{"title":"Deep learning-stochastic ensemble for RUL prediction and predictive maintenance with dynamic mission abort policies","authors":"Faizanbasha A. , U. Rizwan","doi":"10.1016/j.ress.2025.110919","DOIUrl":"10.1016/j.ress.2025.110919","url":null,"abstract":"<div><div>Accurate prediction of Remaining Useful Life (RUL) is crucial for optimizing maintenance strategies in industrial systems. However, existing models often falter under nonlinear and nonstationary degradation conditions with stochastic and abrupt failures, limiting their real-world effectiveness. To address this, we introduce a novel approach that combines advanced deep learning architectures with stochastic modeling and dynamic optimization techniques for more precise RUL prediction. This study has three overarching aims: First, to propose a hybrid ensemble model integrating convolutional neural networks, transformers, long short-term memory networks, and a smooth semi-martingale stochastic layer, a combination not previously explored, to effectively model both deterministic and stochastic degradation processes, thereby enhancing RUL prediction accuracy. Second, to introduce a reinforcement learning-based hyperparameter tuning method that dynamically adjusts model parameters, improving performance and reducing training time, which in turn optimizes the ensemble model’s predictive capabilities. Third, to integrate the refined RUL predictions and time-varying thresholds into a multi-stage optimization framework for mission cycle assignment and resource management. This facilitates real-time decision-making and the development of a dynamic mission abort policy, including mission shifting, re-engagement, post-abortion analysis, mission plan adjustments, and maintenance scheduling. Together, these innovations enhance RUL prediction accuracy, model adaptability, and operational efficiency, ensuring reliable and cost-effective maintenance strategies in mission-critical systems. The proposed model, validated using NASA’s C-MAPSS dataset, demonstrated superior RUL prediction accuracy over state-of-the-art methods, with sensitivity analyses and ablation studies confirming its stability and effectiveness.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"259 ","pages":"Article 110919"},"PeriodicalIF":9.4,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143510492","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Risk propagation mechanisms in railway systems under extreme weather: A knowledge graph-based unsupervised causation chain approach","authors":"Yujie Huang, Zhipeng Zhang, Hao Hu","doi":"10.1016/j.ress.2025.110976","DOIUrl":"10.1016/j.ress.2025.110976","url":null,"abstract":"<div><div>Frequent and intensive adverse weathers can cause severe rail accidents through domino effect, posing significant challenges to railway safety and operational reliability. A detailed elucidation of the risk propagation mechanism across hazardous events is critical for effective risk management in rail transportation. Risk pathways involve various meteorological factors, infrastructure vulnerabilities, and consequences, in which each exhibits distinct causation strengths, trigger probabilities, severity levels, and high-impact points. To disclose the characteristics of weather-related railway domino effect accidents, this paper develops a novel railway causation analysis methodology based on an event logic graph. This framework enhances existing knowledge graph-based methodologies by emphasizing the evolution and logical progression of sequential hazardous events. Besides, an unsupervised accident causation chain linking technique is proposed, which integrates historical accident data into the knowledge graph to build a comprehensive graph database. It facilitates data-driven analysis of both structured and unstructured accident records without requiring laborious annotations. By thoroughly evaluating topological features and statistical indicators via a real-world dataset of weather-related railway accidents, key risk propagation patterns such as risk path dependence, path convergence, and risk escalation curves are recognized. Critical nodes including risk amplifiers, critical junctures, and marginal risk contributors within six critical domino chains are identified. These findings inform targeted risk mitigation strategies to prevent risk propagation and escalation. The proposed methodology and results offer theoretical support and actionable insights for enhancing safety and reliability management of railway systems under extreme weather conditions.