Reliability Engineering & System Safety最新文献_第3页

Multivariate degradation modeling using generalized cauchy process and application in life prediction of dye-sensitized solar cells 基于广义柯西过程的多元退化建模及其在染料敏化太阳能电池寿命预测中的应用

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-17 DOI: 10.1016/j.ress.2024.110651

Ali Asgari , Wujun Si , Wei Wei , Krishna Krishnan , Kunpeng Liu

Recently, the Generalized Cauchy (GC) process has been applied to capture a Long Memory (LM) phenomenon in product degradation modeling and life prediction. Compared with the traditional fractional Brownian motion that captures the LM using a single Hurst parameter, the GC process has two free parameters (Hurst and fractal dimension parameters) that flexibly capture both global LM and local irregularity. However, all existing GC-based degradation models are for a single Degradation Characteristic (DC). In this article, motivated by a real degradation problem of dye-sensitized solar cells that jointly exhibits multiple DCs, global LM, local irregularity and DC-wise cross-correlation, we propose a novel GC-based Multivariate Degradation Model (GC-MDM) to simultaneously capture the aforementioned effects. A maximum likelihood estimation approach is developed to estimate parameters of the GC-MDM. Subsequently, product life prediction based on the GC-MDM is developed. The proposed GC-MDM is validated through a simulation study and a physical experiment of dye-sensitized solar cells. Results show that the proposed GC-MDM fundamentally improves the life prediction accuracy in comparison with conventional degradation models which significantly misestimate the uncertainty of product life.

近年来，广义柯西（GC）过程被应用于产品降解建模和寿命预测中，以捕捉长记忆（LM）现象。与传统分数阶布朗运动使用单个Hurst参数捕获LM相比，GC过程具有两个自由参数（Hurst和分形维数参数），可以灵活地捕获全局LM和局部不规则性。然而，所有现有的基于gc的降解模型都是针对单一的降解特性（DC）。在本文中，基于染料敏化太阳能电池的实际降解问题，我们提出了一种新的基于gc的多元降解模型（GC-MDM），以同时捕捉上述效应。提出了一种最大似然估计方法来估计GC-MDM的参数。随后，提出了基于GC-MDM的产品寿命预测方法。通过染料敏化太阳能电池的模拟研究和物理实验验证了所提出的GC-MDM。结果表明，与传统的降解模型相比，GC-MDM从根本上提高了寿命预测的精度，而传统的降解模型严重错误地估计了产品寿命的不确定性。

{"title":"Multivariate degradation modeling using generalized cauchy process and application in life prediction of dye-sensitized solar cells","authors":"Ali Asgari , Wujun Si , Wei Wei , Krishna Krishnan , Kunpeng Liu","doi":"10.1016/j.ress.2024.110651","DOIUrl":"10.1016/j.ress.2024.110651","url":null,"abstract":"<div><div>Recently, the Generalized Cauchy (GC) process has been applied to capture a Long Memory (LM) phenomenon in product degradation modeling and life prediction. Compared with the traditional fractional Brownian motion that captures the LM using a single Hurst parameter, the GC process has two free parameters (Hurst and fractal dimension parameters) that flexibly capture both global LM and local irregularity. However, all existing GC-based degradation models are for a single Degradation Characteristic (DC). In this article, motivated by a real degradation problem of dye-sensitized solar cells that jointly exhibits multiple DCs, global LM, local irregularity and DC-wise cross-correlation, we propose a novel GC-based Multivariate Degradation Model (GC-MDM) to simultaneously capture the aforementioned effects. A maximum likelihood estimation approach is developed to estimate parameters of the GC-MDM. Subsequently, product life prediction based on the GC-MDM is developed. The proposed GC-MDM is validated through a simulation study and a physical experiment of dye-sensitized solar cells. Results show that the proposed GC-MDM fundamentally improves the life prediction accuracy in comparison with conventional degradation models which significantly misestimate the uncertainty of product life.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110651"},"PeriodicalIF":9.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744675","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 dynamic collaborative adversarial domain adaptation network for unsupervised rotating machinery fault diagnosis 无监督旋转机械故障诊断的动态协同对抗域自适应网络

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-17 DOI: 10.1016/j.ress.2024.110662

Xin Wang, Hongkai Jiang, Mingzhe Mu, Yutong Dong

Acquiring sufficient fault data labels for new tasks in rotating machinery fault diagnosis is tricky. Accurately identifying faults in unlabeled scenarios is a critical and urgent practical need. Unsupervised domain adaptation (UDA) is a mainstream method to address this issue. However, most existing UDA models are static and struggle to dynamically adjust according to changes in the target task, resulting in limited diagnostic performance. To address this limitation, a dynamic collaborative adversarial domain adaptation network (DCADAN) is proposed for unsupervised rotating machinery fault diagnosis. Firstly, a multi-objective dynamic collaborative generator is designed to endow with dynamic characteristics for adjusting its own architecture, enhancing the capture capability of key domain adaptation features. Secondly, a dual-system dynamic collaborative adversarial mode is established to dynamically adjust the network training architecture, forming task-oriented refined diagnostic decision edges to steadily improve domain adaptation diagnostic capability. Finally, a multi-source domain dynamic collaborative loss is developed to match the force of multiple source domains, forming an efficient collaborative diagnostic pattern with dynamic adjustment across multi-source domains. Two case studies indicate that DCADAN demonstrates superior diagnostic performance when executing cross-domain diagnosis tasks without target labels.

