Advances and novel applications in systems reliability and safety engineering (selected papers of the International Conference of SRSE 2022)

Due to the growing complexity of modern engineering systems and people's increasing reliance on these systems, the reliability and safety of modern engineering systems has become a great concern. The International Conference on System Reliability and Safety Engineering (SRSE) has been launched and is held annually for researchers and practitioners in SRSE. The fourth SRSE (SRSE 2022) was held in Guangzhou, Guangdong, China on December 15−18, 2022. The conference was co-sponsored by Sun Yat-Sen University, IEEE, IEEE Reliability Society, supported by National University of Singapore, patrons with Beijing Institute of Technology, Guangdong University of Technology, Harbin Institute of Technology, Nanjing University of Science and Technology, Qingdao University, Shanghai University, Shanghai Jiao Tong University, Northwestern Polytechnical University, Zhejiang Gongshang University, City University of Hong Kong, University of Alberta, The Fifth Electronics Research Institute of Ministry of Industry and Information Technology, organized by School of Intelligent Systems Engineering, Sun Yat-sen University, China. This Special Issue on “Advances and Novel Applications in Systems Reliability and Safety Engineering” of the Quality Reliability Engineering International is motivated to attract the latest ideas and emerging research related to systems reliability and safety engineering. It aims not only to advance the state-of-the-art methodologies but also their novel applications in solving real world problems. In particular, the special issue is open to all papers that have been accepted and presented in SRSE 2022. Since its proposal, this Special Issue has drawn great attention and many papers have been submitted for potential publication. After a rigorous review process, 11 papers have been accepted for publication. To give an explicit reference on these 11 papers, the main contributions of these papers are summarized below. In Ref. 1, the authors propose a novel Goal-Oriented (GO) methodology for multi-state software-intensive systems characterized by degradation accumulation and information transmission attenuation effects. They delineate the impact of degradation accumulation on system availability by establishing non-negative constant and random variable critical thresholds for the degradation state. Subsequently, they devise an effective optical power loss model to assess the reliability of information transmission amidst attenuation phenomena. Building upon these foundations, a robust reliability analysis framework for multi-state software-intensive systems is constructed grounded in the GO methodology. They then exemplify the utility of their proposed approach by applying it to a case study involving a meter and decameter wavelength radio heliograph. Through this application, the authors illustrate the practical applicability and feasibility of their methodology. Importantly, the proposed model furnishes a more precise depiction of system reliability within authentic operational contexts. In Ref. 2, the authors present a joint optimization model of bi-level imperfect preventive maintenance (PM) and spare parts inventory. Upon each system shutdown, the PM level is determined based on the cost-effectiveness ratio. Meanwhile, an improved (s, S) inventory policy is proposed utilizing the information of the remaining number of PM and accumulated consumption of spare parts. Numerical studies show that compared with the traditional (s, S) inventory policy, the proposed joint optimization model can coordinate the ordering quantity of spare parts with the PM policy and reduce the expected maintenance cost. In Ref. 3, the authors present a joint optimization model of condition-based maintenance (CBM) and condition-based reallocation for a non-repairable system with multiple identical and functionally exchangeable components. The problem is formulated as a Markov decision process. The optimal maintenance and reallocation actions for each system state is optimized to minimize the discounted long-run total cost. The authors present a value iteration algorithm to solve the problem. The numerical analysis based on a two-component series system shows that the reallocation can effectively mitigate the risk of system failure through balancing the degradation level of components. Furthermore, compared with the traditional CBM without considering component reallocation, the proposed model can effectively reduce the risk of system failure at a lower cost. In Ref. 4, the authors investigate a preventive replacement policy of a two-component series system considering masked causes of failure. When an unexpected failure occurs, the failed component can be revealed with a diagnosis action and be replaced subsequently, or the whole system can be replaced directly without a diagnosis action. Meanwhile, when a preventive replacement is carried out on a component, the other component can be replaced opportunistically. The authors formulate the problem as a semi-Markov decision process and prove the existence of the stationary optimal policy. The optimal preventive replacement age thresholds for each component and the corresponding optimal maintenance actions upon each failure are jointly obtained to minimize the long-term average maintenance cost per time unit. A comprehensive numerical study is provided to illustrate the effectiveness of our proposed model. In Ref. 5, the author proposes a Bayesian model calibration method for improving the prediction accuracy of surrogate models used in numerical simulation for reliability assessment of engineered systems. In the proposed model calibration framework, a Gaussian process model is used as the surrogate model and a Bayesian method is adopted for model updating and bias correction by considering various uncertainties in the numerical simulation and experimental processes. Various kinds of uncertainty, such as model uncertainty, parameter uncertainty, and measurement uncertainty, can be quantified through the proposed method. High-fidelity experiment data can be leveraged to upgrade the surrogate model with significant prediction accuracy improvement. By utilizing the proposed Bayesian model calibration method, potential discrepancies between the prediction of surrogate model and the actual performance of the system under study can be minimized. Two illustrative cases, that is, a numerical example and an engineering case of NACA0012 airfoil flow simulation, are used to demonstrate the proposed method. In Ref. 6, the authors propose a novel reliability assessment methodology designed for two-level balanced systems with common bus performance sharing (TLBS-CBPS), which considers epistemic uncertainty and transmission loss considerations. Initially, a reliability model is devised based on the operational mode of TLBS-CBPS. Throughout the two-level rebalancing