Abstract:The simulation of stationary non-Gaussian stochastic vector processes is of great importance in the time-domain response analysis of structures under non-Gaussian excitations. When incompatibility between the target cross power spectral density matrix and the marginal probabilistic density function of the stochastic vector process occurs in the sense of translation processes, an iterative simulation scheme must be required. Currently, the Cholesky decomposition-based iterative translation approximation method is widely utilized. However, the use of this method still encounters some difficulties. First, it requires performing Cholesky decompositions three times at each iteration and is therefore cumbersome. Second, a complicated random shuffling procedure is necessary, which renders the final iteration result non-unique. Third, it is very difficult for this method to handle cases with complex-valued spectral matrices. Lastly, it can generally be used to simulate stochastic vector processes with very limited components. To overcome these difficulties, an eigenvalue-based iterative translation approximation method is proposed in this study. This method can not only avoid the cumbersome Cholesky decomposition and complex random shuffling, but can also handle cases with complex-valued and large spectral matrices more easily. Four numerical examples with no incompatibility, significant incompatibility, complex-valued coherence and a large number of component processes are respectively employed to demonstrate the effectiveness of the proposed method. The results show that the proposed method has similar accuracy to the Cholesky decomposition-based iterative method in the first two examples. For the last two examples, the simulation accuracy is also satisfactory. The above results verify that the proposed method is superior to the Cholesky decomposition-based iterative translation approximation method overall.

Simulating strongly non-Gaussian and non-stationary processes using Karhunen–Loève expansion and L-moments-based Hermite polynomial model

Generation of Two Correlated Stationary Gaussian Processes

Results on Local Times of a Class of Multiparameter Gaussian Processes

Hybrid C- and L-Moment–Based Hermite Transformation Models for Non-Gaussian Processes

Conditional simulation of stationary non-Gaussian processes based on unified hermite polynomial model

HYBRID HERMITE MODELS FOR SHORT TERM EXTREMA ESTIMATION OF NON-GAUSSIAN PROCESSES

Similarity Based Robust Probability Latent Variable Regression Model and Its Kernel Extension for Process Monitoring

A Generalized Multi-Fidelity Simulation Method Using Sparse Polynomial Chaos Expansion

Simulation of stationary non-Gaussian stochastic vector processes using an eigenvalue-based iterative translation approximation method

Gaussian processes with skewed Laplace spectral mixture kernels for long-term forecasting

A hierarchical Bayesian modeling framework for identification of Non-Gaussian processes

Harmonizable mixture kernels with variational Fourier features

Time-variant reliability analysis using moment-based equivalent Gaussian process and importance sampling

A Decision-Making Machine Learning Approach in Hermite Spectral Approximations of Partial Differential Equations

Hairer-Quastel universality for KPZ -- polynomial smoothing mechanisms, general nonlinearities and Poisson noise

An Interpolation Method for Karhunen–Loève Expansion of Random Field Discretization

Solving High Frequency and Multi-Scale PDEs with Gaussian Processes

Solving Nonlinear Filtering Problems with Correlated Noise Based on Hermite–Galerkin Spectral Method

Bayesian Non-linear Latent Variable Modeling via Random Fourier Features

ANOVA Gaussian process modeling for high-dimensional stochastic computational models

Nonlinear Online Multioutput Gaussian Process for Multistream Data Informatics