在旋转机械故障诊断中，获取足够的故障数据标签是一个难点。准确识别未标记场景中的故障是一项关键而紧迫的实际需求。无监督域自适应（UDA）是解决这一问题的主流方法。然而，现有的UDA模型大多是静态的，难以根据目标任务的变化进行动态调整，导致诊断性能有限。针对这一局限性，提出了一种用于无监督旋转机械故障诊断的动态协同对抗域自适应网络（DCADAN）。首先，设计了多目标动态协同生成器，赋予其动态特性以调整自身结构，增强了关键领域自适应特征的捕获能力；其次，建立双系统动态协同对抗模式，动态调整网络训练架构，形成面向任务的精细化诊断决策边，稳步提高领域自适应诊断能力；最后，提出了一种多源域动态协同损失模型来匹配多源域的力，形成了一种跨多源域动态调整的高效协同诊断模式。两个案例研究表明，DCADAN在执行无目标标签的跨域诊断任务时表现出优异的诊断性能。

{"title":"A dynamic collaborative adversarial domain adaptation network for unsupervised rotating machinery fault diagnosis","authors":"Xin Wang, Hongkai Jiang, Mingzhe Mu, Yutong Dong","doi":"10.1016/j.ress.2024.110662","DOIUrl":"10.1016/j.ress.2024.110662","url":null,"abstract":"<div><div>Acquiring sufficient fault data labels for new tasks in rotating machinery fault diagnosis is tricky. Accurately identifying faults in unlabeled scenarios is a critical and urgent practical need. Unsupervised domain adaptation (UDA) is a mainstream method to address this issue. However, most existing UDA models are static and struggle to dynamically adjust according to changes in the target task, resulting in limited diagnostic performance. To address this limitation, a dynamic collaborative adversarial domain adaptation network (DCADAN) is proposed for unsupervised rotating machinery fault diagnosis. Firstly, a multi-objective dynamic collaborative generator is designed to endow with dynamic characteristics for adjusting its own architecture, enhancing the capture capability of key domain adaptation features. Secondly, a dual-system dynamic collaborative adversarial mode is established to dynamically adjust the network training architecture, forming task-oriented refined diagnostic decision edges to steadily improve domain adaptation diagnostic capability. Finally, a multi-source domain dynamic collaborative loss is developed to match the force of multiple source domains, forming an efficient collaborative diagnostic pattern with dynamic adjustment across multi-source domains. Two case studies indicate that DCADAN demonstrates superior diagnostic performance when executing cross-domain diagnosis tasks without target labels.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110662"},"PeriodicalIF":9.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744555","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

Lifting wavelet-informed hierarchical domain adaptation network: An interpretable digital twin-driven gearbox fault diagnosis method 提升小波信息分层域自适应网络：可解释的数字孪生驱动齿轮箱故障诊断方法

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-17 DOI: 10.1016/j.ress.2024.110660

Sixiang Jia, Dingyi Sun, Khandaker Noman, Xin Wang, Yongbo Li

Digital twin (DT) has served as a dependable technology for supplementing reliable simulated fault data in gearbox fault diagnosis. However, the vast data distribution discrepancy and insufficient interpretability still significantly limit the industrial application of DT-driven fault diagnosis methods. To solve these problems, a lifting wavelet-informed hierarchical domain adaptation network (LHDAN) is proposed for transferring the diagnostic knowledge between the physical gearbox and DT model. LHDAN improves the interpretability of diagnostic knowledge transfer in terms of parameter initialization, physical constraints on the training process, and feature distribution adaptation. Specifically, LHDAN utilizes a lifting wavelet-informed convolutional neural network (LW-Conv) to mimic the cascade structure of lifting wavelet decomposition, in which the fully learnable prediction and update operators are initialized with existing wavelet bases and further constrained with high-pass and low-pass filters in the training process. Furthermore, a kurtosis-guided attention mechanism is proposed to fuse hierarchical features with diverse transferabilities flexibly. Finally, the fused hierarchical features of the actual gearbox and DT model are explicitly aligned to eliminate the feature distribution discrepancies. A high-fidelity DT model is established based on an industrial gearbox fault test bench. Compared to several state-of-the-art models, LHDAN demonstrates superior interpretability and diagnostic performance.

