Ensuring Slurry Concentration (SC) within a target range is critical for stable and safe dredging operations of dredgers, yet direct SC measurements are often unreliable due to sensor degradation caused by harsh environments. In this context, soft-sensing technology provides a robust and reliable approach to accurately estimate SC during dredging processes. In this work, a soft-sensing method is developed based on a novel self-balancing Physics-informed Long-Short Term Memory (PILSTM) Network. First, the Least Absolute Shrinkage and Selection Operator (LASSO) is employed to analyze the correlation between SC and other monitored dredger signals, thereby identifying the most relevant signals for SC estimation. Second, to address the lack of explicit physical dynamics that governs SC evolution, a Deep Hidden Physics Model (DeepHPM) is leveraged to infer the underlying physical relationships from historical data. This model is then integrated with a Long Short-Term Memory Network (LSTM) to form the PILSTM, enabling accurate and physics-consistent SC estimation. Then, a self-balancing strategy is implemented to automatically weigh the two competing objectives during the PILSTM training process: maximizing physics consistency with DeepHPM and improving estimation accuracy on the training dataset. Finally, two datasets collected from actual in-field dredger operations are utilized to verify the performance of the proposed soft-sensing method. The results demonstrate its superior accuracy and enhanced physics-consistency compared to other state-of-the-art approaches, achieving R 2 coefficients over 0.85 and 0.9 in the two case studies, respectively, which highlights its effectiveness in practical applications.

Industrial Internet of Things applications like aircraft assembly impose stringent demands on Computing Power Networks. Existing deep reinforcement learning (DRL)-based schedulers not only operate under rigid time-step decision mechanisms but also inadequately handle multi-priority tasks owing to oversimplified queue modeling. To resolve these fundamental limitations, we propose an innovative integrated framework that synergistically combines three components. First, the framework establishes a pioneering formalization of preemptive priority scheduling problem, simultaneously optimizing task response time and violation rate. Second, it incorporates the Counterfactual-Aware Semi-Markov Decision Process (CA-SMDP), which employs counterfactual intervention to tackle temporal credit assignment under event-driven decision epochs. Third, we propose a novel Topology-context fusion Event-driven Scheduler (TESer) where specialized modules for latency minimization and SLA assurance collaboratively achieve optimization synergy. Experimental results demonstrate consistent superiority over state-of-the-art baselines across critical scheduling metrics.

Accurately inferring the causality among variables is essential for diagnosing the root causes of industrial process faults. The multimode characteristics of industrial processes augment the complexity of fault evolution processes, which presenting challenges to traditional methods. To this end, the continuous causal learning framework is proposed for comprehensive root cause diagnosis of multimode processes. Firstly, the temporal feature interference aided causal discovery network is designed to achieve collaborative quantification of multivariate causality, which can facilitate continuous learning while avoiding pairwise causal modeling. Then, to address the “catastrophic forgetting" issue in continuous learning, an elastic weight consolidation strategy is introduced to balance the newly emerged modes and previously learned modes through adaptive parameter regularization during sequential model updating. Subsequently, the root cause diagnosis index impact score is designed to locate the real root cause according to the causality variations from normal to faults with Cramer-von Mises test. Additionally, the propagation paths are identified through the SL metric by involving causality strength and time lag simultaneously. Finally, the proposed method significantly outperform the existing methods in terms of root cause diagnosis on both simulated and real-world multimode processes.

As intelligent equipment is increasingly integrated into safety-critical sectors, the demand for its trustworthiness has become paramount. While AI-based virtual sensors can achieve high accuracy in parameter estimation, they often suffer from poor physical consistency and insufficient uncertainty quantification. To address these issues, this paper proposes the Physics-Residual Transformer with Conformal Prediction (PRT-CP) framework for virtual sensing. The PRT-CP framework utilizes a physics-informed residual design to anchor predictions within fundamental physical constraints, effectively preventing physically inconsistent outputs. Furthermore, it integrates a conformal prediction layer to construct uncertainty intervals with a mathematically proven finite-sample coverage guarantee. To evaluate industrial readiness, we introduce the Composite Trustworthiness Index (CTI), which integrates accuracy, physical consistency and uncertainty quantification quality. Validated on Liquefied Petroleum Gas (LPG) monitoring and Combined Cycle Power Plant (CCPP) case studies, PRT-CP consistently achieves the highest CTI scores, reaching up to 91.40%, while maintaining a strict 95% coverage rate with significantly sharper intervals than Bayesian baselines. This work establishes a robust paradigm for deploying reliable, physics-aligned AI in modern industrial systems.

Tabular data is the most widely used data form in real-world applications, and tree-based models are suitable for it due to their model structures. In practice, it is crucial to quantify predictive uncertainty, and several uncertainty estimation methods have been developed for tree-based models. However, some of them compromise point estimation accuracy or incur high computational overhead. To address this limitation, we propose a variance-based method for predictive uncertainty estimation of tabular data using gradient boosting decision tree (VarBoost). VarBoost estimates variance by analytically calculating statistical characteristics based on gradient boosting strategy, ensuring efficient and accurate variance estimation while maintaining point estimation performance. To avoid overfitting or underfitting, hyperparameters are determined via cross-validation using proper scoring rules that balance calibration and sharpness. The effectiveness of VarBoost is validated by comparisons with several state-of-the-art uncertainty estimation methods on a collection of UCI tabular datasets. Experimental results demonstrate that VarBoost achieves superior performance in both point estimation and uncertainty estimation.

