Structural Safety最新文献

{"title":"Accurate variance estimation for subset simulation incorporating intrachain, interchain, and interlevel correlations","authors":"Qingqing Miao,&nbsp;Ying Min Low","doi":"10.1016/j.strusafe.2026.102690","DOIUrl":"10.1016/j.strusafe.2026.102690","url":null,"abstract":"<div><div>Subset simulation (SS) is widely held as a powerful technique for evaluating small failure probabilities. Variance estimation is integral to assessing the uncertainty of the probability estimate. However, variance estimation for SS is complex as samples are generated by Markov chain Monte Carlo (MCMC), resulting in an intricate web of correlations that fall under three categories: (1) within a chain (intrachain), (2) across separate chains (interchain), and (3) between subset levels (interlevel). To date, hardly any advances have been made on this challenging topic. Most studies using SS adopt the conventional variance estimation method, which considers the intrachain correlation but neglects other correlation types. In a recent study, the authors showed that all three correlation types are important, and developed a method that accounts for the intrachain and interchain correlations. This paper presents a theoretical framework for the interlevel correlations, bridging the final gap and illuminating a long-standing unsolved problem. The method utilizes information available from a single SS run. The equations reveal fascinating insights concerning the mechanism of interlevel correlations, valuable to researchers working on enhancing MCMC algorithms for SS. Among other things, it is mathematically proven that if samples within a level are independent, this level and the next must be independent. The new model is integrated with the prior work to produce a variance estimation method that incorporates all sources of correlations. Case studies with multiple independent SS runs show that the proposed method estimates the variance accurately, providing a vast improvement over the conventional method.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102690"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146077805","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Prediction of the flexural behavior of corroded prestressed concrete girders: a probabilistic multi-level approach","authors":"Seungjun Lee ,&nbsp;Chi-Ho Jeon ,&nbsp;Jaebeom Lee ,&nbsp;Young-Joo Lee","doi":"10.1016/j.strusafe.2025.102685","DOIUrl":"10.1016/j.strusafe.2025.102685","url":null,"abstract":"<div><div>This paper introduces a probabilistic multi-level framework for predicting the flexural behavior of corroded prestressed concrete (PSC) girders. The proposed framework employs a hierarchical modeling strategy that progresses from the wire to the girder level and integrates detailed finite element (FE) analysis, surrogate modeling, and Monte Carlo simulations. This computationally efficient framework addresses the challenge of accurately predicting flexural behavior by systematically incorporating the effects of the geometric complexity of the corroded strands and other inherent modeling uncertainties into its probabilistic predictions. The surrogate model constructed from the FE results enables efficient predictions by accounting for material and geometric uncertainties across multiple structural levels. Experimental validation was performed using ten PSC girder specimens, comprising both single- and multi-strand configurations, subjected to controlled corrosion and flexural loading tests. The predicted load–displacement responses, including the 50 %, 95 %, and 99 % prediction ranges, exhibited good agreement with the experimental results, successfully capturing key indicators of structural performance, such as loads and deflections at yield and ultimate. In addition, a global sensitivity analysis identified the dominant sources of uncertainty influencing the variability in the probabilistic predictions. These findings confirm the ability of the proposed framework to accurately model corrosion-induced degradation and reliably quantify the associated uncertainties.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102685"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925965","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A critical review and analysis of the uncertainties involved in fatigue damage assessment and their impact on decision-making","authors":"Somayeh Shojaeikhah ,&nbsp;Baran Yeter ,&nbsp;Mohamed Soliman ,&nbsp;Yordan Garbatov","doi":"10.1016/j.strusafe.2026.102687","DOIUrl":"10.1016/j.strusafe.2026.102687","url":null,"abstract":"<div><div>Fatigue damage is a major driver of interventions in marine and offshore structures. The fluctuating loads applied to these structures can lead to crack initiation and propagation. Accordingly, effective management activities are needed to ensure the safety and reliability of these structures. Predicting the service life under fatigue damage is a crucial step in the effective management of these structures. However, this process is challenged by the presence of significant uncertainties introduced by the natural randomness in sea loading and mechanical behavior. To date, the literature does not present inclusive guidance on quantifying and accounting for these uncertainties in the damage prediction process. To address this need, this paper critically reviews the uncertainties associated with the fatigue damage assessment in ships and offshore structures, including offshore wind turbines, with a specific focus on the reliability and risk as the probabilistic performance indicators. The review covers the S-N approaches and the fracture mechanics-based damage tolerance design and assessment techniques, and discusses the potential discrepancies in their treatment of uncertainties. By systematically evaluating these aspects, this review provides a much-needed insight into existing knowledge gaps and suggests directions for future research on fatigue damage assessment protocols in the marine and offshore engineering domains.