.png){kind=icon}
.png){kind=icon}
Return
Joint probability density function-driven simulation of multivariate non-Gaussian processes
.png){kind=icon}
PRE
AI
M
L
A
DOI：10.1016/j.probengmech.2026.103957.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
This study proposes a data-driven simulation method for multivariate non-Gaussian processes that reproduces not only spectral characteristics but also the joint probability density function (JPDF) of a given target time series. The method comprises three sequential steps that update the power spectral density (PSD), the marginal probabilistic density function (PDF), and the JPDF iteratively. The PSD correction step generates a time series that replicates the template’s Fourier amplitudes and phase differences, whose marginal PDF is then enforced via rank matching. More importantly, non-rigid point set registration based on the Coherent Point Drift (CPD) algorithm is used to align the marginal-PDF-corrected time series with the target’s JPDF. Repeating these steps iteratively yields simulations that jointly match the PSD, marginal CDF, and, importantly, the JPDF of the target time series. The effectiveness of the method is illustrated by a numerical example that simulates the wind pressure field on a low-rise building.
Journal
IF:
3.5
Papers:
1.7K
Citations:
4.1K
Organization
Cited Papers