Single-sided aluminized polyimide films (SAPF) are widely used in thermal management of aerospace systems. Although the reliability of SAPF in space environments has been thoroughly studied, its reliability in ground environments during storage is always ignored, potentially leading to system failure. This paper aims to investigate the reliability of SAPF in storage environments, focusing on the effects of temperature and relative humidity. Firstly, the relationship between the performance degradation of SAPF and aluminum corrosion is identified. Next, considering the presence of two distinct stages in the influence of temperature on aluminum corrosion, a novel degradation model accounting for the degradation mechanism transition is developed. Additionally, a parameter analysis method is proposed for determining SAPF degradation mechanism based on experimental data. Then, a statistical analysis method incorporating an improved rime optimization algorithm is employed for parameter estimation, and the reliability model is established. Experimental results demonstrate that the proposed method effectively identifies two distinct stages in the impact of temperature on SAPF performance degradation. Furthermore, the proposed degradation model outperforms traditional degradation models with unchanged degradation mechanism in terms of degradation prediction accuracy, extrapolation capability and robustness, indicating its suitability for describing the degradation pattern of SAPFs.

Remaining useful life (RUL) prediction is fundamental to prognostics and health management (PHM) in industrial systems. In practical deployment, rotating machinery operates over long service lifecycles under continuously evolving loads and rotational speeds. Such variability, combined with long-term acquisition of monitoring data, gives rise to non-stationary degradation patterns and distributional shifts. These characteristics challenge conventional deep learning-based RUL models, which are typically trained on static datasets and lack adaptability to sequentially arriving operating conditions. Moreover, most existing approaches lack explicit mechanisms for predictive uncertainty quantification in dynamic task environments, restricting their reliability in risk-aware maintenance decision-making. To address these challenges, this paper proposes a hybrid Bayesian learning framework for uncertainty-aware continual RUL prediction in non-stationary industrial settings. The framework employs Bayesian neural networks (BNNs) to jointly model aleatoric and epistemic uncertainties and integrates prior-guided Bayesian knowledge transfer with likelihood-guided generative replay for continual learning. A conditional diffusion-based replay mechanism is introduced to synthesize representative pseudo-samples, together with a dual-uncertainty-driven sample selection strategy that retains informative historical knowledge without storing raw data. Extensive experiments on multiple run-to-failure bearing and gear datasets under diverse operating conditions demonstrate that the proposed method achieves superior RUL prediction accuracy, enhanced robustness to distributional shifts, and more reliable uncertainty quantification in continual learning scenarios, underscoring its suitability for long-term industrial monitoring and predictive maintenance.

Advances in sensor technology and the Industrial Internet of Things (IIoT) have enabled the collection of large-scale monitoring data, facilitating intelligent remaining useful life (RUL) prediction for industrial equipment. However, accurate RUL prediction in distributed environments faces two major challenges. First, the scarcity of data and limited computational resources at edge clients hinder the development of robust RUL models, while privacy constraints prohibit centralized data sharing. Second, distribution shifts across client machines severely limit the model’s ability to generalize to unknown operating conditions (OCs) and equipment. To address these challenges, this article proposes a cloud-edge collaboration (CEC) federated invariance and specificity domain generalization (DG) (CEC-FedISDG) method. CEC-FedISDG integrates both domain-invariant and domain-specific predictive knowledge within a unified cloud-edge federated learning (FL) framework. This design enables the model to exploit the broad generalizability of invariant features while retaining domain-specific predictive capabilities. Specifically, a progressive invariance refinement (PIR) module is developed to gradually strengthen domain-invariant features while preserving privacy through a two-stage learning process. Subsequently, a dynamic specificity selection (DSS) module is designed. It dynamically integrates the outputs of private-domain regressors that retain domain specificity utilizing a domain classifier, adapting weights to test samples, thereby improving RUL prediction accuracy. Experimental evaluations on two bearing datasets and a real-world industrial wind turbine gearbox (WTG) dataset demonstrate that the CEC-FedISDG achieves superior generalization performance while adhering to strict privacy preservation requirements.

Deep Learning (DL) models processing images to recognize the health state of large infrastructure components can exhibit biases and rely on non-causal shortcuts. eXplainable Artificial Intelligence (XAI) can address these issues but manually analyzing explanations generated by XAI techniques is time-consuming and prone to errors. This work proposes a novel framework that combines post-hoc explanations with semi-supervised learning to automatically identify anomalous explanations that deviate from those of correctly classified images and may therefore indicate model abnormal behaviors. This significantly reduces the workload for maintenance decision-makers, who only need to manually reclassify images flagged as having anomalous explanations. The proposed framework is applied to drone-collected images of insulator shells for power grid infrastructure monitoring, considering two different Convolutional Neural Networks (CNNs), GradCAM explanations and Deep Semi-Supervised Anomaly Detection. The average classification accuracy on two faulty classes is improved by 8 % and maintenance operators are required to manually reclassify only 15 % of the images. We compare the proposed framework with a state-of-the-art approach based on the faithfulness metric: the experimental results obtained demonstrate that the proposed framework consistently achieves F 1 scores larger than those of the faithfulness-based approach. Additionally, the proposed framework successfully identifies correct classifications that result from non-causal shortcuts, such as the presence of ID tags printed on insulator shells. • Imaging and Deep Learning (DL) for large infrastructure fault diagnostics has emerged. • eXplainable Artificial Intelligence (XAI) outputs enhance DL models trustworthiness. • Processing XAI output by experts is time consuming and error prone. • We develop a methodology to automatically process XAI outputs. • It identifies misclassifications and shortcuts in classifications of insulator images.

