Simulation of non-Gaussian wind field as a $3^{rd}$-order stochastic wave

Abstract

This paper presents a methodology for the simulation of non-Gaussian wind field as a stochastic wave using the 3rd-order Spectral Representation Method. Traditionally, the wind field is modeled as a stochastic vector process at discrete locations in space. But the simulation of vector process is well-known to be computationally challenging and numerically unstable when modeling wind at a large number of discrete points in space. Recently, stochastic waves have been used to model the field as a continuous process indexed both in time and space. We extend the classical Spectral Representation Method for simulation of Gaussian stochastic waves to a third-order representation modeling asymmetrically skewed non-Gaussian stochastic waves from a prescribed power spectrum and bispectrum. We present an efficient implementation using the fast Fourier transform, which reduces the computational time dramatically. We then apply the method for simulation of a non-Gaussian wind velocity field along a long-span bridge.