Simulation of parameterized random fields, Part II: Non-Gaussian cases

Research output: Contribution to journalArticleResearchpeer review

Authors

External Research Organisations

  • Harbin Institute of Technology
  • University of Liverpool
  • Tongji University
View graph of relations

Details

Original languageEnglish
Article number113386
JournalMechanical Systems and Signal Processing
Volume240
Early online date30 Sept 2025
Publication statusPublished - 1 Nov 2025

Abstract

This paper presents two numerical algorithms to simulate non-Gaussian random fields that are parameterized by random parameters. The simulation of such kind of random fields is very challenging due to their parameterized non-Gaussian properties. For each sample realization of the random parameters, the parameterized non-Gaussian random field degrades into a classical non-Gaussian random field. In the first algorithm, we present a sample-based iterative algorithm to simulate the obtained classical non-Gaussian random field. Initial random samples are first generated to meet the sampled marginal distribution, and an iterative procedure is adopted to change the ranking of the random samples to match the target sampled covariance function. However, this method is computationally expensive since we have to simulate a non-Gaussian random field for each sample realization of the random parameters. To avoid this issue, we develop a reformulation-based algorithm in the second method. Parameterized marginal distributions are reformulated as non-parameterized marginal distributions via a conditional probability integral, and parameterized covariance functions are reformulated as non-parameterized covariance functions via an expectation operation on random parameters. In this way, the original parameterized non-Gaussian random field is transformed into a classical non-Gaussian random field. The sample-based iterative algorithm is then used to simulate the obtained non-Gaussian random field. Moreover, a multi-fidelity approach is presented to further reduce the computational effort of the above iteration by taking advantage of the Karhunen-Loève expansion. Specifically, the expanded random variables in Karhunen-Loève expansion are calculated on a low-fidelity model and the deterministic functions in Karhunen-Loève expansion are calculated on a high-fidelity model. Thus, the method has low computational effort and high fidelity simultaneously. Two numerical examples, including one- and three-dimensional parameterized non-Gaussian random fields, are used to verify the effectiveness of the proposed methods.

Keywords

    Conditional probability integral, Karhunen–Loève expansion, Multi-fidelity iteration, Parameterized non-Gaussian random fields

ASJC Scopus subject areas

Cite this

Simulation of parameterized random fields, Part II: Non-Gaussian cases. / Zheng, Zhibao; Dai, Hongzhe; Beer, Michael et al.
In: Mechanical Systems and Signal Processing, Vol. 240, 113386, 01.11.2025.

Research output: Contribution to journalArticleResearchpeer review

Zheng Z, Dai H, Beer M, Nackenhorst U. Simulation of parameterized random fields, Part II: Non-Gaussian cases. Mechanical Systems and Signal Processing. 2025 Nov 1;240:113386. Epub 2025 Sept 30. doi: 10.1016/j.ymssp.2025.113386
Download
@article{f9b86ade4406457f942c2e3f4a3015a6,
title = "Simulation of parameterized random fields, Part II: Non-Gaussian cases",
abstract = "This paper presents two numerical algorithms to simulate non-Gaussian random fields that are parameterized by random parameters. The simulation of such kind of random fields is very challenging due to their parameterized non-Gaussian properties. For each sample realization of the random parameters, the parameterized non-Gaussian random field degrades into a classical non-Gaussian random field. In the first algorithm, we present a sample-based iterative algorithm to simulate the obtained classical non-Gaussian random field. Initial random samples are first generated to meet the sampled marginal distribution, and an iterative procedure is adopted to change the ranking of the random samples to match the target sampled covariance function. However, this method is computationally expensive since we have to simulate a non-Gaussian random field for each sample realization of the random parameters. To avoid this issue, we develop a reformulation-based algorithm in the second method. Parameterized marginal distributions are reformulated as non-parameterized marginal distributions via a conditional probability integral, and parameterized covariance functions are reformulated as non-parameterized covariance functions via an expectation operation on random parameters. In this way, the original parameterized non-Gaussian random field is transformed into a classical non-Gaussian random field. The sample-based iterative algorithm is then used to simulate the obtained non-Gaussian random field. Moreover, a multi-fidelity approach is presented to further reduce the computational effort of the above iteration by taking advantage of the Karhunen-Lo{\`e}ve expansion. Specifically, the expanded random variables in Karhunen-Lo{\`e}ve expansion are calculated on a low-fidelity model and the deterministic functions in Karhunen-Lo{\`e}ve expansion are calculated on a high-fidelity model. Thus, the method has low computational effort and high fidelity simultaneously. Two numerical examples, including one- and three-dimensional parameterized non-Gaussian random fields, are used to verify the effectiveness of the proposed methods.",
keywords = "Conditional probability integral, Karhunen–Lo{\`e}ve expansion, Multi-fidelity iteration, Parameterized non-Gaussian random fields",
author = "Zhibao Zheng and Hongzhe Dai and Michael Beer and Udo Nackenhorst",
note = "Publisher Copyright: {\textcopyright} 2025 The Author(s)",
year = "2025",
month = nov,
day = "1",
doi = "10.1016/j.ymssp.2025.113386",
language = "English",
volume = "240",
journal = "Mechanical Systems and Signal Processing",
issn = "0888-3270",
publisher = "Academic Press",

}

Download

TY - JOUR

T1 - Simulation of parameterized random fields, Part II

T2 - Non-Gaussian cases

AU - Zheng, Zhibao

AU - Dai, Hongzhe

AU - Beer, Michael

AU - Nackenhorst, Udo

N1 - Publisher Copyright: © 2025 The Author(s)

PY - 2025/11/1

Y1 - 2025/11/1

N2 - This paper presents two numerical algorithms to simulate non-Gaussian random fields that are parameterized by random parameters. The simulation of such kind of random fields is very challenging due to their parameterized non-Gaussian properties. For each sample realization of the random parameters, the parameterized non-Gaussian random field degrades into a classical non-Gaussian random field. In the first algorithm, we present a sample-based iterative algorithm to simulate the obtained classical non-Gaussian random field. Initial random samples are first generated to meet the sampled marginal distribution, and an iterative procedure is adopted to change the ranking of the random samples to match the target sampled covariance function. However, this method is computationally expensive since we have to simulate a non-Gaussian random field for each sample realization of the random parameters. To avoid this issue, we develop a reformulation-based algorithm in the second method. Parameterized marginal distributions are reformulated as non-parameterized marginal distributions via a conditional probability integral, and parameterized covariance functions are reformulated as non-parameterized covariance functions via an expectation operation on random parameters. In this way, the original parameterized non-Gaussian random field is transformed into a classical non-Gaussian random field. The sample-based iterative algorithm is then used to simulate the obtained non-Gaussian random field. Moreover, a multi-fidelity approach is presented to further reduce the computational effort of the above iteration by taking advantage of the Karhunen-Loève expansion. Specifically, the expanded random variables in Karhunen-Loève expansion are calculated on a low-fidelity model and the deterministic functions in Karhunen-Loève expansion are calculated on a high-fidelity model. Thus, the method has low computational effort and high fidelity simultaneously. Two numerical examples, including one- and three-dimensional parameterized non-Gaussian random fields, are used to verify the effectiveness of the proposed methods.

AB - This paper presents two numerical algorithms to simulate non-Gaussian random fields that are parameterized by random parameters. The simulation of such kind of random fields is very challenging due to their parameterized non-Gaussian properties. For each sample realization of the random parameters, the parameterized non-Gaussian random field degrades into a classical non-Gaussian random field. In the first algorithm, we present a sample-based iterative algorithm to simulate the obtained classical non-Gaussian random field. Initial random samples are first generated to meet the sampled marginal distribution, and an iterative procedure is adopted to change the ranking of the random samples to match the target sampled covariance function. However, this method is computationally expensive since we have to simulate a non-Gaussian random field for each sample realization of the random parameters. To avoid this issue, we develop a reformulation-based algorithm in the second method. Parameterized marginal distributions are reformulated as non-parameterized marginal distributions via a conditional probability integral, and parameterized covariance functions are reformulated as non-parameterized covariance functions via an expectation operation on random parameters. In this way, the original parameterized non-Gaussian random field is transformed into a classical non-Gaussian random field. The sample-based iterative algorithm is then used to simulate the obtained non-Gaussian random field. Moreover, a multi-fidelity approach is presented to further reduce the computational effort of the above iteration by taking advantage of the Karhunen-Loève expansion. Specifically, the expanded random variables in Karhunen-Loève expansion are calculated on a low-fidelity model and the deterministic functions in Karhunen-Loève expansion are calculated on a high-fidelity model. Thus, the method has low computational effort and high fidelity simultaneously. Two numerical examples, including one- and three-dimensional parameterized non-Gaussian random fields, are used to verify the effectiveness of the proposed methods.

KW - Conditional probability integral

KW - Karhunen–Loève expansion

KW - Multi-fidelity iteration

KW - Parameterized non-Gaussian random fields

UR - http://www.scopus.com/inward/record.url?scp=105017241045&partnerID=8YFLogxK

U2 - 10.1016/j.ymssp.2025.113386

DO - 10.1016/j.ymssp.2025.113386

M3 - Article

AN - SCOPUS:105017241045

VL - 240

JO - Mechanical Systems and Signal Processing

JF - Mechanical Systems and Signal Processing

SN - 0888-3270

M1 - 113386

ER -

By the same author(s)