The electrification strategies that are being designed to meet sustainability objectives and rising energy demands pose significant challenges for power systems worldwide and require Transmission Expansion Planning (TEP). This study adopts a risk-informed approach to TEP, formulated as a multi-objective optimization problem that concurrently minimizes systemic risks and expansion costs. Given the intractability of this problem with conventional solvers, we turn to artificial intelligence techniques. In particular, we conceptualize power grids as graphs and introduce a goal-oriented graph generation methodology using deep reinforcement learning. We extend welfare-Q learning, a modified variant of Q-learning tailored to yield high rewards across multiple dimensions, by incorporating geometric deep learning for function approximation. This allows us to account for system security while minimizing grid expansion costs. Notably, system risk is evaluated by incorporating a Graph Neural Network (GNN) cascading failure meta-model into the proposed approach. The TEP method is applied to the IEEE 118-bus system, and the efficacy of this novel technique is compared against the state of the art. We conclude that the deep reinforcement learning method can compete with established methods for multi-objective optimization, identifying expansion strategies that improve system security at reduced costs. Furthermore, we test the robustness of the meta-model against topology changes in the transmission network, demonstrating its applicability to novel grid configurations.

In industry, the reliability of rotating machinery is critical for production efficiency and safety. Current methods of prognostics and health management (PHM) often rely on task-specific models, which face significant challenges in handling diverse datasets with varying signal characteristics, fault modes and operating conditions. Inspired by advancements in generative pretrained models, we propose RmGPT, a unified model for diagnosis and prognosis tasks. RmGPT introduces a novel generative token-based framework, incorporating Signal Tokens, Prompt Tokens, Time-Frequency Task Tokens, and Fault Tokens to handle heterogeneous data within a unified model architecture. We leverage self-supervised learning for robust feature extraction and introduce a next signal token prediction pretraining strategy, alongside efficient prompt learning for task-specific adaptation. Extensive experiments demonstrate that RmGPT significantly outperforms state-of-the-art algorithms, achieving near-perfect accuracy in diagnosis tasks and exceptionally low errors in prognosis tasks. Notably, RmGPT excels in few-shot learning scenarios, achieving 82% accuracy in 16-class one-shot experiments, highlighting its adaptability and robustness. This work establishes RmGPT as a powerful PHM foundation model for rotating machinery, advancing the scalability and generalizability of PHM solutions. Code is available at:https://github.com/Pandalin98/RmGPT.

In practical applications, not all nodes in networked swarm systems are cooperative. The noncooperative nodes are transformed from healthy ones because of cyber-attacks launched by malicious adversaries, hardware faults caused by low reliability individuals, or communication delay. In this article, an approximate fault detection method using residual threshold is shown to judge which agent is cooperative or not, and to reconstruct the communication topology. Then, the event-triggered technique is utilized to design the multileaders formation tracking protocol with resilience to noncooperative nodes. The Zeno behavior is considered to constrain the trigger condition. Finally, the comparison simulation results show the effectiveness and advantage for the proposed secure control method.

As the adoption of electric vehicles (EVs) within electrified road networks (ERNs) continues to grow, the criticality of fast-charging stations (FCSs) to support this growth and alleviate range anxiety for EV users is increasingly evident. Due to the complex operating environment of FCSs, their reliable, safe operation faces significant challenges, from exposure to natural disasters, deliberate attacks, and technical failures. This leads to the need to analyze the vulnerability of FCSs within ERNs. This paper provides a mathematical framework to assess the vulnerability of the ERN to intentional attacks and identify critical FCSs within ERNs. We propose a System Optimal Dynamic Mixed Traffic Flow Assignment Model (SODTA) game-theoretical for an ERN where electric and fuel vehicles coexist. Building upon this model, an attacker-defender (AD) framework is established for assessing the vulnerability of FCSs within the ERN. We transform this AD problem into a mixed integer linear programming (MILP) one to solve it efficiently. Two indices are proposed to evaluate the vulnerability and the reallocation of charging loads within the ERN. Finally, we apply the proposed framework to conduct a vulnerability assessment of the ERN in North Carolina, USA. The key findings are as follows: 1) As the EV penetration increases, the studied ERN becomes more vulnerable to attacks. 2) With limited attack resources, mitigating performance losses over time is possible through charging load redistribution. 3) Variation in overall vulnerability levels with increasing attack resources reveals distinct patterns, ranging from mitigable to unmitigable vulnerability.

Federated learning (FL)-based fault diagnosis is being widely developed. However, most of the existing FL methods may suffer from two drawbacks: 1) they are limited to a single diagnosis task, and this may be insufficient when comprehensive health status information is needed and 2) most of them work offline, thus neglecting the useful information contained in newly collected operation data. For this end, this article proposes a multitask federated incremental learning (multitask-FIL) framework. First of all, a multitask feature sharing network is established by assigning the extracted general features to different downstream tasks, so that the joint loss function is obtained for subsequent collaborative training. Then, Q-learning algorithm is used to select the incremental sequences for all the parties from real-time running data, which can facilitate the model performance by involving additional data information and preferred parties. After that, the incremental weight of each party is dynamically adjusted according to the loss depth and sample size in each round of communication, so that the effects of different parties can be quantified throughout the model iteration and aggregation process. Finally, experiments on three challenging cases are performed to show that the proposed method has strong multitask collaboration capability.

