Reliability Engineering & System Safety最新文献_第4页

A large-scale group SLIM considering expert credibility under social network to estimate human error probabilities in the railway driving process 考虑社会网络下专家可信度的大规模群体SLIM估计铁路驾驶过程中的人为错误概率

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.2024.110648

Jian-Lan Zhou , Ya-Lun Zhou , Ren-Bin Xiao

The Large-scale Group Success Likelihood Index Method (LG-SLIM) can eliminate bias caused by a single expert in human error assessment. The traditional LG-SLIM uses trust degrees to cluster and reach a consensus. However, the existing clustering algorithms do not consider the trust degrees between a given pair of experts to be multiple and vary according to different evaluated tasks. Besides, the existing consensus models do not consider various combinations of the evaluated tasks and trusted experts’ professions when managing trust degrees and self-confidence. Therefore, the similarity-trust-based clustering algorithm is improved using the comprehensive trust degree integrated from diverse trust degrees concerning all evaluated tasks. Moreover, expert credibility is proposed to reflect the quality of the expert's evaluation results, determined by self-confidence and trust degree simultaneously according to various combinations of the expert profession and target task. Accordingly, under the social network derived from expert credibility, the incompatible outliers change their opinions by referring to the views of those with the highest expert credibility. Finally, the sensitivity experiment and comparative analysis verify the effectiveness of the proposed model. The proposed LG-SLIM model is useful for human error assessment when critical operations need many experts to obtain reliable and accurate results.

大规模群体成功可能性指数法（large - Group Success Likelihood Index Method, LG-SLIM）可以消除单个专家在人为错误评估中造成的偏差。传统的LG-SLIM使用信任度来聚类并达成共识。然而，现有的聚类算法没有考虑到给定专家对之间的信任程度是多重的，并且会根据评估任务的不同而变化。此外，现有的共识模型在管理信任程度和自信时，没有考虑被评估任务和被信任专家职业的各种组合。因此，对基于相似度信任的聚类算法进行改进，利用对所有被评估任务的不同信任程度进行综合信任程度的整合。此外，根据专家职业和目标任务的各种组合，提出了专家可信度，以反映专家评价结果的质量，由自信程度和信任程度同时确定。因此，在专家可信度衍生的社会网络下，不相容的离群值通过参考专家可信度最高的离群值的观点来改变自己的观点。最后，通过灵敏度实验和对比分析验证了所提模型的有效性。当关键操作需要许多专家才能获得可靠和准确的结果时，所提出的LG-SLIM模型可用于人为错误评估。

{"title":"A large-scale group SLIM considering expert credibility under social network to estimate human error probabilities in the railway driving process","authors":"Jian-Lan Zhou , Ya-Lun Zhou , Ren-Bin Xiao","doi":"10.1016/j.ress.2024.110648","DOIUrl":"10.1016/j.ress.2024.110648","url":null,"abstract":"<div><div>The Large-scale Group Success Likelihood Index Method (LG-SLIM) can eliminate bias caused by a single expert in human error assessment. The traditional LG-SLIM uses trust degrees to cluster and reach a consensus. However, the existing clustering algorithms do not consider the trust degrees between a given pair of experts to be multiple and vary according to different evaluated tasks. Besides, the existing consensus models do not consider various combinations of the evaluated tasks and trusted experts’ professions when managing trust degrees and self-confidence. Therefore, the similarity-trust-based clustering algorithm is improved using the comprehensive trust degree integrated from diverse trust degrees concerning all evaluated tasks. Moreover, expert credibility is proposed to reflect the quality of the expert's evaluation results, determined by self-confidence and trust degree simultaneously according to various combinations of the expert profession and target task. Accordingly, under the social network derived from expert credibility, the incompatible outliers change their opinions by referring to the views of those with the highest expert credibility. Finally, the sensitivity experiment and comparative analysis verify the effectiveness of the proposed model. The proposed LG-SLIM model is useful for human error assessment when critical operations need many experts to obtain reliable and accurate results.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110648"},"PeriodicalIF":9.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744464","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

High-efficient non-iterative reliability-based design optimization based on the design space virtually conditionalized reliability evaluation method 基于设计空间虚拟条件化可靠性评估方法的高效非迭代可靠性设计优化

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.2024.110646

Meng-Ze Lyu , Jia-Shu Yang , Jian-Bing Chen , Jie Li

Dynamic-reliability-based design optimization (DRBDO) is a promising methodology to address the significant challenge posed by the new generation of structural design theories centered around reliability considerations. Solving DRBDO problems typically requires iterations ranging from a dozen to several hundreds, with each iteration dedicated to updating the values of design variables. Furthermore, DRBDO necessitates hundreds of or even more representative structural analyses at each iteration to compute the reliability measure, which serves as a foundation for determining the search direction in the subsequent iteration. This results in the double-loop problem confronted by DRBDO, leading to substantial computational costs for structural re-computing, particularly in the cases involving complex nonlinear stochastic dynamical systems. In the present paper, a non-iterative DRBDO paradigms is proposed by combining a novel virtually conditional reliability evaluation and the newly proposed decoupled multi-probability density evolution method (M-PDEM). By leveraging the decoupled M-PDEM, a series of one-dimensional partial differential equations (PDEs) named Li-Chen equations are solved to calculate the joint PDF of multiple responses. This enables efficient computation of the joint probability density function (PDF) of design variables and extreme response as well as the conditional PDF of the extreme response given the values of design variables based on finite representative structural analyses. Then, the reliability of different designs can be regarded as the integral of the conditional PDF, which yields the reliability feasible domain. For problems that the objective function is monotonic to each design variables, by combining with a direct search technique, this method transforms the optimization process into true iteration-free calculations, and thereby eliminates the significant computational burden associated with structural re-computing at different intermediate designs in optimization iterations. Finally, the accuracy and effectiveness of this novel method are validated through numerical examples.

