A new reliability-based data-driven approach for noisy experimental data with physical constraints

被引:34
作者
Ayensa-Jimenez, Jacobo [1 ,2 ,3 ]
Doweidar, Mohamed H. [1 ,2 ,3 ]
Sanz-Herrera, Jose A. [4 ]
Doblare, Manuel [1 ,2 ,3 ]
机构
[1] Univ Zaragoza, Mech Engn Dept, Zaragoza, Spain
[2] Univ Zaragoza, Aragon Inst Engn Res I3A, Zaragoza, Spain
[3] CIBER BBN, Zaragoza, Spain
[4] Univ Seville, Sch Engn ETSI, Seville, Spain
关键词
Data-driven models; Reliability; Mahalanobis distance; KARHUNEN-LOEVE DECOMPOSITION; DIMENSIONALITY REDUCTION; HOMOGENIZATION ANALYSIS; HETEROGENEOUS MATERIALS; MODEL-REDUCTION; ORDER;
D O I
10.1016/j.cma.2017.08.027
中图分类号
T [工业技术];
学科分类号
08 ;
摘要
Data Science has burst into simulation-based engineering sciences with an impressive impulse. However, data are never uncertainty-free and a suitable approach is needed to face data measurement errors and their intrinsic randomness in problems with well-established physical constraints. As in previous works, this problem is here faced by hybridizing a standard mathematical modeling approach with a new data-driven solver accounting for the phenomenological part of the problem, with the aim of finding a solution point, satisfying some constraints, that minimizes a distance to a given data-set. However, unlike such works that are established in a deterministic framework, we use the Mahalanobis distance in order to incorporate statistical second order uncertainty of data in computations, i.e. variance and correlation. We develop the underlying stochastic theoretical framework and establish the fundamental mathematical and statistical properties. The performance of the resulting reliability-based data-driven procedure is evaluated in a simple but illustrative unidimensional problem as well as in a more realistic solution of a 3D structural problem with a material with intrinsically random constitutive behavior as concrete. The results show, in comparison with other data-driven solvers, better convergence, higher accuracy, clearer interpretation, and major flexibility besides the relevance of allowing uncertainty management with low computational demand. (C) 2017 Elsevier B.V. All rights reserved.
引用
收藏
页码:752 / 774
页数:23
相关论文
共 55 条
[1]   Principal component analysis [J].
Abdi, Herve ;
Williams, Lynne J. .
WILEY INTERDISCIPLINARY REVIEWS-COMPUTATIONAL STATISTICS, 2010, 2 (04) :433-459
[2]  
[Anonymous], 1997, MACH LEARN
[3]   Laplacian eigenmaps for dimensionality reduction and data representation [J].
Belkin, M ;
Niyogi, P .
NEURAL COMPUTATION, 2003, 15 (06) :1373-1396
[4]   A Survey of Projection-Based Model Reduction Methods for Parametric Dynamical Systems [J].
Benner, Peter ;
Gugercin, Serkan ;
Willcox, Karen .
SIAM REVIEW, 2015, 57 (04) :483-531
[5]   Micro-mechanical approach and algorithm for the study of damage appearance in elastomer composites [J].
Brieu, M ;
Devries, F .
COMPOSITE STRUCTURES, 1999, 46 (04) :309-319
[6]  
Cheng SF, 2009, IEEE INT CON AUTO SC, P102, DOI 10.1109/COASE.2009.5234098
[7]   A Short Review on Model Order Reduction Based on Proper Generalized Decomposition [J].
Chinesta, Francisco ;
Ladeveze, Pierre ;
Cueto, Elias .
ARCHIVES OF COMPUTATIONAL METHODS IN ENGINEERING, 2011, 18 (04) :395-404
[8]  
Darema F, 2004, LECT NOTES COMPUT SC, V3038, P662
[9]   The Mahalanobis distance [J].
De Maesschalck, R ;
Jouan-Rimbaud, D ;
Massart, DL .
CHEMOMETRICS AND INTELLIGENT LABORATORY SYSTEMS, 2000, 50 (01) :1-18
[10]   Modeling diffusion in random heterogeneous media: Data-driven models, stochastic collocation and the variational multiscale method [J].
Ganapathysubramanian, Baskar ;
Zabaras, Nicholas .
JOURNAL OF COMPUTATIONAL PHYSICS, 2007, 226 (01) :326-353