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102687"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145925966","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"A transfer learning approach to predict corrosion-induced concrete cracking based on steel weight loss distributions","authors":"Siyi Jia ,&nbsp;Mitsuyoshi Akiyama ,&nbsp;Dan M. Frangopol","doi":"10.1016/j.strusafe.2025.102684","DOIUrl":"10.1016/j.strusafe.2025.102684","url":null,"abstract":"<div><div>Predictive models of corrosion-induced concrete cracking based on steel weight loss (SWL) distributions are crucial for assessing structural degradation and enabling proactive maintenance of reinforced concrete (RC) structures. Data-driven models offer an efficient alternative to predict corrosion-induced concrete cracking, which facilitate real-time monitoring of corrosion damage while capturing cracking mechanisms amid noisy data. However, developing such data-driven predictive models is constrained by limited experimental samples and domain drift across varying corrosion rates. This paper proposes a transfer learning framework to address these challenges. The proposed method first denoises SWL data and extracts principal components via the Karhunen-Loève (KL) transformation. Transfer component analysis (TCA) then aligns joint probability distributions of SWL and cracking states across different corrosion rates in a learned feature space, where K-nearest neighbors (KNN) serves as the base classifier. The proposed KL-TCA model is validated using an existing experimental dataset and achieves a prediction accuracy of 87.32%, which notably outperforms the baseline methods based on TCA and direct KNN. Then, an ensemble learning model is constructed to integrate KL-TCA base learners with different configurations, which achieves a prediction accuracy of 93.89% on the validation set, surpassing that of any individual KL-TCA base learner. Additionally, a damage indicator, defined as the percentage of cracked sections, is computed using the KL-TCA method. Considering the uncertainties in SWL distributions, a Monte Carlo simulation is performed to establish the relationship between the damage indicator and the average SWL. This relationship allows for a preliminary screening of corrosion severity using readily available cracking observations, supporting timely corrosion assessments and maintenance decisions.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102684"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884801","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"On the quantification of robustness and its thresholds","authors":"Alex Sixie Cao ,&nbsp;André T. Beck","doi":"10.1016/j.strusafe.2026.102688","DOIUrl":"10.1016/j.strusafe.2026.102688","url":null,"abstract":"<div><div>Structural systems need to be safe enough against foreseeable loads, but they also need to be robust enough against unforeseeable or abnormal loading. In this paper, a novel entropy-based robustness index for arbitrary perturbations is derived for coherent path-dependent systems, which is consistent with information-theoretic and thermodynamic principles. Using a reliability-based robustness index and the entropy-based robustness index, quantitative robustness thresholds are derived that enable the explicit classification of <em>low</em>, <em>medium</em>, and <em>high</em> robustness based on the sensitivity of the system to arbitrary perturbations. Furthermore, relations between the entropy- and reliability- and risk-based robustness indices are explored, where thresholds for the risk-based robustness indices are provided based on the novel entropy-based robustness index. The use of the various robustness indices and the thresholds are exemplified in three case studies, involving a redundant system subjected to various degrees of damage, damage propagation in frame structures, and a network. For the first time, quantitative thresholds for the robustness of coherent path-dependent systems are provided, which can be applied to structures, networks, and more. This paves the way for providing quantitative guidance on acceptable degrees of robustness in such systems, which may lead to more economic and rational systems with an appropriate degree of robustness.