在齿轮箱故障诊断中，数字孪生（DT）是补充可靠模拟故障数据的可靠技术。然而，巨大的数据分布差异和不足的可解释性仍然极大地限制了 DT 驱动的故障诊断方法的工业应用。为了解决这些问题，我们提出了一种提升小波信息分层域自适应网络（LHDAN），用于在物理齿轮箱和 DT 模型之间传递诊断知识。LHDAN 在参数初始化、训练过程的物理约束和特征分布适应等方面提高了诊断知识转移的可解释性。具体来说，LHDAN 利用提升小波卷积神经网络（LW-Conv）模仿提升小波分解的级联结构，其中完全可学习的预测和更新算子由现有的小波基初始化，并在训练过程中进一步使用高通和低通滤波器进行约束。此外，还提出了一种峰度引导的关注机制，以灵活地融合具有不同转移能力的分层特征。最后，对实际齿轮箱和 DT 模型的融合分层特征进行明确对齐，以消除特征分布差异。基于工业齿轮箱故障测试台建立了高保真 DT 模型。与几种最先进的模型相比，LHDAN 的可解释性和诊断性能更胜一筹。

{"title":"Lifting wavelet-informed hierarchical domain adaptation network: An interpretable digital twin-driven gearbox fault diagnosis method","authors":"Sixiang Jia, Dingyi Sun, Khandaker Noman, Xin Wang, Yongbo Li","doi":"10.1016/j.ress.2024.110660","DOIUrl":"10.1016/j.ress.2024.110660","url":null,"abstract":"<div><div>Digital twin (DT) has served as a dependable technology for supplementing reliable simulated fault data in gearbox fault diagnosis. However, the vast data distribution discrepancy and insufficient interpretability still significantly limit the industrial application of DT-driven fault diagnosis methods. To solve these problems, a lifting wavelet-informed hierarchical domain adaptation network (LHDAN) is proposed for transferring the diagnostic knowledge between the physical gearbox and DT model. LHDAN improves the interpretability of diagnostic knowledge transfer in terms of parameter initialization, physical constraints on the training process, and feature distribution adaptation. Specifically, LHDAN utilizes a lifting wavelet-informed convolutional neural network (LW-Conv) to mimic the cascade structure of lifting wavelet decomposition, in which the fully learnable prediction and update operators are initialized with existing wavelet bases and further constrained with high-pass and low-pass filters in the training process. Furthermore, a kurtosis-guided attention mechanism is proposed to fuse hierarchical features with diverse transferabilities flexibly. Finally, the fused hierarchical features of the actual gearbox and DT model are explicitly aligned to eliminate the feature distribution discrepancies. A high-fidelity DT model is established based on an industrial gearbox fault test bench. Compared to several state-of-the-art models, LHDAN demonstrates superior interpretability and diagnostic performance.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110660"},"PeriodicalIF":9.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697253","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

Risk analysis of weather-related railroad accidents in the United States 美国与天气有关的铁路事故风险分析

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-17 DOI: 10.1016/j.ress.2024.110647

Zhipeng Zhang , Chen-Yu Lin

Global climate change has led to more frequent extreme weather events in recent decades, and this has impacted critical infrastructure such as railway systems. Although train accidents caused by extreme weather events are not uncommon, the level of their risk is foreseen to rise to an unacceptable level. As a result, proper data collection and analysis for train accident related to extreme weather are pertinent to developing an effective railway climate change adaptation plan. This paper presents a comprehensive and quantitative analysis of weather-related railroad accidents in the United States. The analysis comprises time series, spatial, and causal elements to understand the temporal trends of weather-related railroad accidents, the predominant type of weather causes, and the effect of regional meteorological characteristics on them. The results showed that the likelihood of weather-related railroad accidents varies by meteorological regions and does not show a clear increasing or decreasing trend, but their above-average severity indicates opportunities to mitigate the risk in light of the projected increasing frequency. Results of this research contribute to better understanding of railway extreme weather risk and serve as a foundation for future research that addresses the effect of climate change on railroad system and develops proper railway climate change adaptation plans.

近几十年来，全球气候变化导致极端天气事件更加频繁，这影响了铁路系统等关键基础设施。虽然由极端天气事件引起的火车事故并不罕见，但预计其风险水平将上升到不可接受的水平。因此，对与极端天气有关的列车事故进行适当的数据收集和分析，对于制定有效的铁路气候变化适应计划至关重要。本文对美国与天气有关的铁路事故进行了全面和定量的分析。分析包括时间序列、空间和因果因素，以了解与天气有关的铁路事故的时间趋势、主要的天气原因类型以及区域气象特征对天气事故的影响。结果表明，与天气有关的铁路事故的可能性因气象区域而异，没有明显的增加或减少趋势，但其高于平均水平的严重程度表明，根据预测的频率增加，有机会减轻风险。本研究结果有助于更好地理解铁路极端天气风险，并为未来研究气候变化对铁路系统的影响和制定适当的铁路气候变化适应计划奠定基础。

