Neural network ensemble-based sensitivity analysis in structural engineering: Comparison of selected methods and the influence of statistical correlation

Abstract Surrogate model-based sensitivity analysis, especially framed by neural network ensemble (NNE), is an attractive but unresolved issue in structural reliability assessment. In this paper, differing from existing studies, an overview and assessment of typical methods for surrogate model-based parameter sensitivity analysis, namely the input perturbation method, the local analysis of variance, the connection weight method, the non-parametric Spearman rank-order correlation method, and the Sobol indices method, are performed and demonstrated on three illustrative cases of increasing complexity: a simple theoretical instance, an engineering case of midspan deflection of a simply-supported beam, and a real-world practical application of shear failing in a precast concrete girder. Through comprehensive comparisons, several findings are obtained as follows: (i) the NNE is testified a superior surrogate model for sensitivity analysis to a single artificial neural network; (ii) robustness and accuracy of an NNE in sensitivity analysis are demonstrated; (iii) the properties of these parameter sensitivity analysis methods are fully clarified with distinguished merits and limitations; (iv) mechanism of local- and global- sensitivity analysis methods is revealed; and (v) the strategy for sensitivity analysis of correlated descriptive variables are elaborated to address the impact of correlation among random variables in engineering systems.

[1]  Oral Büyüköztürk,et al.  Deep Learning‐Based Crack Damage Detection Using Convolutional Neural Networks , 2017, Comput. Aided Civ. Infrastructure Eng..

[2]  L. Faravelli.  Response‐Surface Approach for Reliability Analysis , 1989 .

[3]  Drahomír Novák,et al.  Polynomial chaos expansion for surrogate modelling: Theory and software , 2018, Beton- und Stahlbetonbau.

[4]  C. Bucher,et al.  A fast and efficient response surface approach for structural reliability problems , 1990 .

[5]  David Lehký,et al.  Neural network ensemble-based parameter sensitivity analysis in civil engineering systems , 2017, Neural Computing and Applications.

[6]  Bruno Sudret,et al.  Global sensitivity analysis using polynomial chaos expansions , 2008, Reliab. Eng. Syst. Saf..

[7]  F. Daneshjoo,et al.  Probabilistic sensitivity analysis of multi-span highway bridges , 2015 .

[8]  Wenxi Lu,et al.  Sobol′ sensitivity analysis of NAPL-contaminated aquifer remediation process based on multiple surrogates , 2014, Comput. Geosci..

[9]  Pizhong Qiao,et al.  Neural network committee-based sensitivity analysis strategy for geotechnical engineering problems , 2008, Neural Computing and Applications.

[10]  Pan Wang,et al.  Copula-based decomposition approach for the derivative-based sensitivity of variance contributions with dependent variables , 2018, Reliab. Eng. Syst. Saf..

[11]  R. Ghanem,et al.  Stochastic Finite Elements: A Spectral Approach , 1990 .

[12]  Maosen Cao,et al.  Structural health monitoring through nonlinear frequency-based approaches for conservative vibratory systems , 2020 .

[13]  Ilya M. Sobol,et al.  Sensitivity Estimates for Nonlinear Mathematical Models , 1993 .

[14]  F. O. Hoffman,et al.  An Examination of Response-Surface Methodologies for Uncertainty Analysis in Assessment Models , 1985 .

[15]  Julian D. Olden,et al.  Illuminating the “black box”: a randomization approach for understanding variable contributions in artificial neural networks , 2002 .

[16]  Thomas D. Gauthier,et al.  Detecting Trends Using Spearman's Rank Correlation Coefficient , 2001 .

[17]  Bruce R. Ellingwood,et al.  A new look at the response surface approach for reliability analysis , 1993 .

[18]  Marijana Hadzima-Nyarko,et al.  A neural network based modelling and sensitivity analysis of damage ratio coefficient , 2011, Expert Syst. Appl..

[19]  Niels Olhoff,et al.  Sensitivity analysis of structures , 1997 .

[20]  David Lehký,et al.  Reliability calculation of time-consuming problems using a small-sample artificial neural network-based response surface method , 2017, Neural Computing and Applications.

[21]  Juan de Oña,et al.  Extracting the contribution of independent variables in neural network models: a new approach to handle instability , 2014, Neural Computing and Applications.

[22]  Herschel Rabitz,et al.  Sixth International Conference on Sensitivity Analysis of Model Output Global Sensitivity Analysis for Systems with Independent and / or Correlated Inputs , 2013 .

[23]  Sarath Das,et al.  Classification of slopes and prediction of factor of safety using differential evolution neural networks , 2011 .

[24]  Andrea Saltelli,et al.  Sensitivity Analysis for Importance Assessment , 2002, Risk analysis : an official publication of the Society for Risk Analysis.

[25]  Nilanjan Dey,et al.  Particle swarm optimization trained neural network for structural failure prediction of multistoried RC buildings , 2016, Neural Computing and Applications.

[26]  Ashu Jain,et al.  Modeling and Analysis of Concrete Slump Using Artificial Neural Networks , 2008 .