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102688"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145976954","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Adaptive machine learning-driven multi-fidelity stratified sampling for failure analysis of nonlinear stochastic systems","authors":"Liuyun Xu,&nbsp;Seymour M.J. Spence","doi":"10.1016/j.strusafe.2025.102673","DOIUrl":"10.1016/j.strusafe.2025.102673","url":null,"abstract":"<div><div>Existing variance reduction techniques used in stochastic simulations for rare event analysis still require a substantial number of model evaluations to estimate small failure probabilities. In the context of complex, nonlinear finite element modeling environments, this can become computationally challenging, particularly for systems subjected to stochastic excitation. To address this challenge, a multi-fidelity stratified sampling scheme with adaptive machine learning metamodels is introduced for efficiently propagating uncertainties and estimating small failure probabilities. In this approach, a high-fidelity dataset generated through stratified sampling is used to train a deep learning-based metamodel, which then serves as a cost-effective and highly correlated low-fidelity model. An adaptive training scheme is proposed to balance the trade-off between approximation quality and computational demand associated with the development of the low-fidelity model. By integrating the low-fidelity outputs with additional high-fidelity results, an unbiased estimate of the strata-wise failure probabilities is obtained using a multi-fidelity Monte Carlo framework. The overall probability of failure is then computed using the total probability theorem. Application to a full-scale high-rise steel building subjected to stochastic wind excitation demonstrates that the proposed scheme can accurately estimate exceedance probability curves for nonlinear responses of interest while achieving significant computational savings compared to single-fidelity variance reduction approaches.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102673"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145738387","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Encoding of decision trees for life-cycle cost and decision value analysis via optimization","authors":"Wellison José de Santana Gomes ,&nbsp;Sebastian Thöns ,&nbsp;André Teófilo Beck","doi":"10.1016/j.strusafe.2026.102689","DOIUrl":"10.1016/j.strusafe.2026.102689","url":null,"abstract":"<div><div>Reliable and cost-effective operation of structural systems over their service life depends on the implementation of Structural Health Monitoring (SHM) and maintenance activities, which influence operational costs, the expected costs of failure and downtime. Decision Value Analysis (DVA) provides a framework to quantify the value of such activities by evaluating their effect on total expected lifecycle costs. While prior studies have focused on isolated decisions or employed heuristic rules to reduce computational demands, an integrated, system-wide, lifetime-based approach is needed to capture interdependencies among components and among decisions, avoiding suboptimal outcomes. This paper addresses the complex problem of optimizing a sequence of SHM and maintenance decisions over the structure’s entire service life, without relying on fixed pre-defined heuristic rules. An approach to encode all decision variables into a single vector of design variables is presented, and an adaptive surrogate modeling strategy is employed to efficiently approximate the total expected cost function, significantly reducing the computational burden. A case study on corrosion in buried steel pipelines is presented, allowing up to nine inspections and the associated repair decisions, resulting in 1533 decision variables and 2¹⁵³³ possible combinations. Results indicate, as expected, that early inspections may be omitted when their cost exceeds the marginal benefit in risk reduction, but also that more frequent inspections can support more effective repair decisions. The proposed approach provides a generalizable and computationally efficient framework for lifecycle DVA, which can be directly applied to more complex problems, and is capable of incorporating multiple inspection and maintenance methods.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102689"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146022916","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Sequential active learning for estimating small failure probabilities in high-dimensional problems: Application to nonlinear vessel responses","authors":"Tomoki Takami ,&nbsp;Masaru Kitahara","doi":"10.1016/j.strusafe.2025.102686","DOIUrl":"10.1016/j.strusafe.2025.102686","url":null,"abstract":"<div><div>A new method for high-dimensional structural reliability analysis is proposed, with particular attention to estimating small failure probabilities. The proposed method is built upon an active learning framework, in which an active subspace for supervised dimensionality reduction and a surrogate model for bypassing the performance function are simultaneously updated. Heteroscedastic Gaussian process (hGP) modeling is employed for this purpose. To effectively address rare event problems, the method further incorporates a sequential sampling strategy based on the subset simulation. The resulting Sequential Active Learning with Active Subspace (SALAS) method is first demonstrated using the Sobol function to illustrate its accuracy and computational efficiency. Following this, its application is extended to specific high-dimensional engineering problems involving nonlinear vessel responses in waves. Two subject vessel responses are studied: vertical bending moment and roll motion of a vessel. A nonlinear strip theory and a two-degree-of-freedom roll motion model are used to analyze these responses, respectively. Comprehensive comparisons with crude Monte Carlo simulation, first order reliability method, and the adaptive active subspace-based heteroscedastic Gaussian process (AaS-hGP) method demonstrates the efficiency and accuracy of the proposed SALAS method, even in estimating rare event probabilities in high-dimensional stochastic spaces.