{"title":"Risk analysis of weather-related railroad accidents in the United States","authors":"Zhipeng Zhang , Chen-Yu Lin","doi":"10.1016/j.ress.2024.110647","DOIUrl":"10.1016/j.ress.2024.110647","url":null,"abstract":"<div><div>Global climate change has led to more frequent extreme weather events in recent decades, and this has impacted critical infrastructure such as railway systems. Although train accidents caused by extreme weather events are not uncommon, the level of their risk is foreseen to rise to an unacceptable level. As a result, proper data collection and analysis for train accident related to extreme weather are pertinent to developing an effective railway climate change adaptation plan. This paper presents a comprehensive and quantitative analysis of weather-related railroad accidents in the United States. The analysis comprises time series, spatial, and causal elements to understand the temporal trends of weather-related railroad accidents, the predominant type of weather causes, and the effect of regional meteorological characteristics on them. The results showed that the likelihood of weather-related railroad accidents varies by meteorological regions and does not show a clear increasing or decreasing trend, but their above-average severity indicates opportunities to mitigate the risk in light of the projected increasing frequency. Results of this research contribute to better understanding of railway extreme weather risk and serve as a foundation for future research that addresses the effect of climate change on railroad system and develops proper railway climate change adaptation plans.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110647"},"PeriodicalIF":9.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744674","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

Emergency Response Inference Mapping (ERIMap): A Bayesian network-based method for dynamic observation processing 应急响应推理映射（ERIMap）：一种基于贝叶斯网络的动态观测处理方法

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-17 DOI: 10.1016/j.ress.2024.110640

Moritz Schneider , Lukas Halekotte , Tina Comes , Daniel Lichte , Frank Fiedrich

In emergencies, high stake decisions often have to be made under time pressure and strain. In order to support such decisions, information from various sources needs to be collected and processed rapidly. The information available tends to be temporally and spatially variable, uncertain, and sometimes conflicting, leading to potential biases in decisions. Currently, there is a lack of systematic approaches for information processing and situation assessment which meet the particular demands of emergency situations. To address this gap, we present a Bayesian network-based method called ERIMap that is tailored to the complex information-scape during emergencies. The method enables the systematic and rapid processing of heterogeneous and potentially uncertain observations and draws inferences about key variables of an emergency. It thereby reduces complexity and cognitive load for decision makers. The output of the ERIMap method is a dynamically evolving and spatially resolved map of beliefs about key variables of an emergency that is updated each time a new observation becomes available. The method is illustrated in a case study in which an emergency response is triggered by an accident causing a gas leakage on a chemical plant site.

在紧急情况下，高风险的决策往往必须在时间压力和紧张的情况下做出。为了支持这种决定，需要迅速收集和处理来自各种来源的信息。可获得的信息往往在时间和空间上都是可变的、不确定的，有时还相互冲突，从而导致决策中的潜在偏见。目前，缺乏系统的处理信息和评估情况的办法，以满足紧急情况的特殊需要。为了解决这一差距，我们提出了一种基于贝叶斯网络的方法，称为ERIMap，该方法是为紧急情况下的复杂信息环境量身定制的。该方法能够系统和快速地处理异构和潜在不确定的观测结果，并对紧急情况的关键变量进行推断。从而降低了决策者的复杂性和认知负荷。ERIMap方法的输出是关于紧急情况关键变量的动态演变和空间分解的信念图，每次有新的观测数据可用时都会更新。该方法在一个案例研究中得到说明，该案例研究是由化学工厂现场发生的事故引起的气体泄漏引发的应急响应。

{"title":"Emergency Response Inference Mapping (ERIMap): A Bayesian network-based method for dynamic observation processing","authors":"Moritz Schneider , Lukas Halekotte , Tina Comes , Daniel Lichte , Frank Fiedrich","doi":"10.1016/j.ress.2024.110640","DOIUrl":"10.1016/j.ress.2024.110640","url":null,"abstract":"<div><div>In emergencies, high stake decisions often have to be made under time pressure and strain. In order to support such decisions, information from various sources needs to be collected and processed rapidly. The information available tends to be temporally and spatially variable, uncertain, and sometimes conflicting, leading to potential biases in decisions. Currently, there is a lack of systematic approaches for information processing and situation assessment which meet the particular demands of emergency situations. To address this gap, we present a Bayesian network-based method called <em>ERIMap</em> that is tailored to the complex information-scape during emergencies. The method enables the systematic and rapid processing of heterogeneous and potentially uncertain observations and draws inferences about key variables of an emergency. It thereby reduces complexity and cognitive load for decision makers. The output of the <em>ERIMap</em> method is a dynamically evolving and spatially resolved map of beliefs about key variables of an emergency that is updated each time a new observation becomes available. The method is illustrated in a case study in which an emergency response is triggered by an accident causing a gas leakage on a chemical plant site.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110640"},"PeriodicalIF":9.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744553","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"OA","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}