This article concentrates on the design of an active resilient formation tracking control strategy for heterogeneous swarm systems (HSS) under malicious cyber-attacks. The attack signals, which are injected into both actuator and sensor randomly, can be detected and estimated by using an observer-based attack estimation scheme. The compromised measured outputs of individuals in the swarm are used to design the distributed secure control protocol and the consensus-based formation condition. For each follower, a compensator is designed to address the secure formation tracking problem for HSS using an approximate model following strategy. Moreover, the event-triggered technique is utilized to achieve the time-varying formation and to track the leader, and saves the communication resource in practice. Finally, a numerous simulation for a heterogeneous swarm system with three different dynamics nodes is presented to verify the effectiveness of the proposed secure approach.

Prognostics and health management (PHM) has garnered significant attention in industrial fields, particularly due to its successful application in managing battery degradation. However, current approaches are inadequate in addressing multiple thresholds, including both theoretical formulation and practical computational complexity. These limitations hinder the development and implementation of threshold-varying assessments, thereby impeding the advancement of PHM application. This article investigates prognostic applications with different failure thresholds and highlights the importance of failure threshold selection. In addition, theoretical evaluation and analysis are provided for multiple threshold settings, encompassing both discrete and continuous series. This introduces a novel technical domain for prognostic applications. The effectiveness of threshold-varying assessment is verified with several different approaches on real battery degradation experiments. Furthermore, we demonstrate the practical significance of threshold-varying assessments in enabling on-demand scheduling for maintenance or replacement of spare parts. Most importantly, to meet the real-time requirements of practical prognostic applications, this article also discusses the computational complexity of threshold-varying assessment and finds an applicable solution for this common difficulty.

To achieve fault diagnosis and prognosis, obtaining adequate and reliable life-cycle data is essential. However, this poses a challenge in current high-reliable Internet of Things (IoT) systems. Fortunately, accelerated degradation testing (ADT) can be employed to overcome this hurdle. Nevertheless, a dependable testing and measuring technique is required to construct an accurate model for ADT. This testing method plays a vital role in evaluating fault diagnosis, prognosis, lifetime, and maintenance decisions for reliable products under operational stress. To ensure effective testing, it is crucial to utilize appropriate models that account for the individual heterogeneity of products. However, the commonly used single stochastic models in ADT overlook the impact of this condition in real-world applications, resulting in misspecification problem. To address this limitation, we propose a novel mixed stochastic process model that integrates multi-Wiener processes and dynamic weights. In addition, we leverage interval analysis to analyze system lifetime, considering the limited data size. The estimation of unknown parameters in our mixed model is achieved using the Metropolis–Hastings algorithm. By analyzing stress relaxation data from electrical connectors, we demonstrate the superior accuracy of our mixed model over conventional single stochastic models in ADT.

Underwater thrusters, vital components in various underwater vehicles, have been extensively studied for fault identification and classification. However, the automatic and unsupervised assessment of fault levels remains largely unexplored. This article presents a novel approach integrating physical information into a data-driven architecture and training process, enabling automated and unsupervised fault identification and evaluation. The process begins with constructing a physical model of the thruster and estimating its low-fidelity current based on the vehicle's velocity and the thruster's rotational speed. An improved SimGAN is then utilized, in combination with the vehicle's motion state, to map this low-fidelity current to its high-fidelity counterpart in the physical space. Statistical features are extracted from the absolute errors between the high-fidelity current and the measured current as conditions for the discriminator. The framework achieves thruster malfunction identification and assessment by analyzing the discriminator's output. Ocean trial data validate the effectiveness of this framework, with experimental results showing its satisfactory performance in fault identification, level evaluation, computational complexity, and robustness when compared with advanced methods.

Accelerated degradation testing (ADT) data typically exhibit a time-stress-dependent structure, as well as random uncertainties due to time-varying effects and unit-to-unit variations. Existing ADT models based on Brownian motion with drift have successfully represented the fault/failure-based degradation behavior and random uncertainty by assuming that the drift parameter follows a Gaussian distribution. However, these models often lack robustness to outliers, leading to distorted analysis, affecting parameter estimation, model accuracy, decision-making, risk assessment, and potentially overlooking the influence of stress factors. A novel robust ADT model based on the Wiener process and its corresponding lifetime analysis method are proposed to address these issues. The proposed approach improves upon traditional ADT models by making the drift parameter follow a$t$-distribution rather than a Gaussian distribution, which can reduce sensitivity to outliers in real degradation processes. In addition, the proposed method allows for the simultaneous consideration of time-stress-dependent factors in the ADT model, facilitating the derivation of a closed-form robust ADT formulation. Subsequently, the lifetime is analyzed based on the ADT model using the first hitting time method in a probabilistic framework. The proposed method is applied to stress relaxation data of electrical connectors and compared to three other common methods.

In this paper, we study network reliability in relation to a periodic time-dependent utility function that reflects the system’s functional performance. When an anomaly occurs, the system incurs a loss of utility that depends on the anomaly’s timing and duration. We analyze the long-term average utility loss by considering exponential anomalies’ inter-arrival times and general distributions of maintenance duration. We show that the expected utility loss converges in probability to a simple form. We then extend our convergence results to more general distributions of anomalies’ inter-arrival times and to particular families of non-periodic utility functions. To validate our results, we use data gathered from a cellular network consisting of 660 base stations and serving over 20k users. We demonstrate the quasi-periodic nature of users’ traffic and the exponential distribution of the anomalies’ inter-arrival times, allowing us to apply our results and provide reliability scores for the network. We also discuss the convergence speed of the long-term average utility loss, the interplay between the different network’s parameters, and the impact of non-stationarity on our convergence results.

