Generalized F‐discrepancy‐based point selection strategy for dependent random variables in uncertainty quantification of nonlinear structures

Abstract

Summary In the performance evaluation of structures under disastrous actions, for example, earthquakes, it is important to take into account the randomness of structural parameters. Generally, these random parameters are treated either as independent or perfectly dependent, but practically they are partly dependent. This article aims at developing a point selection strategy for uncertainty quantification of nonlinear structures involving probabilistically dependent random parameters characterized by copula function. For this purpose, the point selection strategy for structures involving independent basic variables is first revisited. As an improvement, a generalized F‐discrepancy diminishing oriented iterative screening algorithm is proposed. Then, combining with the conditional sampling method, a conditional point set rearrangement method and a conditional iterative screening‐rearrangement method are proposed for probabilistically dependent variables. These new point selection strategies are readily incorporated into the probability density evolution method for uncertainty quantification of nonlinear structures involving dependent random parameters, which is characterized by copula function. The proposed methods are illustrated by two examples including a shear frame with hysteretic restoring forces and a reinforced concrete frame structure with the damage constitutive model of concrete, where the material parameters are probabilistically dependent. The results demonstrate the effectiveness of the proposed method. Problems to be studied are discussed.

Probabilistic screening and behavior of solar cells under Gaussian parametric uncertainty using polynomial chaos representation model

“…Uncertainty quantification through probabilistic assessment is a statistical tool to understand a complex engineering process under extrinsic or intrinsic variations [20,21]. It is important a decision-maker to quantify uncertainties for making a sensible judgement about the model response [22,23].…”

A convergence‐enhanced Timoshenko beam element for the stochastic nonlinear analysis of reinforced concrete components

“…The solution scheme for the probability density evolution method includes sample generation, determined analysis, solution of generalized density evolution equation, and calculation of probability density. For the sample generation, the point selection strategy based on the number theoretic method, 53 the sphere of contact, 54 generalized F‐discrepancy, 55 and vine copula 56 have been successively developed. For the solution of generalized density evolution equation, Li and Chen proposed Lax–Wendroff 52 and total variation diminishing solution schemes, 57 and a family of δ sequences, 58 direct probability integral method 59 are developed based on the idea of point evolution.…”

Physical vulnerability of reinforced concrete buildings under debris avalanche impact based on GF-discrepancy and DEM-FEM

“…Fortunately, a point selection method based on generalized F-discrepancy (GF-discrepancy) can deal with this problem. This method was proposed by Chen et al (2016) and improved by Yang et al (2019), and it has the advantage of handling multidimensional variable point selection with very small deviations. The principle of the method is to sample multidimensional random variables by partitioning their geometric space using Voronoi cells.…”