引用次数: 0

Repairing smarter: Opportunistic maintenance for a closed-loop supply chain with spare parts dependency 更智能的维修：有备件依赖的闭环供应链的机会性维护

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-17 DOI: 10.1016/j.ress.2024.110642

Abdelhamid Boujarif , David W. Coit , Oualid Jouini , Zhiguo Zeng , Robert Heidsieck

Adopting a closed-loop supply chain enhances spare part provisioning through repair, remanufacturing, and recycling. However, poor maintenance of components can have severe consequences. Unlike traditional opportunistic maintenance methods that assume regular inspections or precise degradation monitoring, we propose a model that leverages historical repair data to replace worn components preventively. It considers the real-world workflow where parts are often restored only to a functional level. We study maintenance strategies for repeatedly repaired multi-component systems by applying preventive operations only during corrective repairs. Our model considers component ages, failure time distributions, and structural and economic dependencies, favoring collective over individual replacements for cost efficiency. Stochastic dependencies are mapped using Nataf transformation for component subsets, and a genetic algorithm identifies optimal maintenance strategies to reduce long-term operational costs by balancing maintenance against potential failure penalties. We demonstrate the effectiveness of our approach with a case study on MRI power supply machines, showing that preventive actions can cut early life failures by up to 50% and extend useful life by over a year. Sensitivity analysis reveals that logistic costs, interest rates, and planning horizons influence decisions. Opportunistic maintenance can reduce logistic costs and increase the lifetime of spare parts after repair. Integrating stochastic dependency is computationally efficient for industrial systems and can help predict failures more accurately.

采用闭环供应链，通过维修、再制造和回收提高备件供应。然而，对组件的不良维护可能会产生严重的后果。与传统的机会性维护方法（假设定期检查或精确的退化监测）不同，我们提出了一个利用历史维修数据预防性更换磨损部件的模型。它考虑了现实世界的工作流，其中部件通常只恢复到功能级别。我们研究了重复修复的多部件系统的维护策略，仅在纠正维修期间应用预防性操作。我们的模型考虑了组件的使用年限、故障时间分布、结构和经济依赖关系，为了成本效率，更倾向于集体替换而不是单个替换。随机依赖关系使用Nataf转换映射到组件子集，遗传算法确定最佳维护策略，通过平衡维护和潜在故障惩罚来降低长期运营成本。我们通过MRI电源机器的案例研究证明了我们方法的有效性，表明预防措施可以减少高达50%的早期故障，并将使用寿命延长一年以上。敏感性分析表明，物流成本、利率和规划范围会影响决策。机制化维修可以降低物流成本，增加维修后备件的使用寿命。集成随机依赖对工业系统来说计算效率很高，可以帮助更准确地预测故障。

{"title":"Repairing smarter: Opportunistic maintenance for a closed-loop supply chain with spare parts dependency","authors":"Abdelhamid Boujarif , David W. Coit , Oualid Jouini , Zhiguo Zeng , Robert Heidsieck","doi":"10.1016/j.ress.2024.110642","DOIUrl":"10.1016/j.ress.2024.110642","url":null,"abstract":"<div><div>Adopting a closed-loop supply chain enhances spare part provisioning through repair, remanufacturing, and recycling. However, poor maintenance of components can have severe consequences. Unlike traditional opportunistic maintenance methods that assume regular inspections or precise degradation monitoring, we propose a model that leverages historical repair data to replace worn components preventively. It considers the real-world workflow where parts are often restored only to a functional level. We study maintenance strategies for repeatedly repaired multi-component systems by applying preventive operations only during corrective repairs. Our model considers component ages, failure time distributions, and structural and economic dependencies, favoring collective over individual replacements for cost efficiency. Stochastic dependencies are mapped using Nataf transformation for component subsets, and a genetic algorithm identifies optimal maintenance strategies to reduce long-term operational costs by balancing maintenance against potential failure penalties. We demonstrate the effectiveness of our approach with a case study on MRI power supply machines, showing that preventive actions can cut early life failures by up to 50% and extend useful life by over a year. Sensitivity analysis reveals that logistic costs, interest rates, and planning horizons influence decisions. Opportunistic maintenance can reduce logistic costs and increase the lifetime of spare parts after repair. Integrating stochastic dependency is computationally efficient for industrial systems and can help predict failures more accurately.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110642"},"PeriodicalIF":9.4,"publicationDate":"2024-11-17","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744665","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

Condition-based maintenance for multi-state systems with prognostic and deep reinforcement learning 基于预测和深度强化学习的多状态系统状态维护

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-16 DOI: 10.1016/j.ress.2024.110659

Huixian Zhang, Xiukun Wei, Zhiqiang Liu, Yaning Ding, Qingluan Guan

The utilization of prognostic information in practical engineering is increasing with the development of technology and predictive modeling. Current research on maintenance strategies for complex multi-state systems often neglects prognostic information or assumes complete availability of all component information. This paper investigates the joint maintenance strategies based on condition-based maintenance for complex multi-state systems, in which the predicted remaining useful life of some components is known. Firstly, a maintenance strategy framework is developed and the joint maintenance strategy is proposed for the studied problem. Then the deterioration process of the component, the imperfect maintenance, and prediction error models are constructed. The optimization problem is modeled as a Markov Decision Process to minimize the maintenance cost, and the system reliability constraints are established by using the universal generating function method. In addition, a deep Q-network is designed to solve the optimal maintenance policy. Finally, the traction system of a metro train is taken as an example to verify the applicability of the model and algorithm. The results show that the proposed maintenance strategy reduces the maintenance cost compared to the current maintenance strategy for both fixed maintenance intervals and dynamic maintenance intervals.