基于动态可靠性的优化设计（DRBDO）是一种很有前途的方法，可以解决以可靠性为中心的新一代结构设计理论所带来的巨大挑战。解决 DRBDO 问题通常需要十几次到几百次的迭代，每次迭代都要更新设计变量的值。此外，DRBDO 需要在每次迭代中进行数百次甚至更多的代表性结构分析，以计算可靠性度量，并以此为基础确定后续迭代的搜索方向。这就造成了 DRBDO 所面临的双循环问题，导致结构重新计算所需的大量计算成本，尤其是在涉及复杂非线性随机动力系统的情况下。本文提出了一种非迭代 DRBDO 范式，它结合了新颖的虚拟条件可靠性评估和新提出的解耦多概率密度演化法（M-PDEM）。通过利用解耦 M-PDEM，一系列名为李陈方程的一维偏微分方程（PDE）被求解，以计算多个响应的联合 PDF。这样就能在有限代表性结构分析的基础上，高效计算出设计变量和极端响应的联合概率密度函数 (PDF)，以及给定设计变量值的极端响应的条件 PDF。然后，不同设计的可靠性可视为条件 PDF 的积分，从而得出可靠性可行域。对于目标函数与各设计变量单调的问题，通过与直接搜索技术相结合，该方法将优化过程转化为真正的无迭代计算，从而消除了优化迭代中不同中间设计的结构重新计算所带来的巨大计算负担。最后，通过数值实例验证了这种新方法的准确性和有效性。

{"title":"High-efficient non-iterative reliability-based design optimization based on the design space virtually conditionalized reliability evaluation method","authors":"Meng-Ze Lyu , Jia-Shu Yang , Jian-Bing Chen , Jie Li","doi":"10.1016/j.ress.2024.110646","DOIUrl":"10.1016/j.ress.2024.110646","url":null,"abstract":"<div><div>Dynamic-reliability-based design optimization (DRBDO) is a promising methodology to address the significant challenge posed by the new generation of structural design theories centered around reliability considerations. Solving DRBDO problems typically requires iterations ranging from a dozen to several hundreds, with each iteration dedicated to updating the values of design variables. Furthermore, DRBDO necessitates hundreds of or even more representative structural analyses at each iteration to compute the reliability measure, which serves as a foundation for determining the search direction in the subsequent iteration. This results in the double-loop problem confronted by DRBDO, leading to substantial computational costs for structural re-computing, particularly in the cases involving complex nonlinear stochastic dynamical systems. In the present paper, a non-iterative DRBDO paradigms is proposed by combining a novel virtually conditional reliability evaluation and the newly proposed decoupled multi-probability density evolution method (M-PDEM). By leveraging the decoupled M-PDEM, a series of one-dimensional partial differential equations (PDEs) named Li-Chen equations are solved to calculate the joint PDF of multiple responses. This enables efficient computation of the joint probability density function (PDF) of design variables and extreme response as well as the conditional PDF of the extreme response given the values of design variables based on finite representative structural analyses. Then, the reliability of different designs can be regarded as the integral of the conditional PDF, which yields the reliability feasible domain. For problems that the objective function is monotonic to each design variables, by combining with a direct search technique, this method transforms the optimization process into true iteration-free calculations, and thereby eliminates the significant computational burden associated with structural re-computing at different intermediate designs in optimization iterations. Finally, the accuracy and effectiveness of this novel method are validated through numerical examples.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110646"},"PeriodicalIF":9.4,"publicationDate":"2024-11-15","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697252","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

Modeling nuclear power plant piping reliability by coupling a human reliability analysis-based maintenance model with a physical degradation model 基于人为可靠性分析的维护模型与物理退化模型耦合的核电厂管道可靠性建模

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-14 DOI: 10.1016/j.ress.2024.110655

John Beal, Seyed Reihani, Tatsuya Sakurahara, Ernie Kee, Zahra Mohaghegh

Reliability and availability analysis for repairable components, considering the underlying physical degradation and maintenance, is crucial in support of risk assessment and management. In nuclear power plants (NPPs), reactor coolant piping is a representative example of safety-critical repairable components that are subjected to long-term physical degradation interacting with maintenance activities. The existing methods for piping reliability analysis suffer from a limitation in their capability to analyze the time-dependent physics-maintenance interactions that could occur during the component lifetime and alter the underlying maintenance processes, for instance, an enhancement of maintenance programs based on condition monitoring data or an observed defect. To address this limitation, this paper develops a new piping reliability analysis methodology that couples a physics-of-failure (PoF) model with a maintenance performance analysis model. The contributions of this paper are two-fold: (i) developing a human reliability analysis (HRA)-based maintenance performance analysis model for NPP piping that can quantify maintenance outcomes under multiple types of maintenance programs, including time-based and condition-based preventive maintenance; and (ii) developing a computational methodology to couple the HRA-based maintenance performance analysis model with PoF models. The proposed physics-maintenance coupling methodology is applied to an NPP piping case study.

