Reliability Engineering & System Safety最新文献_第5页

A hybrid deep learning and large language models framework for ship collision accident analysis 船舶碰撞事故分析的混合深度学习和大语言模型框架

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-01-31 DOI: 10.1016/j.ress.2026.112333

Xiyu Zhang , Langxiong Gan , Hailong Cui , Yaqing Shu , Jakub Montewka , Zaili Yang

{"title":"A hybrid deep learning and large language models framework for ship collision accident analysis","authors":"Xiyu Zhang , Langxiong Gan , Hailong Cui , Yaqing Shu , Jakub Montewka , Zaili Yang","doi":"10.1016/j.ress.2026.112333","DOIUrl":"10.1016/j.ress.2026.112333","url":null,"abstract":"<div><div>Innovative artificial intelligence technology has shown potential in ship accident research. Deep learning (DL) is an effective method used to analyse structured accident data. Meanwhile, the development of large language models (LLMs) has brought new breakthroughs to the study of accident occurrence processes in reports. A novel framework combining the two is proposed in this study to provide new solutions to ship collision accident analysis. Firstly, a structured accident analysis method, named Hierarchical graph attention network with Transformer encoding (HGAT-Transformer) model, is proposed. Secondly, a pre-trained LLM is fine-tuned based on low-rank adaptation and a retrieval-augmented generation technology is adopted. Finally, a hybrid framework, named Deep learning-large language model combined framework to Ship Collision Analysis (DLSCA), is developed to guide better accident analysis. Real accident report data from different regions are used for DL model training and LLM fine-tuning. The results show that the generated analysis of the proposed framework is consistent with real accident reports in terms of accident analysis and responsibility assignment. The research demonstrates the potential of LLMs in providing guidance for accident analysis, and provides new insights for research on ship collision accidents.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112333"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174852","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}

引用次数: 0

Hazard and vulnerability analysis of NaTech disasters induced by hydrological events to support probabilistic safety assessment in natural gas pipelines 水文事件诱发的NaTech灾害危害与脆弱性分析，支持天然气管道概率安全评估

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-02-09 DOI: 10.1016/j.ress.2026.112381

Francisco Filipe Cunha Lima Viana , Lucas Borges Leal da Silva , Marcelo Hazin Alencar , Adiel Teixeira de Almeida

{"title":"Hazard and vulnerability analysis of NaTech disasters induced by hydrological events to support probabilistic safety assessment in natural gas pipelines","authors":"Francisco Filipe Cunha Lima Viana , Lucas Borges Leal da Silva , Marcelo Hazin Alencar , Adiel Teixeira de Almeida","doi":"10.1016/j.ress.2026.112381","DOIUrl":"10.1016/j.ress.2026.112381","url":null,"abstract":"<div><div>Climate change is increasing the frequency and intensity of hydrological events, posing significant threats to critical infrastructure such as natural gas pipelines. These pipelines are vulnerable to flooding, which can result in gas releases and trigger Natural Hazard Triggering Technology (NaTech) disasters. The combined forces of floodwaters, floating debris, and soil erosion pose a potential threat to pipelines, highlighting the need for a comprehensive analysis of NaTech events. To support safer, more informed decisions, current research increasingly emphasizes discrete failure modes and vulnerability scenarios, providing a solid basis for understanding uncertainty and for guiding effective risk management. However, a critical gap remains in accounting for the complex interplay between hydrological conditions and the inherent uncertainties in pipeline integrity. This study proposes a systematic approach to analyze the hazards of flood-NaTech events. The proposed probabilistic approach analyzes vulnerability aspects by integrating hydrological modeling with technical analysis of the pipeline, structuring uncertainty scenarios.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112381"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174953","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}

引用次数: 0

Real-time graph-based traffic complexity evaluation for intelligent maritime supervision using multi-source data fusion 基于多源数据融合的智能海事监管交通复杂性实时图评估

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-02-08 DOI: 10.1016/j.ress.2026.112380

