Structural health monitoring (SHM) is the process of detecting damage in structures. Typically, the damage-sensitive feature identification is based on fitting some models to the measured system response data. The parameters of these models or the prediction errors associated with these models then become the desired damage-sensitive features. Real-world structures usually exhibit significant nonlinear behaviors. Data-driven nonlinear system identification (NSI) is essential to obtain accurate models of structural dynamics for reliable damage detection. However, on the one hand, the data acquisition of real-world structures for damage detection is usually difficult or scarce. On the other hand, NSI generally needs to solve a nonlinear optimization problem that requires more computational time for convergence than linear system identification. Data efficiency and computational efficiency become two critical challenges in data-driven NSI. To address these challenges, this work presents a proof-of-principle study on developing a meta learning-based method for efficient data-driven NSI. Specifically, the meta learning algorithm leverages a previously collected database of similar but different systems to learn the meta-knowledge about how to identify a new system, enabling a fast identification/modeling with limited data. Numerical simulations are performed to validate the method on fundamental nonlinear dynamical systems widely seen in civil engineering. It is observed that the presented meta learning-based method outperforms the conventional method in terms of both data and computational efficiency. Limitations of this work are further discussed.
AbadiMAgarwalABarhamP, et al. (2016) Tensorflow: large-scale machine learning on heterogeneous distributed systems. CoRR ArXiv Preprint arXiv:1603.04467. URL: https://arxiv.org/abs/1603.04467.
2.
AndrychowiczMDenilMColmenarejoSG, et al. (2016) Learning to learn by gradient descent by gradient descent. In: Proceedings of the 30th international conference on neural information processing systems, NIPS’16, Red Hook, NY, December 2016, 3988–3996. Curran Associates Inc.
3.
AyalaHVHCoelhoLS (2016) Cascaded evolutionary algorithm for nonlinear system identification based on correlation functions and radial basis functions neural networks. Mechanical Systems and Signal Processing68–69: 378–393. DOI: 10.1016/j.ymssp.2015.05.022. URL: https://www.sciencedirect.com/science/article/pii/S0888327015002757.
4.
BaekJLeeDBHwangSJ (2020) Learning to extrapolate knowledge: transductive few-shot out-of-graph link prediction. In: LarochelleHRanzatoMHadsellR, et al. (eds) Advances in neural information processing systems 33: annual conference on neural information processing systems 2020, NeurIPS 2020, Virtual Conference, 6–12 December 2020. URL: https://proceedings.neurips.cc/paper/2020/hash/0663a4ddceacb40b095eda264a85f15c-Abstract.html.
5.
BengioSBengioYCloutierJ, et al. (2013) On the optimization of a synaptic learning rule. In: Optimality in Biological and Artificial Networks?London, UK: Routledge, 281–303.
6.
dos Santos CoelhoLPessôaMW (2009) Nonlinear identification using a b-spline neural network and chaotic immune approaches. Mechanical Systems and Signal Processing23(8): 2418–2434. DOI: 10.1016/j.ymssp.2009.01.013. URL: https://www.sciencedirect.com/science/article/pii/S0888327009000090.
7.
EhsanFScanlanRH (1990) Vortex-induced vibrations of flexible bridges. Journal of Engineering Mechanics116(6): 1392–1411. DOI: 10.1061/(ASCE)0733-9399(1990)116:6(1392).
8.
FarrarCRWordenK (2007) An introduction to structural health monitoring. Philosophical Transactions of the Royal Society A: Mathematical, Physical & Engineering Sciences365(1851): 303–315. DOI: 10.1098/rsta.2006.1928.
9.
FinnCAbbeelPLevineS (2017a) Model-agnostic meta-learning for fast adaptation of deep networks. In: 34th International Conference on Machine Learning, Sydney, NSW, 06–11 August 2017, 1856–1868. ICML.
10.
FinnCYuTZhangT, et al. (2017b) One-shot visual imitation learning via meta-learning. In: LevineSVanhouckeVGoldbergK (eds) Proceedings of the 1st Annual Conference on Robot Learning, Proceedings of Machine Learning Research, Mountain View, CA, 13–15 November 2017, 357–368. PMLR. URL: https://proceedings.mlr.press/v78/finn17a.html.
11.
HarrisonJSharmaAPavoneM (2020) Meta-learning priors for efficient online bayesian regression. In: MoralesMTapiaLSánchez-AnteG, et al. (eds) Algorithmic Foundations of Robotics XIII. Cham, Switzerland: Springer International Publishing, 318–337.
12.
HornikKStinchcombeMWhiteH (1990) Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networks. Neural Networks3(5): 551–560. DOI: 10.1016/0893-6080(90)90005-6.
13.
