Abstract:This study presents a novel methodology, the enhanced higher-order spectral representation method (HOSRM), designed for the effective simulation of multivariate non-Gaussian stochastic processes. This approach first extends the spectral representation method by introducing the Einstein tensor representation of cross-spectra and cross-trispectra, thereby enabling the modeling of multivariate fourth-order non-Gaussian stochastic processes. To enhance computational efficiency and minimize the memory requirements associated with higher-order spectral matrices, particularly for simulating multivariate scenarios, the non-uniform fast Fourier transform (NUFFT) is used instead of the traditional FFT operation. The acceleration of trigonometric function summation is achieved by employing meticulously designed, non-uniformly distributed frequency points and harnessing the capabilities of the NUFFT algorithm. This algorithm potently reduces the number of necessary segments within the Fourier transform process, thus eliminating the need to simulate unimportant points typically encountered in traditional FFT operations. As a result, computational efficiency is significantly improved, enabling more focused and precise simulations. Meanwhile, a strict proof between the 1st-, 2nd-, and 4th-order statistical characteristics of the simulated 4th-order random vector processes and their corresponding cross-correlation properties is performed. Finally, numerical simulation is performed to validate the efficacy of this approach, using non-Gaussian wind pressures acquired from a wind tunnel experiment. For example, the outcomes discussed in this study reveal that the simulated multivariate non-Gaussian wind pressures' statistical properties, including variance, kurtosis, cross-spectra, and cross-trispectra, align with the intended objectives. These findings provide convincing evidence of the accuracy and reliability of the simulation model employed in this research. Notably, the calculation efficiency of this method is about 30 times that of the FFT method.

Fractional Spectral Moments for Digital Simulation of Multivariate Wind Velocity Fields

Crank-Nicholson Scheme Based Wind Speed Modeling

Simulation of non-Gaussian wind field as a 3rd-order stochastic wave

A Matlab Toolbox for Positive Fractional Time Derivative Modeling of Arbitrarily Frequency-Dependent Viscosity

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

Constructing Transient Response Probability Density of Non-Linear System Through Complex Fractional Moments

Efficient simulation of two-spatial dimensional turbulent wind fields based on the factorization of random functions

A Novel Exact Representation of Stationary Colored Gaussian Processes (Fractional Differential Approach)

NUFFT-enhanced higher-order spectral representation for simulating multivariate non-Gaussian wind pressures

Simulation of ergodic multivariate stochastic processes: An enhanced spectral representation method

Fractional diffusion-reaction stochastic simulations

Multi-domain spectral approach to rational-order fractional derivatives

Multidomain spectral approach to rational‐order fractional derivatives

Numerical Simulation of Non-Gaussian Wave Elevation and Kinematics Based on Two-Dimensional Fourier Transform

A fast algorithm for a three-dimensional synthetic model of intermittent turbulence

Analysis of time dynamics in wind records by means of multifractal detrended fluctuation analysis and Fisher-Shannon information plane

Simulation of a spatially correlated turbulent velocity field using biorthogonal decomposition

THE IMPROVED SIMULATION AND WAVELET ANALYSIS OF WIND VELOCITY TIME SERIES FOR LONG-SPAN SPATIAL STRUCTURES

Numerical simulations of a stochastic dynamics leading to cascades and loss of regularity: applications to fluid turbulence and generation of fractional Gaussian fields

Superstatistical wind fields from point-wise atmospheric turbulence measurements

Intermittent process analysis with scattering moments