Congcong Zhao , Yuxu Lu , Dong Yang, Tsz Leung Yip

{"title":"Real-time graph-based traffic complexity evaluation for intelligent maritime supervision using multi-source data fusion","authors":"Congcong Zhao , Yuxu Lu , Dong Yang, Tsz Leung Yip","doi":"10.1016/j.ress.2026.112380","DOIUrl":"10.1016/j.ress.2026.112380","url":null,"abstract":"<div><div>Ship trajectory data from the Automatic Identification System (AIS) play a crucial role in waterway risk assessment. However, challenges like missing data, ship interactions, and complex regional traffic patterns hinder accurate risk identification. To address this issue, a real-time graph-based evaluation framework is developed for traffic complexity in intelligent maritime supervision, utilizing multi-source data fusion. AIS and video data are processed to extract and reconstruct ship trajectories. Long short-term memory (LSTM) networks are applied to predict ship trajectories, achieving real-time ship behavior data. Graph neural networks (GNN) enable data fusion to identify missing ship data and enhance the precision of traffic complexity evaluation. A homography model reconstructs incomplete paths based on nearby dynamics. This method improves data integrity and captures dynamic interactions. Waterway regionalization represents ships as nodes in a dynamic graph. GNN message passing updates embeddings to account for interactions, and Density-Based Spatial Clustering of Applications with Noise (DBSCAN) clustering segments regions with similar patterns. Traffic complexity is estimated through weighted interaction graphs for each cluster, with adjacency matrices used for risk profiling. Applied to Shanghai’s Huangpu River in eight scenarios and time periods, the framework identifies high-risk areas. Model validation is conducted through comparative analysis with prior studies to ensure reliability. Results validate the methodology and confirm its adaptability, offering actionable guidance for maritime authorities and safety policymakers.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112380"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174918","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}

引用次数: 0

A privacy-enhanced multi-party industrial control systems collaboration framework 一个隐私增强的多方工业控制系统协作框架

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-02-01 DOI: 10.1016/j.ress.2026.112318

Xiangzhen Peng , Yunqi Liu , Chengliang Zheng , Liang Dong , Umer Zukaib , Xiaohui Cui

{"title":"A privacy-enhanced multi-party industrial control systems collaboration framework","authors":"Xiangzhen Peng , Yunqi Liu , Chengliang Zheng , Liang Dong , Umer Zukaib , Xiaohui Cui","doi":"10.1016/j.ress.2026.112318","DOIUrl":"10.1016/j.ress.2026.112318","url":null,"abstract":"<div><div>As a core component of critical infrastructure, industrial control systems (ICS) generate large volumes of fine-grained operational data. Distributed artificial intelligence offers opportunities for secure collaboration across ICS clients, but devices such as programmable logic controllers (PLCs), remote terminal units (RTUs), and supervisory control and data acquisition (SCADA) systems remain limited to basic control and monitoring, lacking capabilities for intelligent analysis. Privacy concerns further hinder effective data utilization. To address these challenges, we propose a distributed, reliability-enhanced collaborative training framework, termed ICS physical equipment–Edge Server–Blockchain (IEEB), which integrates federated learning, edge computing, and blockchain. Edge servers enable localized model training under resource constraints, while blockchain and smart contracts provide decentralized management, process automation, dynamic incentives, and anti-poisoning mechanisms. IEEB ensures auditability, traceability, and secure collaboration among heterogeneous ICS. Implemented with Hyperledger Fabric and evaluated on five types of PLCs, IEEB improved average AUC by 9.1% on the SWaT, WADI, and MSL datasets, reduced training time by 64.7% and batches by 80.3%, and maintained robustness against up to 20% malicious clients.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112318"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174928","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}

引用次数: 0

Proper orthogonal decomposition-assisted polynomial chaos expansion method for fragility analysis of composite shells 复合材料壳脆性分析的正交分解辅助多项式混沌展开法

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-01-20 DOI: 10.1016/j.ress.2026.112248

S. Chandra, V.A. Matsagar

{"title":"Proper orthogonal decomposition-assisted polynomial chaos expansion method for fragility analysis of composite shells","authors":"S. Chandra, V.A. Matsagar","doi":"10.1016/j.ress.2026.112248","DOIUrl":"10.1016/j.ress.2026.112248","url":null,"abstract":"<div><div>A computationally efficient reduced-order metamodel is developed for the fragility assessment of laminated composite shells by employing proper orthogonal decomposition (POD)-assisted generalized polynomial chaos (gPC) expansion method. The developed adaptive POD-gPC expansion method circumvents the conventional computationally intensive stochastic approach of the spatially distributed nodes on the shell. The lightweight composite structures are used in aircraft, space vehicles, and high-speed trains, and infrastructure. Due to the inevitable uncertainty in the material properties and applied load, the probabilistic vulnerability assessment is essential for the safe design of a composite structure. The fragility functions for symmetric and antisymmetric cross-ply laminated composite plates, cylindrical and spherical shells, and hypar shells are determined by using the newly developed metamodel. It is found that the adaptive parameters, i.e., targeted POD-truncation and error tolerance parameter improve the accuracy of the developed metamodel for predicting the spatial distribution of the stochastic response, whereas the target coefficient of variance limits variability in the calculated probability of failure. The study indicates that geometric parameters, i.e., type of curvature, radius of curvature, and lamina sequence, greatly influence the fragility of the composite structures. A lesser probabilistic vulnerability has been observed for the spherical shells as compared to the cylindrical shells and plates. Nonetheless, hypar shells present least vulnerability as compared to the considered plates and shells.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112248"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174931","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}