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"260 ","pages":"Article 110976"},"PeriodicalIF":9.4,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143547940","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Intelligent decision on shield construction parameters based on safety evaluation model and sparrow search algorithm","authors":"Wei Gao, Shuangshuang Ge, Shuang Cui, Xin Chen","doi":"10.1016/j.ress.2025.110973","DOIUrl":"10.1016/j.ress.2025.110973","url":null,"abstract":"<div><div>Determination the suitable shield construction parameters is an important work to ensure the engineering construction safety. Traditional method for determination construction parameters relies on people's experience and lacks precision and adaptability. To overcome this limitation, based on the safety evaluation model using deep learning method of whale optimizing deep belief network (WO-DBN) and optimization method of sparrow search algorithm (SSA), one new method has been proposed. In this method, the WO-DBN, in which the whale optimization algorithm (WOA) has been used to select the suitable parameters of deep belief network (DBN), has been applied to safety evaluation of shield tunnel construction, and SSA has been used to determine the suitable construction parameters. By this method, the suitable construction parameters can be automatically determine by their adjustment sequence according to the sensitivity for safety, and the construction control standard is satisfied simultaneously. Then, the new method has been applied to the shield tunneling of Guangzhou subway line 18 in China to determine the suitable construction parameters for the safety evaluation objectives (ground settlement and segment floating). The results show that by the obtained construction parameters, the safety control standards for ground settlement and segment floating can be improved. This shows that the proposed method has a great contribution to improve the safety of construction by about 68 % and 89 %, respectively. In the future, the physical information and geological uncertainty can be considered in this method, which is the in-depth research to improve its adaptability.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"260 ","pages":"Article 110973"},"PeriodicalIF":9.4,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143529773","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Integrated multi-agent-based outpatient building fire response modeling for risk-driven resource use and retrofitting strategies: A case study","authors":"Aokun Yu , Haichao Bu , Tianyi Luan , Wenmei Gai","doi":"10.1016/j.ress.2025.110970","DOIUrl":"10.1016/j.ress.2025.110970","url":null,"abstract":"<div><div>To evaluate and optimize the evacuation capability of an outpatient building from a risk-driven perspective, a multi-agent-based modeling framework for emergency response that integrates the processes of fire evolution, emergency evacuation and rescue was developed. Then explored evacuation capacity optimization strategies from the perspectives of resource use and building retrofitting. To verify the effectiveness and applicability of the method, an evacuation model that simulates a fire in the outpatient building was developed. It was found that few occupants with restricted mobility took up the most evacuation time, need to improve the vertical mobility of these occupants; To prevent unplanned evacuation network interruptions due to fire, stairwell availability needs to be protected; The use of elevators for evacuation needs to be centralized to speed up the evacuation. The results of the current study will help public emergency authorities develop an effective plan for optimizing the emergency response system in outpatient buildings.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"260 ","pages":"Article 110970"},"PeriodicalIF":9.4,"publicationDate":"2025-02-25","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143529767","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"An adaptive mixture prior in Bayesian convolutional autoencoder for early detecting anomalous degradation behaviors in lithium-ion batteries","authors":"Sun Geu Chae, Suk Joo Bae","doi":"10.1016/j.ress.2025.110926","DOIUrl":"10.1016/j.ress.2025.110926","url":null,"abstract":"<div><div>Accurate and timely detection of anomalies in lithium-ion batteries is crucial for ensuring their reliability and safety. Complex degradation patterns and limited availability of labeled data pose significant challenges in identifying abnormal behaviors in battery usage. This paper proposes an unsupervised adaptive mixture distribution-based Bayesian convolutional autoencoder (AMDBCAE) method for detecting anomalous degradation behaviors in lithium-ion batteries at earlier cycles of reliability test. As the prior for the model parameters, we propose a mixture of the Laplace and Student’s <span><math><mi>t</mi></math></span> distributions by taking uncertainties in the weights of the convolutional network and their heavy-tailed characteristics into account. Using a modified form of the Bayes by backprop algorithm, the parameter of mixture proportion is adaptively updated to capture diverse and complex degradation patterns in battery degradation data more efficiently. Extracted latent features are then processed through unsupervised clustering algorithms to identify abnormal degradation behaviors of lithium-ion batteries. The analyses of two real-world lithium-ion battery datasets demonstrate the efficiency and accuracy of the proposed unsupervised framework with limited number of testing data. The proposed method addresses the limitations of manual feature extraction and the need for extensive experimental knowledge by leveraging the adaptive BCAE model to automatically extract latent features as a virtual health indicator in sparse data environments.