Simulation of non-Gaussian stochastic processes/fields with application in suspension bridge cable strength estimation

In this research work, a novel simulation technique for non-Gaussian stochastic processes and fields is developed. The methodology involves an iterative scheme and is based on the spectral representation method and the concept of translation processes. The major innovation of the proposed methodology is that the underlying Gaussian process/field preserves the Gaussian characteristics throughout the iterative scheme. The immediate consequence is that the translation process concept is preserved along with all accompanying properties of a translation process/field (e.g. crossing rates). The methodology is compared with earlier versions where the concept of translation processes is not preserved throughout the iterative scheme.;This proposed methodology is then implemented in the reliability analysis of suspension bridge parallel-wire cables, which requires the evaluation of the remaining cable capacity using results of tensile strength tests performed on wire samples extracted from the bridge's main cables. The core of this reliability analysis lies in the estimation of the tensile strength of single wires over a prescribed length. Until now, standard methodologies to approach the problem of wire strength estimation usually resorted to uncorrelated random variable models to characterize the spatial variation of the tensile strength along the length of sample wires. However, such a statistical model neglects the information of the spatial correlation of the tensile strength over the wire's length---a real and experimentally measured property of ductile steel wires. In order to offset the aforementioned limitation with the existing standard methodologies, the proposed random field-based simulation methodology is thus adopted to estimate the tensile strength of a single wire segment over an arbitrary length based on a limited experimental data set. The number of parallel wires in the cable cross-section is then considered to estimate the strength of the entire cable.;The capabilities of the proposed methodology are demonstrated through a real-life application involving an experimental data set of 1-ft wire segments extracted from the Williamsburg Bridge. Results of the proposed methodology are compared to corresponding results of the current standard approaches. The paper explains why the results of the proposed methodology are the more accurate ones. Taking into account that the proposed random field-based approach is not requiring any additional experimental testing, is general enough to be applied in a straightforward way to any other suspension bridges, and its computational cost is very reasonable, it should be used for estimating the strength of suspension bridge cables.