引用次数: 0

Vulnerability assessment of storage tanks subject to fires under multi-source uncertainties using a novel ensemble-based non-parametric approach 多源不确定性条件下储罐火灾易损性评估方法研究

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-01-22 DOI: 10.1016/j.ress.2026.112265

Long Ding , Jie Ji , Sining Chen

{"title":"Vulnerability assessment of storage tanks subject to fires under multi-source uncertainties using a novel ensemble-based non-parametric approach","authors":"Long Ding , Jie Ji , Sining Chen","doi":"10.1016/j.ress.2026.112265","DOIUrl":"10.1016/j.ress.2026.112265","url":null,"abstract":"<div><div>Vulnerability assessment of storage tanks under harsh environment, such as fire exposure, encounters various uncertainties which could affect the reliability and credibility of results. A novel ensemble-based non-parametric approach is proposed to quantify uncertainties in vulnerability assessment of storage tanks subject to fires under multi-source uncertainties using tolerance interval. Based on a base model that uses crisp value to characterize vulnerability, the approach firstly obtains the one-sided tolerance interval (upper limit) of time to failure (<em>ttf</em>) and escalation probability (<em>EP</em>) of tank subject to fires using a non-parametric method. Then, using the ensemble-based method, the approach obtains the two-sided tolerance interval of the upper limit of <em>ttf</em> and <em>EP</em> of tank subject to fires. The application of the approach is demonstrated in a case study, which validates the effectiveness of the approach. Besides, the developed approach is validated by comparison with crude Monte Carlo method, and its robustness is also verified. The effects of important parameters on uncertainties of the vulnerability assessment are discussed via sensitivity analysis. The results show that, for the scenario of case study, the relative difference between the lower limit and upper limit are 10.77% and 23.83% of the lower limit, for <em>ttf</em> and <em>EP</em>, respectively, and this reflects the important of uncertainty analysis of vulnerability assessment of storage tanks subject to fires. The proposed approach can serve as an uncertainty analysis tool of vulnerability assessment of storage tanks subject to fires as supplement to commonly used crisp values.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112265"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174761","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}

引用次数: 0

A data-driven system-theoretic Bayesian network framework for probabilistic safety assessment of passenger vessels 客船概率安全评估的数据驱动系统理论贝叶斯网络框架

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-02-02 DOI: 10.1016/j.ress.2026.112350

Luis Alfonso Díaz-Secades, Aitana Sánchez-González

{"title":"A data-driven system-theoretic Bayesian network framework for probabilistic safety assessment of passenger vessels","authors":"Luis Alfonso Díaz-Secades, Aitana Sánchez-González","doi":"10.1016/j.ress.2026.112350","DOIUrl":"10.1016/j.ress.2026.112350","url":null,"abstract":"<div><div>Passenger vessel operations present a high-consequence environment where a paradox has emerged: incident frequency is decreasing, yet catastrophic severity is not. This trend exposes the inadequacy of existing risk models, which are typically localized, and reliant on subjective expert elicitation. This study develops a robust, data-driven risk assessment framework by synergizing System-Theoretic Process Analysis (STPA) with a Bayesian Network (BN), grounded in a novel database of 235 official European accident reports. STPA defines the BN’s causal topology, ensuring theoretical coherence and mitigating the epistemic uncertainty and bias of conventional expert-led modeling. Sensitivity analysis reveals the probabilistic primacy of latent systemic precursors, identifying Structural Failure and Defective Maintenance as dominant risk control points. The analysis moves beyond simplistic attributions of “human error”, revealing how operational failures like COLREGs infringements are symptoms of distinct causal pathways dependent on vessel type and operational conditions. The resulting model is a quantitative instrument that identifies the most probable pathways to catastrophe, offering an objective foundation for transitioning from reactive compliance to proactive, data-driven safety governance.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112350"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174764","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}

