence, accurate prediction of maximum internal stress is critical to predicting time to failure and improving the fracture resistance and reliability of materials. Existing high-fidelity methods, such as the Finite-Discrete Element Model (FDEM), are limited by their high computational cost. Therefore, to reduce computational cost while preserving accuracy, a deep learning model, StressNet, is proposed to predict the entire sequence of maximum internal stress based on fracture propagation and the initial stress data. More specifically, the Temporal Independent Convolutional Neural Network (TI-CNN) is designed to capture the spatial features of fractures like fracture path and spall regions, and the Bidirectional Long Short-term Memory (Bi-LSTM) Network is adapted to capture the temporal features. By fusing these features, the evolution in time of the maximum internal stress can be accurately predicted. Moreover, an adaptive loss function is designed by dynamically integrating the Mean Squared Error (MSE) and the Mean Absolute Percentage Error (MAPE), to reflect the fluctuations in maximum internal stress. After training, the proposed model is able to compute accurate multi-step predictions of maximum internal stress in approximately 20 seconds, as compared to the FDEM run time of 4 h, with an average MAPE of 2% relative to test data.},
doi = {10.1038/s41529-021-00151-y},
journal = {npj Materials Degradation},
number = 1,
volume = 5,
place = {United Kingdom},
year = {Wed Feb 10 00:00:00 EST 2021},
month = {Wed Feb 10 00:00:00 EST 2021}
}
https://doi.org/10.1038/s41529-021-00151-y
Works referenced in this record:
Bidirectional recurrent neural networks
journal, January 1997
- Schuster, M.; Paliwal, K. K.
- IEEE Transactions on Signal Processing, Vol. 45, Issue 11
Feasibility analysis of composite fuselage shape control via finite element analysis
journal, January 2018
- Wen, Yuchen; Yue, Xiaowei; Hunt, Jeffrey H.
- Journal of Manufacturing Systems, Vol. 46
Predictive modeling of dynamic fracture growth in brittle materials with machine learning
journal, June 2018
- Moore, Bryan A.; Rougier, Esteban; O’Malley, Daniel
- Computational Materials Science, Vol. 148
A review of deep learning in the study of materials degradation
journal, November 2018
- Nash, Will; Drummond, Tom; Birbilis, Nick
- npj Materials Degradation, Vol. 2, Issue 1
Virtual assembly and residual stress analysis for the composite fuselage assembly process
journal, July 2019
- Wen, Yuchen; Yue, Xiaowei; Hunt, Jeffrey H.
- Journal of Manufacturing Systems, Vol. 52
Tensor decomposition to compress convolutional layers in deep learning
journal, April 2021
- Wang, Yinan; Guo, Weihong “Grace”; Yue, Xiaowei
- IISE Transactions
On the Properties of Neural Machine Translation: Encoder–Decoder Approaches
conference, January 2014
- Cho, Kyunghyun; van Merrienboer, Bart; Bahdanau, Dzmitry
- Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation
Deep Multi-View Spatial-Temporal Network for Taxi Demand Prediction
journal, April 2018
- Yao, Huaxiu; Wu, Fei; Ke, Jintao
- Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 32, Issue 1
Reduced-order modeling through machine learning and graph-theoretic approaches for brittle fracture applications
journal, February 2019
- Hunter, Abigail; Moore, Bryan A.; Mudunuru, Maruti
- Computational Materials Science, Vol. 157
CityFlow: A Multi-Agent Reinforcement Learning Environment for Large Scale City Traffic Scenario
conference, May 2019
- Zhang, Huichu; Feng, Siyuan; Liu, Chang
- WWW '19: The Web Conference, The World Wide Web Conference
IntelliLight: A Reinforcement Learning Approach for Intelligent Traffic Light Control
conference, July 2018
- Wei, Hua; Zheng, Guanjie; Yao, Huaxiu
- KDD '18: The 24th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, Proceedings of the 24th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining
Numerical Approach: Finite Element Analysis
book, January 2011
- Ashcroft, Ian A.; Mubashar, Aamir; da Silva, Lucas F. M.
