Structural reliability analysis using a hybrid HDMR-ANN method

Abstract

A new hybrid method is proposed to estimate the failure probability of a structure subject to random parameters. The high dimensional model representation (HDMR) combined with artificial neural network (ANN) is used to approximate implicit limit state functions in structural reliability analysis. HDMR facilitates the lower dimensional approximation of the original limit states function. For evaluating the failure probability, a first-order HDMR approximation is constructed by deploying sampling points along each random variable axis and hence obtaining the structural responses. To reduce the computational effort of the evaluation of limit state function, an ANN surrogate is trained based on the sampling points from HDMR. The component of the approximated function in HDMR can be regarded as the input of the ANN and the response of limit state function can be regarded as the target for training an ANN surrogate. This trained ANN surrogate is used to obtain structural outputs instead of directly calling the numerical model of a structure. After generating the ANN surrogate, Monte Carlo simulation (MCS) is performed to obtain the failure probability, based on the trained ANN surrogate. Three numerical examples are used to illustrate the accuracy and efficiency of the proposed method.

Keywords

high dimensional model representation / structural reliability / artificial neural network / failure probability

References

[1] DitlevsenO, MadsenH OStructural reliability methods [M], 1996, Chichester, John Wiley & Sons

[2] BucherCComputational analysis of randomness in structural mechanics [M], 2009, London, CRC Press

[3] LiuP L, DerkiureghianA. Finite element reliability of geometrically nonlinear uncertain structures [J]. Journal of Engineering Mechanics, 1988, 117(8): 164-168

[4] PenmetsaR C, GrandhiR V. Adaptation of fast Fourier transformations to estimate structural failure probability [J]. Finite Elements in Analysis and Design, 2003, 39(5): 473-485

[5] WuY T, TorngT RA fast convolution procedure for probabilistic engineering analysis [C]//Proceeding of the First International Symposium on Uncertainty Modeling and Analysis, 1990, College park, Maryland, USA, IEEE: 670675

[6] ZhaoY G, OnoT. Moment methods for structural reliability [J]. Structural Safety, 2001, 23(1): 47-75

[7] LuZ H, ZhaoY G, AngA H S. Estimation of load and resistance factors based on the fourth moment method [J]. Structural Engineering and Mechanics, 2010, 36(1): 19-36

[8] ZhaoY G, LuZ H. Applicable range of the fourth-moment method for structural reliability [J]. Journal of Asian Architecture & Building Engineering, 2007, 6(1): 151-158

[9] RubinsteinR Y, KrosesD PSimulation and the Monte Carlo method [M], 2017, New Jersey, John Wiley & Sons

[10] SobolI M. Theorems and examples on high dimensional model representations [J]. Reliability Engineering and System Safety, 2003, 79(2): 187-193

[11] SobolI M. Sensitivity analysis for non-linear mathematical models [J]. Mathematical Modeling & Computational Experiment, 1993, 1(4): 407-414

[12] RaoB N, ChowdhuryR. Probabilistic analysis using high dimensional model and fast Fourier transform [J]. International Journal for Computational Methods in Engineering Science and Mechanics, 2008, 9(6): 342-357

[13] ChowdharyR, RaoB N, PrasadA M. High dimensional model representation for structural reliability analysis [J]. Communications in Numerical Methods in Engineering, 2009, 25(4): 301-337

[14] RaoB N, ChowdharyR, SinghC R, KushwahaH SStructural reliability analysis of AHWR and PHWR inner containment using high dimensional model representation [C]//Proceedings of 2nd International Conference on Reliability, Safety & Hazard (ICRESH-2010), 2010, Mumbai, India, IEEE: 7886

[15] WangH, TangL, LiG Y. Adaptive MLS-HDMR metamodeling techniques for high dimensional problems [J]. Expert Systems with Applications, 2011, 38(11): 14117-14126

[16] ShanS, WangGaryDevelopment of adaptive RBF-HDMR model for approximating high dimensional problems [C]//Proceedings of the ASME 2009 International Design Engineering Technical Conferences & Computers and Information in Engineering Conference, 2009, San Diego, California, USA, ASME

[17] ZhouC C, LuZ Z, LiG J. A new algorithm for variance based importance measures and importance kernel sensitivity [J]. Proceedings of the Institution of Mechanical Engineers Part O: Journal of Risk and Reliability, 2013, 227(1): 16-27

[18] LiG, RosenthalC, RabitzH. High dimensional model representation [J]. Journal of Physical Chemistry, 2001, 105(55): 7765-7777

[19] RabitzH, AlisO F. General foundations of high-dimensional model representations [J]. Journal of Mathematical Chemistry, 1999, 25(2): 197-233

[20] CHOWDHURYR, RaoB N, PrasadA M. High dimensional model representation for piece wise continuous function approximation [J]. Communications in Numerical Methods in Engineering, 2008, 24(12): 1587-1609

[21] MacLeodCAn introduction to practical neural networks and genetic algorithms for engineers and scientists [M], 2010, Scotland, The Robert Gordon University

[22] LevenbergK. A method for the solution of certain non-linear problems in least squares [J]. Quarterly of Applied Mathematics, 1944, 2(2): 164-168

[23] MarquardtD W. An algorithm for least square estimation for non-linear parameters [J]. Journal of the Society for Industrial and Applied Mathematics, 1963, 11(2): 431-441

[24] BroomheadD S, LoweD. Multivariable functional interpolation and adaptive networks [J]. Complex Systems, 1988, 2(1): 321-355

[25] MoodyJ, DarkenC JFast learning in networks of locally tuned processing units [J], 1989, 1(2): 281-294

[26] BishopCMNeural networks for pattern recognition [M], 1995, Oxford, Clarendon Press

[27] HaykinSNeural Networks: A comprehensive foundation [M], 1999, New Jersey, Prentice-Hall

[28] KimS H, NaS W. Response surface method using vector projected sampling points [J]. Structural Safety, 1997, 19(1): 3-19