随着技术和预测建模的发展，预测信息在实际工程中的应用越来越多。当前对复杂多状态系统维护策略的研究往往忽略了预测信息或假设所有部件信息完全可用。研究了已知部件预测剩余使用寿命的复杂多状态系统的基于状态维护的联合维护策略。首先，建立了维修策略框架，并针对所研究的问题提出了联合维修策略。然后建立了构件劣化过程、不完善维修和预测误差模型。以维护成本最小化为目标，将优化问题建模为马尔可夫决策过程，采用通用生成函数法建立系统可靠性约束。此外，还设计了一个深度q网络来解决最优维护策略问题。最后，以某地铁列车牵引系统为例，验证了模型和算法的适用性。结果表明，与现有的固定维修间隔和动态维修间隔的维修策略相比，所提出的维修策略降低了维修成本。

{"title":"Condition-based maintenance for multi-state systems with prognostic and deep reinforcement learning","authors":"Huixian Zhang, Xiukun Wei, Zhiqiang Liu, Yaning Ding, Qingluan Guan","doi":"10.1016/j.ress.2024.110659","DOIUrl":"10.1016/j.ress.2024.110659","url":null,"abstract":"<div><div>The utilization of prognostic information in practical engineering is increasing with the development of technology and predictive modeling. Current research on maintenance strategies for complex multi-state systems often neglects prognostic information or assumes complete availability of all component information. This paper investigates the joint maintenance strategies based on condition-based maintenance for complex multi-state systems, in which the predicted remaining useful life of some components is known. Firstly, a maintenance strategy framework is developed and the joint maintenance strategy is proposed for the studied problem. Then the deterioration process of the component, the imperfect maintenance, and prediction error models are constructed. The optimization problem is modeled as a Markov Decision Process to minimize the maintenance cost, and the system reliability constraints are established by using the universal generating function method. In addition, a deep Q-network is designed to solve the optimal maintenance policy. Finally, the traction system of a metro train is taken as an example to verify the applicability of the model and algorithm. The results show that the proposed maintenance strategy reduces the maintenance cost compared to the current maintenance strategy for both fixed maintenance intervals and dynamic maintenance intervals.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110659"},"PeriodicalIF":9.4,"publicationDate":"2024-11-16","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744551","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

Multi-label domain adversarial reinforcement learning for unsupervised compound fault recognition 用于无监督复合故障识别的多标签域对抗强化学习

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-15 DOI: 10.1016/j.ress.2024.110638

Zisheng Wang , Jianping Xuan , Tielin Shi , Yan-Fu Li

Compound fault composed of coinstantaneous multiple faults frequently causes the failure of a manufacturing system, which greatly reduces the reliability. When measuring the compound fault, two difficulties generally exist: (1) the complex correlation between different single faults, and (2) collected target samples without labels. To accomplish the cross-domain unsupervised compound fault recognition, this study proposes a multi-label domain adversarial reinforcement learning (ML-DARL) framework that implements two multi-label deep reinforcement learning (ML-DRL) models with adversarial domain adaptation. First, a source ML-DRL model is adopted to train a source feature network (SFN) and a policy network by using a dataset with labels (source domain). Then, a discriminator and a target ML-DRL model that includes a target feature network (TFN) are jointly trained with adversarial adaptation by simultaneously using the dataset without labels (target domain) and the source domain. Specifically, two outputs of TFN and SFN are regarded as fake and real components, respectively. Notably, the reward function in the target ML-DRL model is related inversely to the output of the discriminator for the fake component. Finally, a cross-speed case and a cross-location case are executed to verify the adaptation ability of the proposed method on cross-domain unsupervised compound fault recognition.