引用次数: 0

Travel time resilience of national multimodal transport systems under extreme events: A passenger-oriented framework 极端事件下国家多式联运系统的旅行时间弹性：一个以乘客为导向的框架

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-01-19 DOI: 10.1016/j.ress.2026.112256

Hui Zhang , Min Ouyang , Min Xu , Wenjing Sun , Liu Hong , Duantengchuan Li

{"title":"Travel time resilience of national multimodal transport systems under extreme events: A passenger-oriented framework","authors":"Hui Zhang , Min Ouyang , Min Xu , Wenjing Sun , Liu Hong , Duantengchuan Li","doi":"10.1016/j.ress.2026.112256","DOIUrl":"10.1016/j.ress.2026.112256","url":null,"abstract":"<div><div>National multimodal transport systems (NMTSs) are vital to intercity mobility and economic development, yet increasingly vulnerable to extreme events. Existing resilience assessments largely focus on single-mode or regional-scale systems and fail to address the resilience of NMTSs, considering their large-scale, multi-modal, timetable-dependent nature. In response to this challenge, the paper proposes a passenger-oriented framework to assess the resilience of NMTSs under extreme events. A dynamic functionality network is constructed, integrating scheduled services, intermodal transfers, and time-dependent travel paths. Based on this network, a resilience model is formulated using daily passenger demand and minimum travel time matrices. Two computational methods, a critical nodes-based method and a segment computation-based method, are designed to ensure computational efficiency. The framework is applied to the coupled road, high-speed rail, and airline system in mainland China under COVID-19 lockdowns and a heavy rainstorm in Zhengzhou. Results show that the daily proportion of affected passengers during city lockdowns ranges from 0.01% to 10.07%, with impacts strongly correlated with city GDP and network size. In the rainstorm case, 6.44% of passengers were affected, with an average delay of 167.12 min. The proposed framework offers a transferable and computationally efficient approach for national-scale resilience assessment and transport planning.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112256"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174843","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}

引用次数: 0

Analysis of systems with dependent components through a variance-based index and regression importance signature 通过基于方差的指标和回归重要性特征对具有相关成分的系统进行分析

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-02-06 DOI: 10.1016/j.ress.2026.112357

Antonio Di Crescenzo , Giulia Pisano , Alfonso Suárez-Llorens

引用次数: 0

Remaining useful life prediction of aero-engine using pyramid temporal convolutional network with fused complementary attention 基于融合互补注意力的金字塔时间卷积网络的航空发动机剩余使用寿命预测

IF 11 1区工程技术

Reliability Engineering & System Safety Pub Date : 2026-09-01 Epub Date: 2026-01-19 DOI: 10.1016/j.ress.2026.112254

Kun Wang , Ai He , Jiashuai Liu , Qifan Zhou , Zhongzhi Hu , Jinshen Wang

{"title":"Remaining useful life prediction of aero-engine using pyramid temporal convolutional network with fused complementary attention","authors":"Kun Wang , Ai He , Jiashuai Liu , Qifan Zhou , Zhongzhi Hu , Jinshen Wang","doi":"10.1016/j.ress.2026.112254","DOIUrl":"10.1016/j.ress.2026.112254","url":null,"abstract":"<div><div>Accurate remaining useful life (RUL) prediction for aero-engines is a key technology for prognostics and health management (PHM). However, challenges such as cross-domain and operating condition adaptation, personalized adaptation, and modeling of temporal-spatial coupling characteristics hinder the reliability of existing RUL prediction methods. To address these, this paper presents a novel RUL prediction framework—pyramid temporal convolutional network with fused complementary attention (PTCN-FCA). The framework employs a multi-scale pyramidal temporal convolutional network to systematically extract state-related features and dynamic patterns, constructing a robust baseline model. Personalized parameter adaptation with a lightweight auxiliary branch enables efficient individualized models, and a degradation pattern-switching mechanism triggers RUL prediction. A complementary attention mechanism enhances the model’s ability to capture critical degradation dynamics. Experiments on the N<img>CMAPSS and real-world quick access recorder datasets validate the effectiveness, showing significant improvements in accuracy and generalization, with Scores of 0.33, 0.41, and 0.39 on DS01, DS04, and DS06, respectively. These results demonstrate the practical potential of the proposed approach for aero-engine health management.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"273 ","pages":"Article 112254"},"PeriodicalIF":11.0,"publicationDate":"2026-09-01","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"146174950","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}

引用次数: 0