考虑到潜在的物理退化和维护，可修复部件的可靠性和可用性分析对于支持风险评估和管理至关重要。在核电站（NPPs）中，反应堆冷却剂管道是安全关键的可修复部件的典型例子，这些部件受到长期物理退化的影响，与维护活动相互作用。现有的管道可靠性分析方法在分析部件寿命期间可能发生的随时间变化的物理维护相互作用和改变潜在维护过程（例如，根据状态监测数据或观察到的缺陷加强维护计划）方面存在局限性。为了解决这一限制，本文开发了一种新的管道可靠性分析方法，该方法将失效物理模型（PoF）与维护性能分析模型相结合。本文的贡献有两个方面：(i)开发了一个基于人类可靠性分析（HRA）的核电厂管道维修性能分析模型，该模型可以量化多种维修计划下的维修结果，包括基于时间和基于状态的预防性维修；（ii）开发一种计算方法，将基于hra的维护性能分析模型与PoF模型相结合。提出的物理-维护耦合方法应用于核电厂管道的案例研究。

{"title":"Modeling nuclear power plant piping reliability by coupling a human reliability analysis-based maintenance model with a physical degradation model","authors":"John Beal, Seyed Reihani, Tatsuya Sakurahara, Ernie Kee, Zahra Mohaghegh","doi":"10.1016/j.ress.2024.110655","DOIUrl":"10.1016/j.ress.2024.110655","url":null,"abstract":"<div><div>Reliability and availability analysis for repairable components, considering the underlying physical degradation and maintenance, is crucial in support of risk assessment and management. In nuclear power plants (NPPs), reactor coolant piping is a representative example of safety-critical repairable components that are subjected to long-term physical degradation interacting with maintenance activities. The existing methods for piping reliability analysis suffer from a limitation in their capability to analyze the time-dependent physics-maintenance interactions that could occur during the component lifetime and alter the underlying maintenance processes, for instance, an enhancement of maintenance programs based on condition monitoring data or an observed defect. To address this limitation, this paper develops a new piping reliability analysis methodology that couples a physics-of-failure (PoF) model with a maintenance performance analysis model. The contributions of this paper are two-fold: (i) developing a human reliability analysis (HRA)-based maintenance performance analysis model for NPP piping that can quantify maintenance outcomes under multiple types of maintenance programs, including time-based and condition-based preventive maintenance; and (ii) developing a computational methodology to couple the HRA-based maintenance performance analysis model with PoF models. The proposed physics-maintenance coupling methodology is applied to an NPP piping case study.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110655"},"PeriodicalIF":9.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744559","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

Machine remaining useful life prediction method based on global-local attention compensation network 基于全局-局部注意力补偿网络的机器剩余使用寿命预测方法

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-14 DOI: 10.1016/j.ress.2024.110652

Zhixiang Chen

Accurate remaining useful life (RUL) prediction is essential for ensuring the safe operation of machinery. The extraction of high-level features that contain both global dependencies and local refinements can effectively improve the accuracy of RUL predictions. In order to extract high-level features, this paper proposes a global-local attention compensation network (GLACN) for RUL prediction. The proposed network integrates a global interaction-feature (GIF) mechanism, a long short-term memory network (LSTM), and a local attention enhanced residual compensation (LAERC) mechanism. Initially, the GIF mechanism is used to processed selected signals from multiple sensors to facilitate global information interaction and allocate channel attention weights. Subsequently, the LSTM is employed to extract global temporal features and establish long-term dependencies among them. Finally, the global temporal features extracted by LSTM are further refined by LAERC to mine local features. To address the potential weakening of long-term dependencies during feature refinement, the global temporal features from the last hidden layer of LSTM are utilized as compensation, concatenated with refined features to generate final features. The effectiveness of the designed model for RUL prediction is tested by two benchmark datasets. The results illustrate that the prediction performance of the GLACN outperforms some of some state-of-the-art (SOTA) methods.

准确的剩余使用寿命（RUL）预测是保证机械安全运行的必要条件。对包含全局依赖和局部细化的高级特征的提取可以有效地提高RUL预测的准确性。为了提取高级特征，本文提出了一种用于RUL预测的全局-局部注意补偿网络（GLACN）。该网络集成了全局交互特征（GIF）机制、长短期记忆网络（LSTM）和局部注意增强剩余补偿（LAERC）机制。首先，利用GIF机制对来自多个传感器的选定信号进行处理，以促进全局信息交互和分配通道关注权。然后，利用LSTM提取全局时间特征并建立它们之间的长期依赖关系。最后，LSTM提取的全局时间特征通过LAERC进一步细化，挖掘局部特征。为了解决特征精化过程中可能弱化长期依赖关系的问题，利用LSTM最后一层隐藏层的全局时态特征作为补偿，与精化特征连接生成最终特征。通过两个基准数据集验证了所设计模型在RUL预测中的有效性。结果表明，GLACN的预测性能优于一些最先进的（SOTA）方法。