由同时发生的多个故障组成的复合故障经常导致制造系统失灵，从而大大降低了可靠性。在测量复合故障时，一般存在两个困难：（1）不同单一故障之间存在复杂的相关性；（2）采集的目标样本没有标签。为了实现跨域无监督复合故障识别，本研究提出了一种多标签域对抗强化学习（ML-DARL）框架，该框架实现了两个具有对抗域适应性的多标签深度强化学习（ML-DRL）模型。首先，采用源 ML-DRL 模型，利用带标签的数据集（源域）训练源特征网络（SFN）和策略网络。然后，通过同时使用无标签数据集（目标域）和源域，联合训练包含目标特征网络（TFN）的判别器和目标 ML-DRL 模型。具体来说，TFN 和 SFN 的两个输出分别被视为假成分和真成分。值得注意的是，目标 ML-DRL 模型中的奖励函数与假分量判别器的输出成反比关系。最后，通过跨速度案例和跨位置案例验证了所提方法对跨域无监督复合故障识别的适应能力。

{"title":"Multi-label domain adversarial reinforcement learning for unsupervised compound fault recognition","authors":"Zisheng Wang , Jianping Xuan , Tielin Shi , Yan-Fu Li","doi":"10.1016/j.ress.2024.110638","DOIUrl":"10.1016/j.ress.2024.110638","url":null,"abstract":"<div><div>Compound fault composed of coinstantaneous multiple faults frequently causes the failure of a manufacturing system, which greatly reduces the reliability. When measuring the compound fault, two difficulties generally exist: (1) the complex correlation between different single faults, and (2) collected target samples without labels. To accomplish the cross-domain unsupervised compound fault recognition, this study proposes a multi-label domain adversarial reinforcement learning (ML-DARL) framework that implements two multi-label deep reinforcement learning (ML-DRL) models with adversarial domain adaptation. First, a source ML-DRL model is adopted to train a source feature network (SFN) and a policy network by using a dataset with labels (source domain). Then, a discriminator and a target ML-DRL model that includes a target feature network (TFN) are jointly trained with adversarial adaptation by simultaneously using the dataset without labels (target domain) and the source domain. Specifically, two outputs of TFN and SFN are regarded as fake and real components, respectively. Notably, the reward function in the target ML-DRL model is related inversely to the output of the discriminator for the fake component. Finally, a cross-speed case and a cross-location case are executed to verify the adaptation ability of the proposed method on cross-domain unsupervised compound fault recognition.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110638"},"PeriodicalIF":9.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697250","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

Service risk evaluation of telecommunication core network: A perspective of routing resilience 电信核心网业务风险评估：路由弹性的视角

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-15 DOI: 10.1016/j.ress.2024.110629

Zongqi Xue, Zhenglin Liang

The increasing dependence on telecommunication networks for digitalization raises apprehensions about their vulnerability to disruptions, potentially impacting service performance. Classic risk analysis for telecommunication networks often overlooks the compensating effect of rerouting and its resulting resilience. Our study is devoted to conducting an evaluation of service risk within telecommunication core networks, with a particular emphasis on rerouting resilience. Telecommunication core networks operate on a large scale, where various autonomous systems interconnect through pivotal anchor points. Its routing is typically governed by multiple protocols, including Interior Gateway Protocol and Border Gateway Protocol. We apply percolation theory and Monte Carlo Tree Search to effectively analyze the potential service risk and sequential rerouting effects caused by disruptions marked by a noteworthy probability and substantial impact. Our findings highlight the dual nature of rerouting: it enhances service connectivity while potentially increasing service flows in nearby routers, leading to heightened time delays and packet loss ratios. To comprehensively assess network service risk considering this double-edged sword impact, we devise a performance vector covering service connectivity, time delay, and packet loss across various data transmissions and disruptions. The overall approach is tested on two telecommunication core networks.

越来越多地依赖电信网络实现数字化，这引发了人们对电信网络易受中断影响的担忧，这可能会影响服务性能。传统的电信网络风险分析往往忽略了路由变更的补偿效应及其产生的弹性。我们的研究致力于对电信核心网络中的服务风险进行评估，特别强调重新路由弹性。电信核心网络在大规模运行，其中各种自治系统通过关键锚点相互连接。它的路由通常由多个协议控制，包括内部网关协议和边界网关协议。我们运用渗透理论和蒙特卡罗树搜索，有效地分析了具有显著概率和重大影响的中断所引起的潜在服务风险和顺序改道效应。我们的研究结果强调了重路由的双重性质：它增强了服务连接，同时潜在地增加了附近路由器的业务流，导致时间延迟和丢包率升高。考虑到这种双刃剑影响，为了全面评估网络服务风险，我们设计了一个性能向量，涵盖了各种数据传输和中断的服务连接、时间延迟和数据包丢失。在两个电信核心网络上对整个方法进行了测试。