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"259 ","pages":"Article 110926"},"PeriodicalIF":9.4,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143488960","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Knowledge embedded spatial–temporal graph convolutional networks for remaining useful life prediction","authors":"Xiao Cai , Dingcheng Zhang , Yang Yu , Min Xie","doi":"10.1016/j.ress.2025.110928","DOIUrl":"10.1016/j.ress.2025.110928","url":null,"abstract":"<div><div>Accurate prediction of remaining useful life (RUL) is crucial for prognostics and health management of equipment. Deep learning methods have gained significant attention in this field, leveraging the abundance of monitoring data captured from sensor networks. However, these methods often overlook the spatial interactions among sensor signals. Moreover, they primarily focus on pattern extraction from sensor data and neglect the utilization of available prior knowledge that could enhance prediction accuracy and stability. To address these limitations, a knowledge-embedded spatial–temporal graph convolutional networks (KEST-GCN) method is proposed. In KEST-GCN, the relationship triplets are established based on the system structure knowledge and sensor position information. Then, these triplets are transformed into low-dimensional vector embeddings using an energy-based knowledge embedded algorithm. After that, the graph dataset is generated, where the embeddings of sensors are utilized to construct the graph edges and weighted adjacency matrix. The weights are dynamically updated by an attention mechanism. Finally, a GCN layer with a multi-head attention mechanism, an LSTM layer and a fully connected layer are employed to extract spatial–temporal degradation patterns and obtain the RUL prediction results. The effectiveness and stability of our proposed method is demonstrated using an aero-engine dataset and a cutting tool dataset, respectively.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"259 ","pages":"Article 110928"},"PeriodicalIF":9.4,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143511316","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Reliability improvement of rolling stock planning with maintenance requirements for high-speed railway","authors":"Jiaxin Niu, Ke Qiao, Peng Zhao","doi":"10.1016/j.ress.2025.110972","DOIUrl":"10.1016/j.ress.2025.110972","url":null,"abstract":"<div><div>The rolling stock planning problem is a key step in the high-speed railway transit planning process. When trip delays, railway departments may need to incur considerable additional costs for plan adjustments to quickly resume operations. Therefore, we consider the impact of trip delay probabilities during rolling stock planning, aiming to minimise the total travel cost and maximise plan robustness to enhance the reliability of plan execution. A space–time–state network is established to describe the operation of rolling stocks considering accumulated mileage and running time constraints for maintenance, representing the rolling stock planning problem as a mixed-integer nonlinear programming model. Then, an alternating direction method of multipliers (ADMM)-based decomposition mechanism that decomposes the model into a set of rolling stock route selection subproblems is introduced, where each subproblem is efficiently solved by a maintenance-constrained shortest path algorithm related to reliability. A set of different scale real-life cases based on trips managed by a depot in China are used to verify the effectiveness of the proposed model and algorithm. The results show that the ADMM considerably outperforms the traditional Lagrangian relaxation (LR) method. On a set of larger-scale cases, the proposed ADMM with enhancements obtains an optimality gap of 2.42 % on average. This result is substantially better than LR, which provides optimality gaps of 32.55 % on average. Finally, the model in this paper effectively enhances the probability of successful rolling stock route execution in trip delay scenarios, resulting in a rolling stock plan with improved reliability.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"259 ","pages":"Article 110972"},"PeriodicalIF":9.4,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143510493","PeriodicalName":null,"FirstCategoryId":null,"ListUrlMain":null,"RegionNum":1,"RegionCategory":"工程技术","ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":"","EPubDate":null,"PubModel":null,"JCR":null,"JCRName":null,"Score":null,"Total":0}
{"title":"Creating an incident investigation framework for a complex socio-technical system: Application of multi-label text classification and Bayesian network structure learning","authors":"Mohammadreza Karimi Dehkordi, Fereshteh Sattari, Lianne Lefsrud","doi":"10.1016/j.ress.2025.110971","DOIUrl":"10.1016/j.ress.2025.110971","url":null,"abstract":"<div><div>The power distribution sector presents a complex socio-technical system where accidents frequently occur from various technical, human, environmental, and organizational factors, resulting in fatalities and substantial economic losses. The dynamic operational environment and complex interactions among the causal factors further complicate effective risk management and accident prevention. This research proposes a methodology to identify various risk factors and develop causal networks representing the complex relationships among these factors in power distribution incident reports. First, machine learning multi-label text