{"title":"Machine remaining useful life prediction method based on global-local attention compensation network","authors":"Zhixiang Chen","doi":"10.1016/j.ress.2024.110652","DOIUrl":"10.1016/j.ress.2024.110652","url":null,"abstract":"<div><div>Accurate remaining useful life (RUL) prediction is essential for ensuring the safe operation of machinery. The extraction of high-level features that contain both global dependencies and local refinements can effectively improve the accuracy of RUL predictions. In order to extract high-level features, this paper proposes a global-local attention compensation network (GLACN) for RUL prediction. The proposed network integrates a global interaction-feature (GIF) mechanism, a long short-term memory network (LSTM), and a local attention enhanced residual compensation (LAERC) mechanism. Initially, the GIF mechanism is used to processed selected signals from multiple sensors to facilitate global information interaction and allocate channel attention weights. Subsequently, the LSTM is employed to extract global temporal features and establish long-term dependencies among them. Finally, the global temporal features extracted by LSTM are further refined by LAERC to mine local features. To address the potential weakening of long-term dependencies during feature refinement, the global temporal features from the last hidden layer of LSTM are utilized as compensation, concatenated with refined features to generate final features. The effectiveness of the designed model for RUL prediction is tested by two benchmark datasets. The results illustrate that the prediction performance of the GLACN outperforms some of some state-of-the-art (SOTA) methods.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110652"},"PeriodicalIF":9.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744557","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

Pseudo-label assisted contrastive learning model for unsupervised open-set domain adaptation in fault diagnosis 用于故障诊断中无监督开放集域适应的伪标签辅助对比学习模型

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-14 DOI: 10.1016/j.ress.2024.110650

Weicheng Wang , Chao Li , Zhipeng Zhang , Jinglong Chen , Shuilong He , Yong Feng

The operation of mechanical equipment is frequently characterized by complexity and variability, leading to signal domain shifts. This phenomenon underscores the significance of cross-domain fault diagnosis for maintaining the reliability and safety of mechanical systems. Due to the absence of labeled data in many operational contexts, there's a clear need for an unsupervised domain adaptation technique that does not rely on labeled information. Moreover, traditional domain adaptation methods presuppose identical label distributions across source and target domains. Nevertheless, real-world engineering scenarios often present novel fault categories out of distribution, thereby challenging the efficacy of established domain adaption methods. To address these challenges, we proposed a pseudo-label assisted contrastive learning model (PLA-CLM) for Unsupervised Open-set Domain Adaptation. Based on contrastive learning, the proposed model effectively minimizes the discrepancy between samples of identical pseudo-label across domains, while simultaneously integrating distance, density, and entropy to isolate out-of-distribution samples. After training, the model adaptively identifies known faults and detects OOD faults using thresholds calculated based on sample distribution. Experimental results on two datasets demonstrate that our method surpasses existing approaches, ensuring enhanced reliability of mechanical systems’ operation and maintenance.

机械设备的运行往往具有复杂性和多变性的特点，从而导致信号域的偏移。这种现象凸显了跨域故障诊断对于维护机械系统的可靠性和安全性的重要意义。由于在许多运行环境中缺乏标记数据，因此显然需要一种不依赖标记信息的无监督域适应技术。此外，传统的域适应方法预先假定源域和目标域的标签分布完全相同。然而，现实世界的工程场景中经常会出现分布不均的新故障类别，从而对已有的域自适应方法的有效性提出了挑战。为了应对这些挑战，我们提出了一种用于无监督开放集域自适应的伪标签辅助对比学习模型（PLA-CLM）。基于对比学习，所提出的模型能有效地最小化不同领域中相同伪标签样本之间的差异，同时整合距离、密度和熵来隔离分布外样本。训练完成后，该模型会自适应地识别已知故障，并使用根据样本分布计算出的阈值检测 OOD 故障。在两个数据集上的实验结果表明，我们的方法超越了现有方法，确保提高机械系统运行和维护的可靠性。

{"title":"Pseudo-label assisted contrastive learning model for unsupervised open-set domain adaptation in fault diagnosis","authors":"Weicheng Wang , Chao Li , Zhipeng Zhang , Jinglong Chen , Shuilong He , Yong Feng","doi":"10.1016/j.ress.2024.110650","DOIUrl":"10.1016/j.ress.2024.110650","url":null,"abstract":"<div><div>The operation of mechanical equipment is frequently characterized by complexity and variability, leading to signal domain shifts. This phenomenon underscores the significance of cross-domain fault diagnosis for maintaining the reliability and safety of mechanical systems. Due to the absence of labeled data in many operational contexts, there's a clear need for an unsupervised domain adaptation technique that does not rely on labeled information. Moreover, traditional domain adaptation methods presuppose identical label distributions across source and target domains. Nevertheless, real-world engineering scenarios often present novel fault categories out of distribution, thereby challenging the efficacy of established domain adaption methods. To address these challenges, we proposed a pseudo-label assisted contrastive learning model (PLA-CLM) for Unsupervised Open-set Domain Adaptation. Based on contrastive learning, the proposed model effectively minimizes the discrepancy between samples of identical pseudo-label across domains, while simultaneously integrating distance, density, and entropy to isolate out-of-distribution samples. After training, the model adaptively identifies known faults and detects OOD faults using thresholds calculated based on sample distribution. Experimental results on two datasets demonstrate that our method surpasses existing approaches, ensuring enhanced reliability of mechanical systems’ operation and maintenance.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110650"},"PeriodicalIF":9.4,"publicationDate":"2024-11-14","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697246","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

