AK-PDEMi: A failure-informed enrichment algorithm for improving the AK-PDEM in reliability analysis

Zhou, Tong; Marelli, Stefano; Sudret, Bruno; Peng, Yongbo

doi:10.1016/j.ymssp.2022.109435

Cited by 13 publications

(4 citation statements)

References 37 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…To obtain the evolutionary PDF p a (a, v) the Generalized Density Evolution Equations (GDEEs) are solved via the Probability Density Evolution Method. Following solving procedure containing 4 stages as presented in [8,15,20] is utilised in this study:…”

Section: The Probability Density Evolution Methodsmentioning

confidence: 99%

“…The following notation and sequence of establishing a first-passage failure probability problem utilizing PDEM is strongly following the considerations in [8] and supplementary by [17]. Let…”

Section: Stochastic Representation and Reliability Statementmentioning

confidence: 99%

“…For this purpose a performance criterion regarding the maximum deck acceleration is established as in [5,6,7]. Different approaches to assess the structural reliability of a specific system exist and can be categorized in approximate analytical methods, sampling techniques, surrogate methods and numerical integration based methods [8,9]. Approximation methods aim to describe the performance function directly, well known methods in this field are e.g.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Efficient time-dependent reliability analysis for a railway bridge model

Bittner,

Fritsch,

Hirzinger

et al. 2024

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

This paper proposes a framework for efficient time-dependent reliability analysis for a parametrized stochastic dynamic system, namely a train bridge load model with uncertain design properties. The Probability Density Evolution Method is utilized to explore the multidimensional random space, identify specific failure paths contributing to the failure region, and provide a full probabilistic output of the desired target quantity. The framework is tested on an uncertain railway bridge subjected to train transit (moving loads). The peak acceleration as a function of the train speed in a certain interval is analysed and utilised as performance criteria. The main sources of uncertainties are the damping and the bridge’s moment of inertia. The full evolutionary Probability Density Function of the bridge’s maximum deck acceleration is obtained, the reliability is assessed and a probability of failure estimated. The results show that in the considered speed intervals, the velocities contributing to the failure region are depending on the underlying sampling method. The Probability Density Evolution Method offers additional insight on the evolution of the critical peak accelerations while at the same time performing a reasonable amount of full model evaluations. The study concludes that further discussion is needed to determine the appropriate prediction of the train speeds that may or may not significantly contribute to the probability of failure in this bridge train model.

show abstract

Section: The Probability Density Evolution Methodsmentioning

confidence: 99%

Section: Stochastic Representation and Reliability Statementmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Efficient time-dependent reliability analysis for a railway bridge model

Bittner,

Fritsch,

Hirzinger

et al. 2024

J. Phys.: Conf. Ser.

View full text Add to dashboard Cite

show abstract

“…Active learning is a method of adding learning points to the training data by evaluating a learning function. AK-MCS by Echard et al 3 and AK-SS by Huang et al 4 are examples of previous active learning methods, and AK-PDEMi by Zhou et al 5 has recently been proposed. Most of these surrogate modeling studies have considered uncertainties in parameters of physical properties, but has not considered uncertainties in loading, especially dynamic loading due to earthquakes.…”

Section: Introductionmentioning

confidence: 99%

Gaussian process regression surrogate model for dynamic analysis to account for uncertainties in seismic loading

Saida¹,

Nishio

2023

Sensors and Smart Structures Technologies for Civil, Mechanical, and Aerospace Systems 2023

View full text Add to dashboard Cite

Reliability assessment of civil structures under seismic loads requires probabilistic evaluation considering the uncertainty of input ground motion and material properties due to deterioration. However, Monte Carlo calculation for the structural reliability analysis is computationally expensive. This study develops the deep kernel learning surrogate model that can not only reduce the computational cost but also provide explainability for the prediction results. The model extracts the features of seismic loads by the convolutional neural network (CNN) and considers the uncertainty of seismic loads and material properties by the Gaussian process regression with the automatic relevance determination (ARD) kernel. By the incorporating gradient-weighted class activation mapping (Grad-CAM) in the CNN, the parts of seismic load response spectra, where contribute to the constructed surrogate model, can be visualized. The model can also provide which input uncertain parameters of structural properties has relatively influence on the output response by the estimated ARD kernel weights. The developed surrogate model is verified by applying it to the seismic performance analysis of a concrete bridge pier with a seismic rubber bearing under various earthquake loads with different intensity and response spectra. The results show that the developed surrogate model can predict accurate distributions of maximum displacements and can provide reasonable contributions of uncertain inputs to enhance the explainability.

show abstract