An adaptive Kriging method with PCE-based dimension reduction for structural reliability analysis

The PCE-Kriging method integrates the advantages of Polynomial Chaos Expansion (PCE) and Kriging by employing PCE as the trend function to capture the global behavior of the performance function, while Kriging provides unbiased predictions together with associated uncertainty estimates. However, this method suffers from the curse of dimensionality, the assumption of mutual independence among input random variables, and a strong dependence of model accuracy and efficiency on the choice of the learning function. To overcome these limitations, an adaptive Kriging method with PCE-based dimension reduction (AK-DRPCE) is proposed for structural reliability analysis. Based on the PCE-Kriging framework, the proposed method introduces a dimension reduction parameter s to limit the interaction order of polynomial basis functions. Furthermore, the existing learning function $${\psi _m}$$ is enhanced by incorporating an adaptive weighting mechanism between the sample-distance and the prediction-variance penalty terms, as well as a local density control strategy for samples based on the k-nearest neighbor principle, resulting in a new learning function $$\psi _{m}^{\prime }$$ . Finally, the accuracy and efficiency of the proposed method are validated through a numerical example. Furthermore, probability models of vehicle load and temperature load are constructed based on bridge structural health monitoring (SHM) data, and the load-carrying capacity reliability of an in-service bridge is evaluated. The results demonstrate that, while maintaining accurate failure probability estimates, the AK-DRPCE+ $$\psi _{m}^{\prime }$$ method can reduce the number of performance function evaluations by up to 25% compared with existing methods, exhibiting high computational efficiency and strong engineering applicability.

Authors

Institutions

Publication Details

Journal
Advances in Bridge Engineering
Published
2026-11-15
DOI
https://doi.org/10.1186/s43251-026-00235-2
Primary Topic
Probabilistic and Robust Engineering Design
Type
article
Field-Weighted Citation Impact
0.00

Funders

Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
article

An adaptive Kriging method with PCE-based dimension reduction for structural reliability analysis

Yu Hong, 蒲黔辉, Z.J. Zhang
Advances in Bridge Engineering
Probabilistic and Robust Engineering Design
article

An adaptive Kriging method with PCE-based dimension reduction for structural reliability analysis

Yu Hong, 蒲黔辉, Z.J. Zhang
article en

Abstract

The PCE-Kriging method integrates the advantages of Polynomial Chaos Expansion (PCE) and Kriging by employing PCE as the trend function to capture the global behavior of the performance function, while Kriging provides unbiased predictions together with associated uncertainty estimates. However, this method suffers from the curse of dimensionality, the assumption of mutual independence among input random variables, and a strong dependence of model accuracy and efficiency on the choice of the learning function. To overcome these limitations, an adaptive Kriging method with PCE-based dimension reduction (AK-DRPCE) is proposed for structural reliability analysis. Based on the PCE-Kriging framework, the proposed method introduces a dimension reduction parameter s to limit the interaction order of polynomial basis functions. Furthermore, the existing learning function $${\psi _m}$$ is enhanced by incorporating an adaptive weighting mechanism between the sample-distance and the prediction-variance penalty terms, as well as a local density control strategy for samples based on the k-nearest neighbor principle, resulting in a new learning function $$\psi _{m}^{\prime }$$ . Finally, the accuracy and efficiency of the proposed method are validated through a numerical example. Furthermore, probability models of vehicle load and temperature load are constructed based on bridge structural health monitoring (SHM) data, and the load-carrying capacity reliability of an in-service bridge is evaluated. The results demonstrate that, while maintaining accurate failure probability estimates, the AK-DRPCE+ $$\psi _{m}^{\prime }$$ method can reduce the number of performance function evaluations by up to 25% compared with existing methods, exhibiting high computational efficiency and strong engineering applicability.

Advances in Bridge EngineeringVol. 7(1)
Southwest Jiaotong University (CN)
National Natural Science Foundation of China
Industry, innovation and infrastructure
Probabilistic and Robust Engineering Design
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon