Application of deep learning neural network to identify collision load conditions based on permanent plastic deformation of shell structures
- 21 Downloads
In this work, we have developed a novel deep learning inverse solution or identification method to determine and identify the impact load conditions of shell structures based on their final state of damage or inelastic deformation. This artificial intelligence approach offers a practical solution to solve the inverse problem of engineering failure analysis based on final material and structure damage state and permanent plastic deformation. More precisely, the machine learning inverse problem solver may provide a practical solution to characterize failure load parameters and conditions based on the final permanent plastic deformation distribution of the shell structure that is under examination. In this work, we have demonstrated that the proposed deep learning method can accurately identify a “practically unique” static loading condition as well as the impact dynamic loading condition for a hemispherical shell structure based the permanent plastic deformation after the impact event as the forensic signatures. The data-driven based method developed in this work may provide a powerful tool for forensically diagnosing, determining, and identifying damage loading conditions for engineering structures in various accidental failure events, such as car crashes, pressure vessel failure, or thin-walled infrastructure structure collapses. The machine learning inverse problem solver developed here in this work may have potential impacts on general forensic material and structure failure analysis based on final permanent plastic deformations.
KeywordsArtificial intelligence (AI ) Deep learning Forensic materials engineering Neural network Plastic deformation Structural failure analysis
Mr. G. Chen is supported by a graduate fellowship from Chinese Scholar Council (CSC), and this support is gratefully appreciated.
- 1.Ren S, Chen G, Li T, Chen Q, Li S (2018) A deep learning-based computational algorithm for identifying damage load condition: an artificial intelligence inverse problem solution for failure analysis. Comput. Model. Eng. Sci. 117(3):287–307Google Scholar
- 8.Mena J (2016) Machine learning forensics for law enforcement, security, and intelligence. Auerbach Publications, Boca RatonGoogle Scholar
- 15.Principal component analysis (2019). https://en.wikipedia.org/wiki/Principal_component_analysis. Accessed 28 Mar 2019
- 16.Shlens J (2014) A tutorial on principal component analysis. arXiv preprint arXiv:1404.1100
- 17.Lu X, Giovanis DG, Yvonnet J, Papadopoulos V, Detrez F, Bai J (2018) A data-driven computational homogenization method based on neural networks for the nonlinear anisotropic electrical response of graphene/polymer nanocomposites. Comput. Mech. https://doi.org/10.1007/s00466-018-1643-0 Google Scholar
- 28.Nair V, Hinton GE (2010) Rectified linear units improve restricted Boltzmann machines. In: Proceedings of the 27th international conference on machine learning (ICML-10), pp 807–814Google Scholar
- 29.Kingma DP, Ba J (2014) Adam: a method for stochastic optimization. arXiv preprint arXiv:1412.6980
- 30.Hibbitt H, Karlsson B, Sorensen P (2011) Abaqus analysis user’s manual version 6.10. Dassault Systmes Simulia Corp., ProvidenceGoogle Scholar
- 33.TensorFlow. Tf.losses.Reduction – TensorFlow. https://www.tensorflow.org/api_docs/python/tf/losses/Reduction. Accessed 27 Mar 2019