Browse by author
Lookup NU author(s): Professor Emilio Porcu
Full text for this publication is not currently held within this repository. Alternative links are provided below where available.
This paper reports on a generalization of Lamb's problem to a linear elastic, infinite half-space with random fields (RFs) of mass density, subject to a normal line load. Both, uncorrelated and correlated (with fractal and Hurst characteristics) RFs without any weak noise restrictions, are proposed. Cellular automata (CA) is used to simulate the wave propagation. CA is a local computational method which, for rectangular discretization of spatial domain, is equivalent to applying the finite difference method to the governing equations of classical elasticity. We first evaluate the response of CA to an uncorrelated mass density field, more commonly known as white-noise, of varying coarseness as compared to CA's node density. We then evaluate the response of CA to multiscale mass density RFs of Cauchy and Dagum type; these fields are unique in that they are able to model and decouple the field's fractal dimension and Hurst parameter. We focus on stochastic imperfection sensitivity; we determine to what extent the fractal or the Hurst parameter is a significant factor in altering the solution to the planar stochastic Lamb's problem by evaluating the coefficient of variation of the response when compared with the coefficient of variation of the RF.
Author(s): Nishawala CC, Ostoja-Starzewski M, Leamy MJ, Porcu E
Publication type: Article
Publication status: Published
Journal: Proceedings of the Royal Society of London A: Mathematical, Physical and Engineering Sciences
Year: 2016
Volume: 472
Issue: 2196
Online publication date: 07/12/2016
Acceptance date: 04/11/2016
ISSN (print): 1364-5021
ISSN (electronic): 1471-2946
Publisher: The Royal Society Publishing
URL: https://doi.org/10.1098/rspa.2016.0638
DOI: 10.1098/rspa.2016.0638
Altmetrics provided by Altmetric