{"title":"Service risk evaluation of telecommunication core network: A perspective of routing resilience","authors":"Zongqi Xue, Zhenglin Liang","doi":"10.1016/j.ress.2024.110629","DOIUrl":"10.1016/j.ress.2024.110629","url":null,"abstract":"<div><div>The increasing dependence on telecommunication networks for digitalization raises apprehensions about their vulnerability to disruptions, potentially impacting service performance. Classic risk analysis for telecommunication networks often overlooks the compensating effect of rerouting and its resulting resilience. Our study is devoted to conducting an evaluation of service risk within telecommunication core networks, with a particular emphasis on rerouting resilience. Telecommunication core networks operate on a large scale, where various autonomous systems interconnect through pivotal anchor points. Its routing is typically governed by multiple protocols, including Interior Gateway Protocol and Border Gateway Protocol. We apply percolation theory and Monte Carlo Tree Search to effectively analyze the potential service risk and sequential rerouting effects caused by disruptions marked by a noteworthy probability and substantial impact. Our findings highlight the dual nature of rerouting: it enhances service connectivity while potentially increasing service flows in nearby routers, leading to heightened time delays and packet loss ratios. To comprehensively assess network service risk considering this double-edged sword impact, we devise a performance vector covering service connectivity, time delay, and packet loss across various data transmissions and disruptions. The overall approach is tested on two telecommunication core networks.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110629"},"PeriodicalIF":9.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744679","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 novel sim2real reinforcement learning algorithm for process control 用于过程控制的新型模拟真实强化学习算法

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-15 DOI: 10.1016/j.ress.2024.110639

Huiping Liang , Junyao Xie , Biao Huang , Yonggang Li , Bei Sun , Chunhua Yang

While reinforcement learning (RL) has potential in advanced process control and optimization, its direct interaction with real industrial processes can pose safety concerns. Model-based pre-training of RL may alleviate such risks. However, the intricate nature of industrial processes complicates the establishment of entirely accurate simulation models. Consequently, RL-based controllers relying on simulation models can easily suffer from model-plant mismatch. On the one hand, utilizing offline data for pre-training of RL can also mitigate safety risks. However, it requires well-represented historical datasets. This is demanding because industrial processes mostly run under a regulatory mode with basic controllers. To handle these issues, this paper proposes a novel sim2real reinforcement learning algorithm. First, a state adaptor (SA) is proposed to align simulated states with real states to mitigate the model-plant mismatch. Then, a fix-horizon return is designed to replace traditional infinite-step return to provide genuine labels for the critic network, enhancing learning efficiency and stability. Finally, applying proximal policy optimization (PPO), the SA-PPO method is introduced to implement the proposed sim2real algorithm. Experimental results show that SA-PPO improves performance in MSE by 1.96% and in R by 21.64% on average for roasting process simulation. This verifies the effectiveness of the proposed method.

虽然强化学习（RL）在先进过程控制和优化方面具有潜力，但它与实际工业过程的直接互动可能会带来安全问题。对 RL 进行基于模型的预训练可以减轻这种风险。然而，工业流程的复杂性使得建立完全精确的模拟模型变得复杂。因此，基于仿真模型的 RL 控制器很容易出现模型与工厂不匹配的问题。一方面，利用离线数据对 RL 进行预训练也能降低安全风险。不过，这需要代表性良好的历史数据集。这一点要求很高，因为工业流程大多是在使用基本控制器的调节模式下运行的。为了解决这些问题，本文提出了一种新颖的 sim2real 强化学习算法。首先，提出了一种状态适配器（SA），使模拟状态与真实状态保持一致，以减少模型与工厂之间的不匹配。然后，设计了一个固定视距回归器来取代传统的无限步回归器，为批判网络提供真正的标签，从而提高学习效率和稳定性。最后，应用近端策略优化（PPO），引入 SA-PPO 方法来实现所提出的 sim2real 算法。实验结果表明，在焙烧过程仿真中，SA-PPO 的 MSE 平均提高了 1.96%，R 平均提高了 21.64%。这验证了所提方法的有效性。

{"title":"A novel sim2real reinforcement learning algorithm for process control","authors":"Huiping Liang , Junyao Xie , Biao Huang , Yonggang Li , Bei Sun , Chunhua Yang","doi":"10.1016/j.ress.2024.110639","DOIUrl":"10.1016/j.ress.2024.110639","url":null,"abstract":"<div><div>While reinforcement learning (RL) has potential in advanced process control and optimization, its direct interaction with real industrial processes can pose safety concerns. Model-based pre-training of RL may alleviate such risks. However, the intricate nature of industrial processes complicates the establishment of entirely accurate simulation models. Consequently, RL-based controllers relying on simulation models can easily suffer from model-plant mismatch. On the one hand, utilizing offline data for pre-training of RL can also mitigate safety risks. However, it requires well-represented historical datasets. This is demanding because industrial processes mostly run under a regulatory mode with basic controllers. To handle these issues, this paper proposes a novel sim2real reinforcement learning algorithm. First, a state adaptor (SA) is proposed to align simulated states with real states to mitigate the model-plant mismatch. Then, a fix-horizon return is designed to replace traditional infinite-step return to provide genuine labels for the critic network, enhancing learning efficiency and stability. Finally, applying proximal policy optimization (PPO), the SA-PPO method is introduced to implement the proposed sim2real algorithm. Experimental results show that SA-PPO improves performance in MSE by 1.96% and in R by 21.64% on average for roasting process simulation. This verifies the effectiveness of the proposed method.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110639"},"PeriodicalIF":9.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697245","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