HuangYBeckJLLiH (2017) Bayesian system identification based on hierarchical sparse bayesian learning and gibbs sampling with application to structural damage assessment. Computer Methods in Applied Mechanics and Engineering318: 382–411.
14.
JinXChengPChenWL, et al. (2018) Prediction model of velocity field around circular cylinder over various Reynolds numbers by fusion convolutional neural networks based on pressure on the cylinder. Physics of Fluids30(4): 047105. DOI: 10.1063/1.5024595.
15.
LakeBMSalakhutdinovRTenenbaumJB (2015) Human-level concept learning through probabilistic program induction. Science350(6266): 1332–1338. DOI: 10.1126/science.aab3050. URL: https://science.sciencemag.org/.
16.
LiDYangYSongYZ, et al. (2018a) Learning to generalize: meta-learning for domain generalization. In: In: 32nd AAAI Conference on Artificial Intelligence, New Orleans, LA, February 2018, 3490–3497. URL: https://arxiv.org/abs/1710.03463.
17.
LiSLaimaSLiH (2018b) Data-driven modeling of vortex-induced vibration of a long-span suspension bridge using decision tree learning and support vector regression. Journal of Wind Engineering and Industrial Aerodynamics172: 196–211. DOI: 10.1016/j.jweia.2017.10.022. URL: https://linkinghub.elsevier.com/retrieve/pii/S0167610517301599.
18.
LiWLaimaSJinX, et al. (2020) A novel long short-term memory neural-network-based self-excited force model of limit cycle oscillations of nonlinear flutter for various aerodynamic configurations. Nonlinear Dynamics100(3): 2071–2087. DOI: 10.1007/s11071-020-05631-5. URL: https://link.springer.com/article/10.1007/s11071-020-05631-5.
19.
LiSLiSLaimaS, et al. (2021) Data-driven modeling of bridge buffeting in the time domain using long short-term memory network based on structural health monitoring. Structural Control and Health Monitoring28(8): e2772. DOI: 10.1002/stc.2772. URL: https://onlinelibrary.wiley.com/doi/abs/10.1002/stc.2772.
20.
MeteredHBonelloPOyadijiS (2010) The experimental identification of magnetorheological dampers and evaluation of their controllers. Mechanical Systems and Signal Processing24(4): 976–994. DOI: 10.1016/j.ymssp.2009.09.005. URL: https://www.sciencedirect.com/science/article/pii/S0888327009002659.
21.
PanSDuraisamyK (2018) Long-time predictive modeling of nonlinear dynamical systems using neural networks. Complexity2018: 4801012. DOI: 10.1155/2018/4801012.
22.
PathakJHuntBGirvanM, et al. (2018) Model-free prediction of large spatiotemporally chaotic systems from data: a reservoir computing approach. Physical Review Letters120: 024102. DOI: 10.1103/PhysRevLett.120.024102.
23.
RaviSLarochelleH (2017) Optimization as a model for few-shot learning In: 5th International Conference on Learning Representations, ICLR 2017, Toulon, France, 24–26 April 2017. URL: https://openreview.net/forum?id=rJY0-Kcll.
24.
RoussVCharonWCirrincioneG (2009) Neural model of the dynamic behaviour of a non-linear mechanical system. Mechanical Systems and Signal Processing23(4): 1145–1159. DOI: 10.1016/j.ymssp.2008.09.004. URL: https://www.sciencedirect.com/science/article/pii/S0888327008002215.
25.
SahooDPhamQLuJ, et al. (2018) Online Deep Learning: Learning Deep Neural Networks on the Fly. San Francisco, CA: IJCAI International Joint Conference On Artificial Intelligence, 2660–2666.
26.
TengQZhangL (2019) Data driven nonlinear dynamical systems identification using multi-step CLDNN. AIP Advances9(8): 085311. DOI: 10.1063/1.5100558.
WittersMSweversJ (2010) Black-box model identification for a continuously variable, electro-hydraulic semi-active damper. Mechanical Systems and Signal Processing24(1): 4–18. DOI: 10.1016/j.ymssp.2009.03.013. URL: https://www.sciencedirect.com/science/article/pii/S0888327009001241.
29.
WordenKWongCParlitzU, et al. (2007) Identification of pre-sliding and sliding friction dynamics: grey box and black-box models. Mechanical Systems and Signal Processing21(1): 514–534. DOI: 10.1016/j.ymssp.2005.09.004. URL: https://www.sciencedirect.com/science/article/pii/S0888327005001408.
30.
WordenKStaszewskiWJHensmanJJ (2011) Natural computing for mechanical systems research: a tutorial overview. Mechanical Systems and Signal Processing25(1): 4–111. DOI: 10.1016/j.ymssp.2010.07.013. URL: https://www.sciencedirect.com/science/article/pii/S0888327010002499.