- Handbook of Adhesion Technology, p. 629-660
CDC: Convolutional-De-Convolutional Networks for Precise Temporal Action Localization in Untrimmed Videos
conference, July 2017
- Shou, Zheng; Chan, Jonathan; Zareian, Alireza
- 2017 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)
Using machine learning and a data-driven approach to identify the small fatigue crack driving force in polycrystalline materials
journal, July 2018
- Rovinelli, Andrea; Sangid, Michael D.; Proudhon, Henry
- npj Computational Materials, Vol. 4, Issue 1
A generalized anisotropic deformation formulation for geomaterials
journal, October 2015
- Lei, Z.; Rougier, Esteban; Knight, E. E.
- Computational Particle Mechanics, Vol. 3, Issue 2
Accelerating high-strain continuum-scale brittle fracture simulations with machine learning
journal, January 2021
- Fernández-Godino, M. Giselle; Panda, Nishant; O’Malley, Daniel
- Computational Materials Science, Vol. 186
Spatial Temporal Graph Convolutional Networks for Skeleton-Based Action Recognition
journal, April 2018
- Yan, Sijie; Xiong, Yuanjun; Lin, Dahua
- Proceedings of the AAAI Conference on Artificial Intelligence, Vol. 32, Issue 1
Surrogate Model-Based Control Considering Uncertainties for Composite Fuselage Assembly
journal, February 2018
- Yue, Xiaowei; Wen, Yuchen; Hunt, Jeffrey H.
- Journal of Manufacturing Science and Engineering, Vol. 140, Issue 4
Superiorities of support vector machine in fracture prediction and gassiness evaluation
journal, October 2008
- Shi, Guang-ren
- Petroleum Exploration and Development, Vol. 35, Issue 5
Finite element analysis of residual stresses and thin plate distortion after face milling
conference, January 2015
- Ma, Y.; Liu, S.; Feng, P. F.
- 2015 12th International Bhurban Conference on Applied Sciences and Technology (IBCAST)
Box-Jenkins Seasonal Forecasting: Problems in a Case-Study
journal, January 1973
- Chatfield, C.; Prothero, D. L.
- Journal of the Royal Statistical Society. Series A (General), Vol. 136, Issue 3
Strain rate concentration and dynamic stress concentration for double-edge-notched specimens subjected to high-speed tensile loads: STRAIN RATE CONCENTRATION AND DYNAMIC STRESS CONCENTRATION
journal, July 2014
- Noda, N. -A.; Ohtsuka, H.; Zheng, H.
- Fatigue & Fracture of Engineering Materials & Structures, Vol. 38, Issue 1
Brittle-material failures as indicators of stress-concentration factors: Test program leads to development of method for determining stress-concentration factors in components of complicated shapes. Plaster-silica mortar is used for test models
journal, March 1962
- Durelli, A. J.; Jacobson, R. H.
- Experimental Mechanics, Vol. 2, Issue 3
Long Short-Term Memory
journal, November 1997
- Hochreiter, Sepp; Schmidhuber, Jürgen
- Neural Computation, Vol. 9, Issue 8
Learning to fail: Predicting fracture evolution in brittle material models using recurrent graph convolutional neural networks
journal, May 2019
- Schwarzer, Max; Rogan, Bryce; Ruan, Yadong
- Computational Materials Science, Vol. 162
Stress Field Prediction in Cantilevered Structures Using Convolutional Neural Networks
journal, September 2019
- Nie, Zhenguo; Jiang, Haoliang; Kara, Levent Burak
- Journal of Computing and Information Science in Engineering, Vol. 20, Issue 1
Skeleton-Based Action Recognition with Spatial Reasoning and Temporal Stack Learning
book, January 2018
- Si, Chenyang; Jing, Ya; Wang, Wei
- Computer Vision – ECCV 2018: 15th European Conference, Munich, Germany, September 8-14, 2018, Proceedings, Part I, p. 106-121
Optimal Integration of Supervised Tensor Decomposition and Ensemble Learning for In Situ Quality Evaluation in Friction Stir Blind Riveting
journal, January 2021
- Gao, Zhe; Guo, Weihong; Yue, Xiaowei
- IEEE Transactions on Automation Science and Engineering, Vol. 18, Issue 1
Tensor Mixed Effects Model With Application to Nanomanufacturing Inspection
journal, June 2019
- Yue, Xiaowei; Park, Jin Gyu; Liang, Zhiyong
- Technometrics, Vol. 62, Issue 1