Trustworthy Bayesian deep learning framework for uncertainty quantification and confidence calibration: Application in machinery fault diagnosis 基于可信赖贝叶斯深度学习框架的不确定性量化与置信度定标：在机械故障诊断中的应用

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-13 DOI: 10.1016/j.ress.2024.110657

Hao Li, Jinyang Jiao, Zongyang Liu, Jing Lin, Tian Zhang, Hanyang Liu

Reliable and accurate machinery fault diagnosis is crucial for ensuring operational safety and reducing downtime in industrial settings. Traditional intelligent diagnosis methods only focus on improving the accuracy of in-distribution samples, but neglect the trustworthiness evaluation of diagnosis results. To address these issues, this paper developed a novel trustworthy machinery fault diagnosis (TMFD) method, which integrates Bayesian deep learning techniques with model calibration strategies. Specifically, TMFD regards a Bayesian convolutional neural network framework as the backbone. Then, we introduce α-divergence to facilitate the decomposition and quantification of epistemic uncertainty and aleatoric uncertainty, ultimately achieving out-of-distribution sample detection through epistemic uncertainty. Then, the ante-calibration loss constraint and the compositional post-calibration operation are jointly applied to promote data-efficient and high expressive calibration for in-distribution sample diagnosis confidence. Finally, TMFD is validated using three experimental datasets, demonstrating its effectiveness and robustness in machinery fault diagnosis.

可靠和准确的机械故障诊断对于确保操作安全和减少工业环境中的停机时间至关重要。传统的智能诊断方法只注重提高分布样本的准确性，而忽视了诊断结果的可信度评价。为了解决这些问题，本文开发了一种新的可信机械故障诊断（TMFD）方法，该方法将贝叶斯深度学习技术与模型校准策略相结合。具体而言，TMFD以贝叶斯卷积神经网络框架为骨干。然后，我们引入α-散度，便于对认知不确定性和任意不确定性进行分解和量化，最终通过认知不确定性实现分布外样本检测。然后，结合校正前损失约束和组合校正后操作，实现分布样本诊断置信度的数据高效、高表达校正。最后，利用三个实验数据集对TMFD进行了验证，验证了其在机械故障诊断中的有效性和鲁棒性。

{"title":"Trustworthy Bayesian deep learning framework for uncertainty quantification and confidence calibration: Application in machinery fault diagnosis","authors":"Hao Li, Jinyang Jiao, Zongyang Liu, Jing Lin, Tian Zhang, Hanyang Liu","doi":"10.1016/j.ress.2024.110657","DOIUrl":"10.1016/j.ress.2024.110657","url":null,"abstract":"<div><div>Reliable and accurate machinery fault diagnosis is crucial for ensuring operational safety and reducing downtime in industrial settings. Traditional intelligent diagnosis methods only focus on improving the accuracy of in-distribution samples, but neglect the trustworthiness evaluation of diagnosis results. To address these issues, this paper developed a novel trustworthy machinery fault diagnosis (TMFD) method, which integrates Bayesian deep learning techniques with model calibration strategies. Specifically, TMFD regards a Bayesian convolutional neural network framework as the backbone. Then, we introduce α-divergence to facilitate the decomposition and quantification of epistemic uncertainty and aleatoric uncertainty, ultimately achieving out-of-distribution sample detection through epistemic uncertainty. Then, the ante-calibration loss constraint and the compositional post-calibration operation are jointly applied to promote data-efficient and high expressive calibration for in-distribution sample diagnosis confidence. Finally, TMFD is validated using three experimental datasets, demonstrating its effectiveness and robustness in machinery fault diagnosis.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110657"},"PeriodicalIF":9.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744561","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

Aerodynamic robustness optimization of aeroengine fan performance based on an interpretable dynamic machine learning method 基于可解释动态机器学习方法的航空发动机风扇性能气动鲁棒性优化

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-13 DOI: 10.1016/j.ress.2024.110654

Hongzhi CHENG , Ziqing ZHANG , Xingen LU , Penghao DUAN , Junqiang ZHU

Aeroengines and gas turbines are susceptible to uncertainties during manufacturing and operation, leading to reduced efficiency and dispersed performance. Current engine design system often produces deterministic performance databases that cannot be effectively used to guide the uncertainty analysis and robust design process of turbomachinery. This paper proposes an interpretable dynamic machine learning method for sensitivity analysis and robust optimization of turbomachinery blades. A dynamic extreme gradient boosting (XGBoost) is trained to predict fan aerodynamic performance, and the SHapley additional explanation (SHAP) method is introduced to explain regression model behavior and identify the impact of uncertain variables. On this basis, the Lipschitz-based trust region (MAXLIPO-TR) optimization algorithm is used to obtain the optimal configuration with the best robustness performance. Finally, the method is applied to data mining for design guidelines of robustness performance enhancement of an aeroengine fan. The results show that maximum camber and tangential stacking have major effects on fan performance dispersion. The standard deviation of the isentropic efficiency, pressure ratio and mass flow rate of the optimized configuration are reduced by 42.4%, 35.6% and 22.7% respectively at design conditions. The proposed data mining method has scientific significance and industrial application value in the robust design of advanced turbomachinery.