This article focuses on multitask selective maintenance (SM) for multistate complex systems, with the goal of selecting subsets of feasible maintenance actions on multitask systems simultaneously due to limited resources. For each task, system characteristic comprises of various configurations, such as series, parallel, bridge, and complex, Weibull distribution, and multiple states; maintenance characteristic includes perfect maintenance, imperfect maintenance (IM), and minimal repair. Considering these realistic issues, this article introduces a reliability evaluation approach, including Markov chain, universal generating function, and IM age reduction model. The challenge of solving such kind of problems lies not only in the reliability estimation, but also in the solution method. Since it is the first time to solve the multitask SM problem, this article tailors a novel multifactorial evolutionary algorithm, with an improved associate mating. In our algorithm, a similarity-based task selection mechanism tries to determine the intensity between intertask self-evolution and intertask knowledge transfer, based on the relatedness between tasks; a feedback-based task transfer mechanism adjusts the transfer intensity, with regard to convergence and diversity. Numerical experiments verify the effectiveness of the proposed method compared with the original one.

Prognostics and Health Management (PHM) aims at assessing and predicting the degradation of structures, systems and components (SSCs), so as to allow anticipating failures and, thus, avoiding accidents. In this contribution, we recognize the efforts by the researchers and experts of the Reliability Society in the development of PHM methods and briefly point at some of the challenges for the application of PHM in practice.

Reliability optimization aims at maximizing system reliability and related metrics, while minimizing the cost allocated to maximize the reliability and respecting other design constraints like weight and volume. It has been an active research domain since the 1960s, and various optimization problems have been formulated and solution techniques proposed. In this contribution, we briefly recall the evolution of the approaches to reliability optimization from the early analytical ones for simple systems to the current advanced computational and empirical techniques addressing more practical problems. Much of this evolution is the result of the efforts by the researchers and experts of the Reliability Society.

Uncertainty is a practical issue in system design optimization because some characteristics of components, such as reliability and cost, cannot be determined precisely in many situations. Considering the imprecise characteristics of components, few works have focused on the multi-objective optimization for the redundancy allocation due to the challenges of comparing multi intervals. To tackle the issue, a novel angle-based bi- objective redundancy allocation algorithm is proposed in this study, introducing three original contributions: 1) An angle-based interval crowding distance (ICA) is especially designed for effective performance and reduced computational time; 2) Two techniques are applied to tackle the problem: An elite selection for mutation is presented for generating better offsprings; A penalty-guided constraint handling technique is introduced for converting the problem into an unconstrained one. 3) Since a set of optimal solutions is obtained by the proposed method and no preference on uncertainties is provided, this paper proposes a novel knee interval method to help DMs make a decision. To be specific, the proposed ICA can describe the distribution of the whole population intuitively and effectively, considering not only the angle between two compared individuals but also the angle range of the interval values. The computational results from two typical experiments demonstrate that the proposed algorithm is more efficient than other state-of-the-art algorithms, generating Pareto sets with less repeating individuals, stronger convergence, wider distribution, less imprecision, and reduced computational time. Note to Practitioners—This article is motivated by two practical problems in multi-objective redundancy allocation in presence of interval uncertainty: First, this paper tries to solve the multi-objective redundancy allocation problem with the imprecise characteristics of components, which is rarely considered in the field of reliability optimization design. Second, the calculation of the crowding distance needs extra time cost and is less efficient. To tackle this issue, an interval crowding angle is especially designed, considering not only the angle between two compared individuals, but also the angle range of the interval values. The proposed method can be embedded in most multi-objective interval evolutionary algorithms to compute the diversity of the individuals. The goal of this study is to allocate the economy and high-reliable components for practitioners. The computational results verify its effectiveness and efficiency. Besides, in many cases the practitioners know only few or no preferences, this paper proposes a knee point analysis of interval values that allows practitioners to select the optimal solution with large hypervolume and less imprecision among a set of solutions.

Deep learning methods of fault diagnostics require the availability of a large amount of labeled data for training, i.e., signal values corresponding to known degradation and fault states. Furthermore, the distribution of the training data should be similar to that of the (test) data collected in the field. Since these conditions are typically not satisfied in most industrial applications, this article develops a deep multiadversarial conditional domain adaptation network. The main original contribution lies in a novel method to align, class by class, the weighted marginal data distributions using multiple domain discriminators. The network allows overtaking the classification underperformance caused by the problem of negative transfer, which is typically encountered when only few training data of some of the classes are available. The proposed method is shown to outperform other state-of-the-art methods on two cross-domain fault diagnostic case studies, verified by applying Friedman and Holm post-hoc tests.

Degradation modeling and prognostics serve as the basis for system health management. Recently, various sensors provide plentiful monitoring data that can reflect the system status. A multitude of feature fusion techniques based on multisensor data have been proposed to generate a composite health index (HI) for prognostics, which can represent the underlying degradation mechanism. Most existing methods have used linear fusion models and neglected the practical requirements for HI construction, which are insufficient to reveal the nonlinear relations among features and difficult to obtain accurate HIs for complicated systems. This study proposes a novel feature fusion-based HI construction method with deep learning and multiobjective optimization. Multiple degradation features are fused by a deep neutral network (DNN). Several desired properties that the HIs should have for prognostics are adopted to formulate the objective functions of DNN training. To balance the spatial complexity and performance of the fusion model, a multiobjective optimization model is generated for training the DNN. Then, a generalized nonlinear Wiener process model is used to predict the remaining useful life with the resulted HIs. Finally, two cases are analyzed to illustrate the effectiveness and robustness of the proposed method.