classification identifies the risk factors from the incident reports. Then, the relationship among these factors is determined by integrating experts’ domain knowledge and data-driven Bayesian network structure learning approaches. Finally, the most influential causal factors and their direct/indirect effects on the incidents are identified, and proper risk control measures are recommended. The proposed methodology is applied to an incident database from a Canadian power distribution company, covering power outages, injuries, environmental issues, and near misses collected from 2013 to 2020. The results highlight that human and technical factors are the most influential and affected by organizational and environmental factors. Considering their complex interaction, implementing targeted risk management for high-risk direct/indirect causal factors could prevent further incidents and improve the companies’ overall safety.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"260 ","pages":"Article 110971"},"PeriodicalIF":9.4,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143529768","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}
Aqeel Afzal Chaudhry , Chao Zhang , Oliver G. Ernst , Thomas Nagel
{"title":"Effects of inhomogeneity and statistical and material anisotropy on THM simulations","authors":"Aqeel Afzal Chaudhry , Chao Zhang , Oliver G. Ernst , Thomas Nagel","doi":"10.1016/j.ress.2025.110921","DOIUrl":"10.1016/j.ress.2025.110921","url":null,"abstract":"<div><div>When modeling the material properties of host rocks for thermo-hydro-mechanical simulations in barrier integrity investigations for deep geological disposal of radioactive waste, numerous modeling aspects must be considered. If complete information were available, the material properties would be functions of space, with inhomogeneity and anisotropy expressed by spatially varying and tensor-valued coefficients. In practice, uncertainty is present in particular related to spatial variability of physical properties. This variability can be modeled by random fields, whose realizations are functions of space. A common choice is a Gaussian random field, determined by its mean and two-point covariance function. Anisotropy can occur both in the statistical covariance structure, resulting in different correlation lengths along principal axes, and in the physical properties themselves, leading to tensor-valued random fields. In this study, we focus on both cases, considering dominant material properties such as thermal conductivity, intrinsic permeability, and Young’s modulus, and present numerical simulations illustrating the effects of inhomogeneity, randomness, and anisotropy. Since spatial variability is a key feature in the analysis of in-situ data, this study quantifies the individual contribution of each of the listed effects in a well-controlled synthetic case and discusses them in the context of scale.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"260 ","pages":"Article 110921"},"PeriodicalIF":9.4,"publicationDate":"2025-02-24","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143529772","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}
{"title":"Optimal operation and maintenance scheduling in generalized repairable m-out-of-n standby systems with common shocks","authors":"Gregory Levitin , Liudong Xing , Yuanshun Dai","doi":"10.1016/j.ress.2025.110967","DOIUrl":"10.1016/j.ress.2025.110967","url":null,"abstract":"<div><div>This paper contributes by modeling a new class of repairable, dynamic <span><math><mi>m</mi></math></span>-out-of-<span><math><mi>n</mi></math></span> standby systems operating in random shock environments. Operating components are exposed to a common shock process and can fail due to external shocks and/or internal deterioration, causing the failure of the entire mission. Therefore, it is pivotal to implement an operation and maintenance schedule (OMS), according to which any operating component may be preventively replaced by a standby component to undergo perfect maintenance during the mission. Due to heterogeneity of system components, different OMSs incur different expected mission cost (EMC) and mission success probability (MSP). We formulate a new optimization problem to determine the optimal OMS that minimizes the EMC while satisfying a certain level of MSP. The solution methodology encompasses a new recursive procedure to evaluate MSP and the realization of genetic algorithm. A case study of a chemical reactor cooling system is conducted to showcase the proposed model and study the effects of component heterogeneity as well as several key model parameters on the system performance. The mission cost sensitivity analysis is also demonstrated, providing insights on the most cost-effective component performance or shock resistance improvement. The proposed model extends the OMS study of standby systems in literature from non-shock to shock operating environments.</div></div>","PeriodicalId":54500,"journal":{"name":"Reliability Engineering & System Safety","volume":"260 ","pages":"Article 110967"},"PeriodicalIF":9.4,"publicationDate":"2025-02-23","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":null,"resultStr":null,"platform":"Semanticscholar","paperid":"143529770","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}