航空发动机和燃气轮机在制造和运行过程中容易受到不确定因素的影响，导致效率降低和性能分散。目前的发动机设计系统通常生成确定性的性能数据库，无法有效用于指导涡轮机械的不确定性分析和鲁棒性设计过程。本文提出了一种可解释的动态机器学习方法，用于透平机械叶片的敏感性分析和鲁棒性优化。通过训练动态极梯度提升（XGBoost）来预测风机气动性能，并引入 SHapley 附加解释（SHAP）方法来解释回归模型行为并识别不确定变量的影响。在此基础上，使用基于 Lipschitz 的信任区域（MAXLIPO-TR）优化算法来获得鲁棒性能最佳的最优配置。最后，将该方法应用于数据挖掘，为增强航空发动机风扇的鲁棒性能提供设计指导。结果表明，最大外倾和切向堆叠对风扇性能分散有很大影响。在设计条件下，优化配置的等熵效率、压力比和质量流量的标准偏差分别降低了 42.4%、35.6% 和 22.7%。所提出的数据挖掘方法对先进透平机械的稳健设计具有重要的科学意义和工业应用价值。

{"title":"Aerodynamic robustness optimization of aeroengine fan performance based on an interpretable dynamic machine learning method","authors":"Hongzhi CHENG , Ziqing ZHANG , Xingen LU , Penghao DUAN , Junqiang ZHU","doi":"10.1016/j.ress.2024.110654","DOIUrl":"10.1016/j.ress.2024.110654","url":null,"abstract":"<div><div>Aeroengines and gas turbines are susceptible to uncertainties during manufacturing and operation, leading to reduced efficiency and dispersed performance. Current engine design system often produces deterministic performance databases that cannot be effectively used to guide the uncertainty analysis and robust design process of turbomachinery. This paper proposes an interpretable dynamic machine learning method for sensitivity analysis and robust optimization of turbomachinery blades. A dynamic extreme gradient boosting (XGBoost) is trained to predict fan aerodynamic performance, and the SHapley additional explanation (SHAP) method is introduced to explain regression model behavior and identify the impact of uncertain variables. On this basis, the Lipschitz-based trust region (MAXLIPO-TR) optimization algorithm is used to obtain the optimal configuration with the best robustness performance. Finally, the method is applied to data mining for design guidelines of robustness performance enhancement of an aeroengine fan. The results show that maximum camber and tangential stacking have major effects on fan performance dispersion. The standard deviation of the isentropic efficiency, pressure ratio and mass flow rate of the optimized configuration are reduced by 42.4%, 35.6% and 22.7% respectively at design conditions. The proposed data mining method has scientific significance and industrial application value in the robust design of advanced turbomachinery.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110654"},"PeriodicalIF":9.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697249","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

An uncertainty-incorporated active data diffusion learning framework for few-shot equipment RUL prediction 不确定性融入式主动数据扩散学习框架，用于少发设备 RUL 预测

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-13 DOI: 10.1016/j.ress.2024.110632

Chao Zhang , Daqing Gong , Gang Xue

In predicting the remaining useful life (RUL) of critical equipment, the challenge of obtaining degradation data and the limitation of data volume lead to few-shot problems that significantly impact prediction accuracy. To address this issue, this paper introduces a reinforcement learning feedback loop mechanism for predicting the RUL of critical equipment. Initially, the framework uses a data diffusion model to generate a dataset that closely approximates the distribution of the labeled samples for data augmentation. Subsequently, Bayesian deep learning and Monte Carlo (MC) dropout inference provide uncertainty quantifications for RUL interval predictions. An active learning strategy, which is based on uncertainty and diversity, converts unlabeled samples into labeled samples, thereby selecting the most effective training dataset. In each iteration, the model adjusts its strategy for selecting and generating data based on the current state of learning, dynamically adapting to the needs of the learning process via Bayesian methods. The proposed prediction framework was validated through experiments using the C-MAPSS and NASA battery datasets. The results indicate that the application of data diffusion and active learning strategies significantly enhances prediction performance, increasing confidence by 42 %. Comparative experiments with other benchmark methods demonstrate that the proposed method reduces prediction uncertainty by at least 15 %.