The uncertainties in the degradation process have always been regarded as a major challenge in the practical applications of Prognostics and Health Management. This article discusses the uncertainties in the degradation process of Lithium-ion batteries, points out their potential consequences in practical applications, and then we summarizes some commonly adopted aftertreatment solutions for them. To proceed realistic analysis, we present a non-accelerated degradation experiment with several Lithium-ion batteries. The experiment lasted for more than ten months and the data highlighted the uncertain fluctuations and periodic waves in SOH degradation process. Furthermore, this article reveals the delayed correlation between SOH degradation and changing environmental temperature. Finally, we provide some possible solutions for guiding practical applications of our finding.

The papers in this special section focus on increasing interests in the development and implementation of advanced artificial intelligence (AI) and machine learning (ML) methods for tackling the reliability and system health prognostics challenges in various industrial applications.

Due to the scarcity of abnormal condition data in components of transportation systems, only normal condition data are typically used to train models for anomaly detection. One of the main challenges is the difficulty of properly representing the data distribution which is typically non-smooth, high-dimensional and on a manifold. This work develops an anomaly detection model based on an Auto-Encoder (AE) formed by the generator of a Generative Adversarial Network (GAN) and an auxiliary encoder to capture the sophisticated data structure. The reconstruction error of the AE is, then, used as anomaly score to detect anomalies. Additionally, an adaptive noise is added to the data to make easier the GAN optimization, an AdaBoost-based ensemble learning scheme is used to improve detection performance and a new approach for setting the hyperparameters of the AE-GAN model based on the derivation of a lower bound of the Jensen-Shannon divergence between generator and normal condition data distributions is developed. The method has been applied to synthetic and real data collected from automatic doors of high-speed trains.

The multimission selective maintenance problem (MSMP) for repairable systems has received increasing attention in recent years. The problem amounts to selecting a subset of feasible maintenance actions, in view of the resource limitations. For considering the realistic case of the imperfect maintenance, this article introduces a hybrid imperfect maintenance model, which is more realistic to evaluate the system reliability. The challenge of solving such kind of problems lies not only in the reliability estimation, but also in the solution method of the maintenance selection. Such decision-making problem can be effectively formulated using the Markov decision process, but it is difficult to apply current methods for solving the engineering systems with large action decision spaces. In order to solve this issue, this work puts forth a novel hybrid algorithm for the MSMP in a large multicomponent system. In the proposed method, a discrete differential evolution algorithm is developed for searching the optimal maintenance action in large-scale discrete action spaces and the deep Q-network method is utilized to approximate the effectiveness of maintenance actions and facilitate the agent training. The experiments, based on a large-scale coal transportation system, verify the effectiveness of the proposed method compared with LSDQN and differential evolution.

With the increasing penetration of electric vehicles (EVs), more and more interactions appear between the transportation system and the power system, which might provide new hazards and channels for the proliferation of failures across the boundaries of the individual systems. In this context, this paper proposes an integrated risk assessment framework for an electric power system, considering scenarios that involve the electrified transportation system enabled by EVs charging technology in New York (NY) State. Firstly, scenarios in the transportation network of NY State, e.g. of reduced capacity and incident, are generated by a Monte Carlo non-sequential algorithm. Then, the cell transmission model (CTM) is used to simulate the evolution of the traffic flows under such scenarios. This allows evaluating the spatial-temporal EV charging loads in different areas of the electrified transportation system of NY State. Correspondingly, the running parameters in the studied power system are updated by the alternative current (AC) power flow model. Finally, the risk for the power system coming from the transportation system scenarios is assessed within a probabilistic risk analysis framework. The proposed integrated risk assessment framework is able to model the propagation of the effects of scenarios in the transportation system onto the power system of NY State and quantify the consequences. A real test case is used to illustrate the proposed framework.

This article presents a new two-stage stochastic model for the simultaneous optimization of component test plan and redundancy allocation. The optimal number of tests is determined for each component of the system in the first stage, and the system redundancy configuration is optimized in the second stage, after the realization of the system working condition. Scenario analysis is used for modeling variations in working conditions. A multiobjective reliability problem is formulated for series-parallel systems, and the allocation of redundancy is optimized by a genetic algorithm. The results of the feedback control system demonstrate the performance of the proposed model.

The Industry 4.0 paradigm is boosting the relevance of predictive maintenance (PdM) for manufacturing and production industries. PdM strongly relies on Internet of Things (IoT), which digitalizes the physical actions allowing human-to-human, human-to-machine, and machine-to-machine connections for intelligent perception. Several issues still need to be addressed for reaching the maturity stage for the widespread application of PdM. To do this, IoT needs to be empowered with data science capabilities, to reach the ultimate objective of digitalization, which is supporting decision making to optimally act on the physical systems. In this article, we present a comprehensive outlook of the current PdM issues, with the final aim of providing a deeper understanding of the limitations and strengths, challenges and opportunities of this dynamic maintenance paradigm. This is done through extensive research and analysis of the scientific and technical literature. On this basis, this article outlines some main research issues to be addressed for the successful development and deployment of IoT-enabled PdM in industry.