process, the performance transmission dynamics among modules, between modules and components, and among components are delineated as three nonlinear programming challenges. Subsequently, the authors introduce the belief universal generating function (BUGF) method to evaluate the epistemic uncertainty associated with system reliability, leveraging the Dempster-Shafer evidence theory. To address the issue of combinatorial explosion, a streamlined BUGF algorithm is proposed, integrating merging, judgment, labeling, and realization expansion techniques. The efficacy of the proposed methodology is validated through two case studies involving lithium-ion battery packs, with the accuracy of results confirmed via Monte Carlo simulation. Sensitivity analysis reveals that the transmission capacity between modules exerts the most substantial influence on system reliability, whereas the transmission loss rate between modules has a comparatively minor impact. These findings offer valuable insights for enhancing the design and configuration of such systems in practical applications. In Ref. 7, the authors propose to utilize the plant-level real-time monitoring information of nuclear power plants and to use gate-based recurrent neural network (GRU) and its variants for predicting the variations of operating condition parameters in nuclear power plant loss of coolant accidents. Two prediction methods for the trend of accident development base on the GRU networks with fixed rolling step feedback loop and fixed time step sliding delay are introduced, which are proposed to deal with insufficient data for some loss of coolant accidents. Furthermore, a GRU-D model based on the decay mechanism is proposed for operating parameter prediction, which is particular for handling with the limitation of available data, the probability of data missing as well as the contamination of noises in some loss of coolant accidents. Simulation data of transient behavior of the coolant system in a light water reactor during accidents are used to demonstrate the proposed methods. In Ref. 8, the authors propose a computationally efficient approach for resilience assessment of rail transit system under disruptions. To address the computational challenge caused by the complexity of rail transit system, an improved linear programming model is presented to characterize commuter flows and a four-step approach is proposed to estimate system statuses. A novel framework for quantifying and assessing the resilience of large-scale rail transit network is proposed by modeling commuter flows under normal and disrupted conditions, respectively. The proposed approach is demonstrated on the core part of Hangzhou rail transit network, considering more than 20,000 OD pairs involving over 80,000 commuters. Through this case study, the results indicate that all the resulting large models for resilience assessment of the system can be efficiently solved in hours and the proposed commuter flow models can provide critical information about disruption events and affected commuters. In Ref. 9, the authors propose a fault location method for drive circuits in nuclear reactors by combining the empirical modal decomposition (EMD) algorithm and sparse convolutional autoencoder (SCAE). As a critical part of instrumentation and control systems in nuclear reactors, the drive circuit directly influences the operation of nuclear reactors. The soft faults caused by the degradation of electronic components in the drive circuit are very hard to detect, which are often misclassified as normal conditions by traditional fault diagnosis methods. To improve the fault location precision of this kind of soft faults, the proposed method leverages SCAE to extract features from the condition monitoring data and further to adopt the EMD algorithm to improve the distinguishability of fault features associated with these soft faults. A deep learning-based classifier is then designed to locate soft faults in the driver circuit. The proposed method is demonstrated via a simulation study of drive circuit fault location in nuclear reactors. In Ref. 10, the authors propose a novel reliability analysis and evaluation model tailored for the intricacies of complex space phased mission systems (SPMSs), leveraging event chain, and Bayesian information fusion methodologies. The authors meticulously outline the implementation process of this integrated reliability modeling approach, providing a comprehensive framework for identifying weak links and conducting reliability assessments. Illustrating the practical utility of their model, Liu et al. conduct a detailed reliability evaluation within the context of the entry, descent, and landing process of the Tianwen-1 Mars Probe. The results of this application demonstrate the efficacy and engineering applicability of the proposed reliability analysis and evaluation model, showcasing its potential for extending to and benefiting the reliability modeling, analysis, and evaluation of diverse complex SPMSs. In Ref. 11, the authors propose both an explicit and an implicit method based on cognitive reliability and error analysis method (CREAM) to enhance the evaluation of reliability in man-machine systems (MMSs) vulnerable to probabilistic common cause errors (PCCE). The explicit method models common causes as basic events within a fault tree framework, offering a straightforward approach. However, it may face computational inefficiencies when applied to large-scale systems and is constrained to systems subject to independent common causes. On the other hand, the implicit method accommodates diverse statistical relationships among common causes, whether they are independent, mutually exclusive, or dependent. Both methodologies have demonstrated feasibility and effectiveness in the presented example. Future research will extend the investigation to PCCEs in phase-mission processes, encompassing multiple consecutive phases of operation. This expanded inquiry aims to elucidate the impact of common cause errors on various systems within complex operational contexts. The Guest Editors would like to thank all the authors for submitting their valuable papers to this special issue and would like to extend our deepest appreciation to all the anonymous reviewers for spending their great time and efforts on the thorough review of all the submitted papers. In addition, the Guest Editors would like to express our sincere gratitude to the Co-Editor-in-Chief of the Quality Reliability Engineering International, Prof. Loon Ching Tang for providing this valuable opportunity to organize the Special Issue as well as his continuous support, and to the Journal Administrator, Dr. Jeanette Loos, for her professional assistance throughout the entire preparation of the Special Issue.

Récupéré en direct depuis OpenAlex et désinversé. Les résumés ne sont pas conservés dans cette base de données : les index inversés représentent 8,6 Go des 9,3 Go de texte de la base, et le serveur dispose de 13 Go libres.