在预测关键设备的剩余使用寿命（RUL）时，获取降解数据的挑战和数据量的限制导致了少数问题，严重影响了预测精度。为解决这一问题，本文引入了一种强化学习反馈循环机制，用于预测关键设备的剩余使用寿命。首先，该框架使用数据扩散模型生成一个数据集，该数据集近似于用于数据增强的标记样本分布。随后，贝叶斯深度学习和蒙特卡洛（MC）遗漏推理为 RUL 间隔预测提供了不确定性量化。基于不确定性和多样性的主动学习策略将未标记样本转换为标记样本，从而选择最有效的训练数据集。在每次迭代中，模型都会根据当前的学习状态调整其选择和生成数据的策略，通过贝叶斯方法动态适应学习过程的需要。通过使用 C-MAPSS 和 NASA 电池数据集进行实验，验证了所提出的预测框架。结果表明，数据扩散和主动学习策略的应用大大提高了预测性能，置信度提高了 42%。与其他基准方法的对比实验表明，所提出的方法至少降低了 15% 的预测不确定性。

{"title":"An uncertainty-incorporated active data diffusion learning framework for few-shot equipment RUL prediction","authors":"Chao Zhang , Daqing Gong , Gang Xue","doi":"10.1016/j.ress.2024.110632","DOIUrl":"10.1016/j.ress.2024.110632","url":null,"abstract":"<div><div>In predicting the remaining useful life (RUL) of critical equipment, the challenge of obtaining degradation data and the limitation of data volume lead to few-shot problems that significantly impact prediction accuracy. To address this issue, this paper introduces a reinforcement learning feedback loop mechanism for predicting the RUL of critical equipment. Initially, the framework uses a data diffusion model to generate a dataset that closely approximates the distribution of the labeled samples for data augmentation. Subsequently, Bayesian deep learning and Monte Carlo (MC) dropout inference provide uncertainty quantifications for RUL interval predictions. An active learning strategy, which is based on uncertainty and diversity, converts unlabeled samples into labeled samples, thereby selecting the most effective training dataset. In each iteration, the model adjusts its strategy for selecting and generating data based on the current state of learning, dynamically adapting to the needs of the learning process via Bayesian methods. The proposed prediction framework was validated through experiments using the C-MAPSS and NASA battery datasets. The results indicate that the application of data diffusion and active learning strategies significantly enhances prediction performance, increasing confidence by 42 %. Comparative experiments with other benchmark methods demonstrate that the proposed method reduces prediction uncertainty by at least 15 %.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110632"},"PeriodicalIF":9.4,"publicationDate":"2024-11-13","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142697247","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 three-stage bearing transfer fault diagnosis method for large domain shift scenarios 针对大域移动场景的三阶段轴承传递故障诊断方法

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-12 DOI: 10.1016/j.ress.2024.110641

Kai Huang , Zhijun Ren , Linbo Zhu , Tantao Lin , Yongsheng Zhu , Li Zeng , Jin Wan

In recent years, significant progress has been achieved in the intelligent fault diagnosis of bearings based on transfer learning. However, existing methods overlook the presence of domain-specific features that are non-transferable when aligning domain distributions. Additionally, the reliability of subdomain alignment has not been adequately evaluated. This severely restricts the diagnostic performance of transfer learning, especially in scenarios of large domain shifts. To address these issues, this paper proposes a novel approach based on three-stage transfer alignment. In the first stage, two private encoders, and a shared encoder are designed to eliminate domain-specific features, thus maximizing the effectiveness and transferability of shared encoded features. Subsequently, in the second stage, a deep adversarial domain adaptation method is introduced to adapt the global distributions between the two domains. Lastly, the third stage presents a novel soft pseudo-label distillation method, based on adaptive entropy weighting. This enhances alignment between subdomains, further bridging the distribution gap between the two domains. A series of comprehensive experiments under two types of large domain shift scenarios validate that the proposed method has a superior performance and could boost 6.93 % and 6.14 % accuracy than the state-of-the-art methods, respectively.

近年来，基于迁移学习的轴承智能故障诊断取得了重大进展。然而，现有方法在对齐域分布时忽略了不可迁移的特定域特征的存在。此外，子域对齐的可靠性也未得到充分评估。这严重限制了迁移学习的诊断性能，尤其是在领域发生巨大变化的情况下。为了解决这些问题，本文提出了一种基于三阶段转移对齐的新方法。在第一阶段，设计了两个私有编码器和一个共享编码器，以消除特定领域的特征，从而最大限度地提高共享编码特征的有效性和可转移性。随后，在第二阶段，引入了一种深度对抗域适应方法，以适应两个域之间的全局分布。最后，第三阶段提出了一种基于自适应熵加权的新型软伪标签蒸馏方法。这增强了子域之间的一致性，进一步缩小了两个域之间的分布差距。在两类大型域转移场景下进行的一系列综合实验验证了所提出的方法性能优越，比最先进的方法分别提高了 6.93% 和 6.14% 的准确率。

{"title":"A three-stage bearing transfer fault diagnosis method for large domain shift scenarios","authors":"Kai Huang , Zhijun Ren , Linbo Zhu , Tantao Lin , Yongsheng Zhu , Li Zeng , Jin Wan","doi":"10.1016/j.ress.2024.110641","DOIUrl":"10.1016/j.ress.2024.110641","url":null,"abstract":"<div><div>In recent years, significant progress has been achieved in the intelligent fault diagnosis of bearings based on transfer learning. However, existing methods overlook the presence of domain-specific features that are non-transferable when aligning domain distributions. Additionally, the reliability of subdomain alignment has not been adequately evaluated. This severely restricts the diagnostic performance of transfer learning, especially in scenarios of large domain shifts. To address these issues, this paper proposes a novel approach based on three-stage transfer alignment. In the first stage, two private encoders, and a shared encoder are designed to eliminate domain-specific features, thus maximizing the effectiveness and transferability of shared encoded features. Subsequently, in the second stage, a deep adversarial domain adaptation method is introduced to adapt the global distributions between the two domains. Lastly, the third stage presents a novel soft pseudo-label distillation method, based on adaptive entropy weighting. This enhances alignment between subdomains, further bridging the distribution gap between the two domains. A series of comprehensive experiments under two types of large domain shift scenarios validate that the proposed method has a superior performance and could boost 6.93 % and 6.14 % accuracy than the state-of-the-art methods, respectively.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"254 ","pages":"Article 110641"},"PeriodicalIF":9.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142657364","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