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102686"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145884800","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Fusing experimental and FEM-based knowledge: a transfer learning model for inferring steel corrosion in reinforced concrete structures","authors":"Siyi Jia ,&nbsp;Mitsuyoshi Akiyama ,&nbsp;Dan M. Frangopol","doi":"10.1016/j.strusafe.2025.102672","DOIUrl":"10.1016/j.strusafe.2025.102672","url":null,"abstract":"<div><div>Corrosion-induced crack width (CCW) can be readily obtained through visual inspection of reinforced concrete structures and serves as a proxy indicator of steel corrosion based on its relationship with steel weight loss (SWL). Although accelerated corrosion tests and FEM simulations provide a cost-effective data source to derive the relationship between CCW and SWL, the calibrated models often fail to generalize under natural corrosion conditions. This study proposes a transfer learning-based model that fuses knowledge from accelerated corrosion tests and FEM simulations to predict CCW from SWL distribution under natural corrosion conditions. This model is then used as the forward model in a Bayesian inference scheme to estimate SWL distributions. Specifically, the proposed approach first uses an unsupervised transfer learning model that combines the geodesic flow kernel (GFK) and transfer component analysis (TCA) to align the marginal distributions of SWL across different corrosion conditions. This unsupervised model is extended to a semi-supervised approach by introducing FEM-generated pseudo labels for CCW under natural corrosion conditions, which embeds the statistical dependence between SWL and CCW into the TCA projection in a structure-aware manner. Numerical results demonstrate that the proposed transfer learning models effectively transfer the SWL-CCW mapping learned from a combination of accelerated corrosion tests using the galvanostatic method and FEM simulations to natural corrosion conditions represented by the artificial chloride environment (ACE) method. The unsupervised GFK-TCA method improves CCW prediction accuracy under ACE conditions by 115% compared to the non-transfer learning baseline. Furthermore, the semi-supervised GFK-TCA method achieves an additional 18.0% improvement in accuracy over the unsupervised GFK-TCA method. Based on the transfer learning model, Bayesian inference yields a range estimate of SWL distribution that covers nearly 70.0% of the observations in the RC specimens corroded by ACE method.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102672"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145693130","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

{"title":"Reliability sensitivity with response gradient","authors":"Siu-Kui Au ,&nbsp;Zi-Jun Cao","doi":"10.1016/j.strusafe.2025.102683","DOIUrl":"10.1016/j.strusafe.2025.102683","url":null,"abstract":"<div><div>Engineering risk is concerned with the likelihood of failure and the scenarios when it occurs. The sensitivity of failure probability to change in system parameters is relevant to risk-informed decision making. Computing sensitivity is at least one level more difficult than the probability itself, which is already challenged by a large number of input random variables, rare events and implicit nonlinear ‘black-box’ response. Finite difference with Monte Carlo probability estimates is spurious, requiring the number of samples to grow with the reciprocal of step size to suppress estimation variance. Many existing works gain efficiency by exploiting a specific class of input variables, sensitivity parameters, or response in its exact or surrogate form. For general systems, this work presents a theory and Monte Carlo strategy for computing sensitivity using response values and gradients with respect to sensitivity parameters. It is shown that the sensitivity at a given response threshold can be expressed via the expectation of response gradient conditional on the threshold. Determining the expectation requires conditioning on the threshold that is a zero-probability event, but it can be resolved by kernel smoothing. The proposed method offers sensitivity estimates for all response thresholds generated in a Monte Carlo run. It is investigated in a number of examples featuring sensitivity parameters of different nature. As response gradient becomes increasingly available, it is hoped that this work can provide the basis for embedding sensitivity calculations with reliability in the same Monte Carlo run.</div></div>","PeriodicalId":21978,"journal":{"name":"Structural Safety","volume":"120 ","pages":"Article 102683"},"PeriodicalIF":6.3,"publicationDate":"2026-05-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"145791571","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}