[27]  David Lehký,et al.  A comparison of sensitivity analyses for selected prestressed concrete structures , 2019 .

[28]  Dave Winkler,et al.  Bayesian Regularization of Neural Networks , 2009, Artificial Neural Networks.

[29]  Heng Li,et al.  Towards reliability evaluation involving correlated multivariates under incomplete probability information: A reconstructed joint probability distribution for isoprobabilistic transformation , 2017 .

[30]  Victor Picheny,et al.  Adaptive Designs of Experiments for Accurate Approximation of a Target Region , 2010 .

[31]  I. Sobola,et al.  Global sensitivity indices for nonlinear mathematical models and their Monte Carlo estimates , 2001 .

[32]  Maurice Lemaire,et al.  Assessing small failure probabilities by combined subset simulation and Support Vector Machines , 2011 .

[33]  J. Xavier,et al.  Determination of mode I dynamic fracture toughness of IM7-8552 composites by digital image correlation and machine learning , 2019, Composite Structures.

[34]  Bertrand Iooss,et al.  Advanced Methodology for Uncertainty Propagation in Computer Experiments with Large Number of Inputs , 2018, Nuclear Technology.

[35]  Siddhartha Kumar Khaitan,et al.  Deep Convolutional Neural Networks with transfer learning for computer vision-based data-driven pavement distress detection , 2017 .

[36]  Hao Yu,et al.  Levenberg—Marquardt Training , 2011 .

[37]  Bruno Sudret,et al.  Adaptive sparse polynomial chaos expansion based on least angle regression , 2011, J. Comput. Phys..

[38]  R. Pires dos Santos,et al.  Identifying the relative importance of predictive variables in artificial neural networks based on data produced through a discrete event simulation of a manufacturing environment , 2018, International Journal of Modelling and Simulation.

[39]  Jack P. C. Kleijnen,et al.  Sensitivity analysis of simulation models: an overview , 2010 .

[40]  Liborio Cavaleri,et al.  Prediction of the Fundamental Period of Infilled RC Frame Structures Using Artificial Neural Networks , 2015, Comput. Intell. Neurosci..

[41]  A. Kiureghian,et al.  STRUCTURAL RELIABILITY UNDER INCOMPLETE PROBABILITY INFORMATION , 1986 .

[42]  Murat Kayri,et al.  Predictive Abilities of Bayesian Regularization and Levenberg–Marquardt Algorithms in Artificial Neural Networks: A Comparative Empirical Study on Social Data , 2016 .

[43]  Fabrice Gamboa,et al.  Local Polynomial Estimation for Sensitivity Analysis on Models With Correlated Inputs , 2008, Technometrics.

[44]  R. Lebrun,et al.  An innovating analysis of the Nataf transformation from the copula viewpoint , 2009 .

[45]  Jurgita Antucheviciene,et al.  Solving Civil Engineering Problems by Means of Fuzzy and Stochastic MCDM Methods: Current State and Future Research , 2015 .

[46]  Richard J. Beckman,et al.  A Comparison of Three Methods for Selecting Values of Input Variables in the Analysis of Output From a Computer Code , 2000, Technometrics.

[47]  Reza Noroozi,et al.  Determining the Relative Importance of Parameters Affecting Concrete Pavement Thickness , 2015 .

[48]  Amira S. Ashour,et al.  Neural-based prediction of structural failure of multistoried RC buildings , 2016 .

[49]  M. D. McKay,et al.  A comparison of three methods for selecting values of input variables in the analysis of output from a computer code , 2000 .

[50]  Aminul Islam Laskar,et al.  Prediction of unconfined compressive strength of geopolymer stabilized clayey soil using Artificial Neural Network , 2015 .

[51]  Mehdi Nikoo,et al.  Multiple linear regression, artificial neural network, and fuzzy logic prediction of 28 days compressive strength of concrete , 2017, Frontiers of Structural and Civil Engineering.

[52]  Michele Scardi,et al.  Developing an empirical model of phytoplankton primary production: a neural network case study , 1999 .

[53]  M. Rosenblatt.  Remarks on a Multivariate Transformation , 1952 .

[54]  Abbas Majdi,et al.  Genetic programming approach for estimating the deformation modulus of rock mass using sensitivity analysis by neural network , 2010 .

[55]  M. Gevrey,et al.  Review and comparison of methods to study the contribution of variables in artificial neural network models , 2003 .

[56]  M. Eldred,et al.  Efficient Global Reliability Analysis for Nonlinear Implicit Performance Functions , 2008 .

[57]  Paul Ziehl,et al.  Bayesian demand model based seismic vulnerability assessment of a concrete girder bridge , 2020 .

[58]  Wei Tang,et al.  Ensembling neural networks: Many could be better than all , 2002, Artif. Intell..

[59]  Emanuele Borgonovo,et al.  Sensitivity analysis: A review of recent advances , 2016, Eur. J. Oper. Res..

[60]  Alfred Strauss,et al.  Correlation among selected fracture‐mechanical parameters of concrete obtained from experiments and inverse analyses , 2016 .