IF 9.4 1区工程技术 Q1 ENGINEERING, INDUSTRIAL

Reliability Engineering & System Safety

Pub Date : 2024-11-12 DOI: 10.1016/j.ress.2024.110643

Tao Wang , Shin Yee Khoo , Zhi Chao Ong , Pei Yi Siow , Teng Wang

Entropy-based methods are widely used in machinery fault diagnosis for characterizing system disorder and complexity. However, conventional entropy techniques often fail to capture local signal variations when analyzing relationships between vectors, especially in complex settings. This leads to incomplete representations of subtle features and dynamic behaviors, resulting in inaccurate estimations of system complexity and affecting diagnostic accuracy and reliability. To address these limitations, a novel distance similarity entropy (DSEn) is proposed in this paper: (1) It leverages element-wise distance to precisely capture local shifts and subtle distortions between subsequences. (2) It employs a Gaussian kernel function for vector similarity, enhancing signal pattern analysis by preserving subtle differences and mitigating the impact of outliers. (3) It uses probability density estimation of distance similarities between adjacent vectors to track changes in internal signal patterns, enabling more accurate and sensitive estimations of signal complexity. Synthetic signal experiments demonstrate that DSEn excels in detecting dynamic time series changes and characterizing signal complexity. Tests on two bearing datasets reveal that DSEn's extracted features show significant differences, highlighted by Hedges’ g effect size. Compared to other commonly used entropies (SampEn, PermEn, FuzzEn, DistEn, etc.), DSEn shows superior fault identification accuracy, computational efficiency, and noise resistance.

基于熵的故障诊断方法被广泛应用于机械故障诊断中，以表征系统的无序性和复杂性。然而，在分析向量之间的关系时，传统的熵技术往往无法捕获局部信号变化，特别是在复杂的环境中。这将导致细微特征和动态行为的不完整表示，从而导致对系统复杂性的不准确估计，并影响诊断的准确性和可靠性。为了解决这些限制，本文提出了一种新的距离相似熵（DSEn）：(1)它利用元素之间的距离来精确捕获子序列之间的局部移位和微妙扭曲。(2)采用高斯核函数进行向量相似性，通过保留细微差异和减轻异常值的影响来增强信号模式分析。(3)利用相邻向量之间距离相似度的概率密度估计来跟踪内部信号模式的变化，使信号复杂度的估计更加准确和灵敏。合成信号实验表明，DSEn在检测动态时间序列变化和表征信号复杂度方面表现优异。对两个轴承数据集的测试表明，DSEn提取的特征显示出显着差异，突出显示了Hedges的g效应大小。与其他常用熵（SampEn、PermEn、FuzzEn、DistEn等）相比，DSEn具有更好的故障识别精度、计算效率和抗噪声能力。

{"title":"Distance similarity entropy: A sensitive nonlinear feature extraction method for rolling bearing fault diagnosis","authors":"Tao Wang , Shin Yee Khoo , Zhi Chao Ong , Pei Yi Siow , Teng Wang","doi":"10.1016/j.ress.2024.110643","DOIUrl":"10.1016/j.ress.2024.110643","url":null,"abstract":"<div><div>Entropy-based methods are widely used in machinery fault diagnosis for characterizing system disorder and complexity. However, conventional entropy techniques often fail to capture local signal variations when analyzing relationships between vectors, especially in complex settings. This leads to incomplete representations of subtle features and dynamic behaviors, resulting in inaccurate estimations of system complexity and affecting diagnostic accuracy and reliability. To address these limitations, a novel distance similarity entropy (DSEn) is proposed in this paper: (1) It leverages element-wise distance to precisely capture local shifts and subtle distortions between subsequences. (2) It employs a Gaussian kernel function for vector similarity, enhancing signal pattern analysis by preserving subtle differences and mitigating the impact of outliers. (3) It uses probability density estimation of distance similarities between adjacent vectors to track changes in internal signal patterns, enabling more accurate and sensitive estimations of signal complexity. Synthetic signal experiments demonstrate that DSEn excels in detecting dynamic time series changes and characterizing signal complexity. Tests on two bearing datasets reveal that DSEn's extracted features show significant differences, highlighted by Hedges’ <em>g</em> effect size. Compared to other commonly used entropies (SampEn, PermEn, FuzzEn, DistEn, etc.), DSEn shows superior fault identification accuracy, computational efficiency, and noise resistance.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"255 ","pages":"Article 110643"},"PeriodicalIF":9.4,"publicationDate":"2024-11-12","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"142744556","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