We present a novel method for concept drift detection, based on: 1) the development and continuous updating of online sequential extreme learning machines (OS-ELMs) and 2) the quantification of how much the updated models are modified by the newly collected data. The proposed method is verified on two synthetic case studies regarding different types of concept drift and is applied to two public real-world data sets and a real problem of predicting energy production from a wind plant. The results show the superiority of the proposed method with respect to alternative state-of-the-art concept drift detection methods. Furthermore, updating the prediction model when the concept drift has been detected is shown to allow improving the overall accuracy of the energy prediction model and, at the same time, minimizing the number of model updatings.

A sequential Bayesian approach is presented for remaining useful life (RUL) prediction of dependent competing failure processes (DCFP). The DCFP considered comprises of soft failure processes due to degradation and hard failure processes due to random shocks, where dependency arises due to the abrupt changes to the degradation processes brought by the random shocks. In practice, random shock processes are often unobservable, which makes it difficult to accurately estimate the shock intensities and predict the RUL. In the proposed method, the problem is solved recursively in a two-stage framework: in the first stage, parameters related to the degradation processes are updated using particle filtering, based on the degradation data observed through condition monitoring; in the second stage, the intensities of the random shock processes are updated using the Metropolis-Hastings algorithm, considering the dependency between the degradation and shock processes, and the fact that no hard failure has occurred. The updated parameters are, then, used to predict the RUL of the system. Two numerical examples are considered for demonstration purposes and a real dataset from milling machines is used for application purposes. Results show that the proposed method can be used to accurately predict the RUL in DCFP conditions.

A modeling and optimization framework for the maintenance of systems under epistemic uncertainty is presented in this paper. The component degradation processes, the condition-based preventive maintenance, and the corrective maintenance are described through piecewise-deterministic Markov processes in consideration of degradation dependence among degradation processes. Epistemic uncertainty associated with component degradation processes is treated by considering interval-valued parameters. This leads to the formulation of a multi-objective optimization problem whose objectives are the lower and upper bounds of the expected maintenance cost, and whose decision variables are the periods of inspections and the thresholds for preventive maintenance. A solution method to derive the optimal maintenance policy is proposed by combining finite-volume scheme for calculation, differential evolution, and nondominated sorting differential evolution for optimization. An industrial case study is presented to illustrate the proposed methodology.

Interval availability, defined as the fraction of time that a system is operational during a period of time [0, T], is an important indicator of system performance, especially for industries with a Service Level Agreement, e.g., telecommunication industry, computer industry, etc. Most existing methods to compute interval availability are based on numerical simulations. In this paper, we present a new analytical solution for interval availability of Markov repairable systems. Three interval availability indexes, i.e., interval availability, interval availability in a general interval, and interval availability in multiple intervals, are considered. The interval availability indexes are derived based on aggregated stochastic processes and the results are obtained in closed form using Laplace transforms. A numerical example is presented and the results are compared with those of Monte Carlo simulation. The developed methods are applied to calculate the interval availability of a fault-tolerant database system from the literature.

The development of data-driven models for the identification of the degradation state of industrial components is challenged by several issues, such as the unavailability of large datasets containing historical data, the presence of measurement noise, and the intrinsic stochasticity of the degradation process. The proposed method merges the degradation state assessments provided by self-organizing maps (SOMs) trained using data collected from the component under test (component-based) with that provided by SOMs trained using data collected from a fleet of similar components (population-based). Within an ensemble approach, the outcomes of the SOM models are then aggregated using a dynamic weights proportional method based on the individual model local performances in situations similar to the one under test. The proposed SOM-based ensemble approach has been verified with respect to an experimental case study concerning the identification of the degradation state of insulated-gate bipolar transistors, which are known as one of the most critical components in power systems.

Traditional quantitative risk assessment methods (e.g., event tree analysis) are static in nature, i.e., the risk indexes are assessed before operation, which prevents capturing time-dependent variations as the components and systems operate, age, fail, are repaired and changed. To address this issue, we develop a dynamic risk assessment (DRA) method that allows online estimation of risk indexes using data collected during operation. Two types of data are considered: statistical failure data, which refer to the counts of accidents or near misses from similar systems and condition-monitoring data, which come from online monitoring the degradation of the target system of interest. For this, a hierarchical Bayesian model is developed to compute the reliability of the safety barriers and a Bayesian updating algorithm, which integrates particle filtering (PF) with Markov Chain Monte Carlo, is developed to update the reliability evaluations based on both the statistical and condition-monitoring data. The updated safety barriers reliabilities, are, then, used in an event tree (ET) for consequence analysis and the risk indexes are updated accordingly. A case study on a high-flow safety system is conducted to demonstrate the developed methods. A comparison to the DRA method which only uses statistical failure data shows that by introducing condition-monitoring data on the system degradation process, it is possible to capture the system-specific characteristics, and, therefore, provide a more complete and accurate description of the risk of the target system.

This paper proposes an adaptive incremental ensemble of extreme learning machines for fault diagnosis. The diagnostic system contains a data processing unit which aims to progressively generate discriminant features from the vibration signals for decision making. The decision making unit receives a few sets of labeled discriminant features in a chunk by chunk manner, incrementally learns the features-faults relations, dynamically diagnoses multiple bearing defects, and adaptively adjusts itself to learn new concept classes. This adaptive ensemble system is based on incremental learning of multiple extreme learning machines that are able to consult together and adjust themselves based on their confidence in the decision making. Extreme learning machines are used to construct the hybrid ensemble due to their good controllability and fast learning rate. Experimental results show the efficiency of the hybrid diagnostic system. The proposed diagnostic system is applied to diagnosing bearing defects in an induction motor.

A variety of prognostic and health management (PHM) algorithms have been developed in the last years and some metrics have been proposed to evaluate their performances. However, a general framework that allows us to quantify the benefit of PHM depending on these metrics is still lacking. We propose a general, time-variant, analytical model that conservatively evaluates the increase in system availability achievable when a component is equipped with a PHM system of known performance metrics. The availability model builds on metrics of literature and is applicable to different contexts. A simulated case study is presented concerning crack propagation in a mechanical component. A simplified cost model is used to compare the performance of predictive maintenance based on PHM with corrective and scheduled maintenance.

In accelerated degradation testing (ADT), test data from higher than normal stress conditions are used to find stochastic models of degradation, e.g., Wiener process, Gamma process, and inverse Gaussian process models. In general, the selection of the degradation model is made with reference to one specific product and no consideration is given to model uncertainty. In this paper, we address this issue and apply the Bayesian model averaging (BMA) method to constant stress ADT. For illustration, stress relaxation ADT data are analyzed. We also make a simulation study to compare the s-credibility intervals for single model and BMA. The results show that degradation model uncertainty has significant effects on the p-quantile lifetime at the use conditions, especially for extreme quantiles. The BMA can well capture this uncertainty and compute compromise s-credibility intervals with the highest coverage probability at each quantile.

This paper proposes a methodology for run-time safety management in Smart Work Environments (SWE), based on a semantic assessment of the risks, based on defined indicators. To assess SWE safety, we define a novel ontological model for representing its characteristics and introduce a metric for run-time evaluation of the SWE safety level. A description logic reasoner is used to automatically check the consistency of the proposed ontology. The novelty of the approach lies in treating safety at run-time.

The increasing liberalization of European electricity markets, the growing proportion of intermittent renewable energy being fed into the energy grids, and also new challenges in the patterns of energy consumption (such as electric mobility) require flexible and intelligent power grids capable of providing efficient, reliable, economical, and sustainable energy production and distribution. From the supplier side, particularly, the integration of renewable energy sources (e.g., wind and solar) into the grid imposes an engineering and economic challenge because of the limited ability to control and dispatch these energy sources due to their intermittent characteristics. Time-series prediction of wind speed for wind power production is a particularly important and challenging task, wherein prediction intervals (PIs) are preferable results of the prediction, rather than point estimates, because they provide information on the confidence in the prediction. In this paper, two different machine learning approaches to assess PIs of time-series predictions are considered and compared: 1) multilayer perceptron neural networks trained with a multiobjective genetic algorithm and 2) extreme learning machines combined with the nearest neighbors approach. The proposed approaches are applied for short-term wind speed prediction from a real data set of hourly wind speed measurements for the region of Regina in Saskatchewan, Canada. Both approaches demonstrate good prediction precision and provide complementary advantages with respect to different evaluation criteria.

This paper presents the statistical characterization of the oxidation degradation mechanism affecting the nozzles of turbines operated in Oil&Gas utilities. The degradation mechanism is modeled as a four-state, continuous-time semi-Markov process with Weibull distributed transition times. Maximum likelihood estimation is used to infer the parameters of the model from an available set of field data, whereas a numerical approach to estimate the Fisher information matrix is used to characterize the uncertainty in the estimates. The estimates obtained are, then, utilized to compute the probabilities of occupying the four degradation states over time and the corresponding uncertainties. A case study is shown, dealing with real field data.

In this paper, we propose two metrics, i.e., the optimal repair time and the resilience reduction worth, to measure the criticality of the components of a network system from the perspective of their contribution to system resilience. Specifically, the two metrics quantify: 1) the priority with which a failed component should be repaired and re-installed into the network and 2) the potential loss in the optimal system resilience due to a time delay in the recovery of a failed component, respectively. Given the stochastic nature of disruptive events on infrastructure networks, a Monte Carlo-based method is proposed to generate probability distributions of the two metrics for all of the components of the network; then, a stochastic ranking approach based on the Copeland's pairwise aggregation is used to rank components importance. Numerical results are obtained for the IEEE 30-bus test network and a comparison is made with three classical centrality measures.

In this paper, we propose a method for online assessing the performance of a prognostic approach in situations of very poor knowledge on the degradation process. In particular, we deal with cases in which the entire degradation process, from the beginning of the operation until failure, has never been observed and, thus, the traditional offline performance metrics cannot be applied. The proposed method is applied on a prognostic approach based on a particle filter and optimized tuning kernel smoothing (PF-OTKS). Case studies regarding the degradation of turbine blade and aluminum electrolytic capacitor are considered.

Component importance measures (IMs) are widely used to rank the importance of different components within a system and guide allocation of resources. The criticality of a component may vary over time, under the influence of multiple dependent competing degradation processes and maintenance tasks. Neglecting this may lead to inaccurate estimation of the component IMs and inefficient related decisions (e.g., maintenance, replacement, etc.). The work presented in this paper addresses the issue by extending the mean absolute deviation IM by taking into account: 1) the dependency of multiple degradation processes within one component and among different components; 2) discrete and continuous degradation processes; and 3) two types of maintenance tasks: condition-based preventive maintenance via periodic inspections and corrective maintenance. Piecewise-deterministic Markov processes are employed to describe the stochastic process of degradation of the component under these factors. A method for the quantification of the component IM is developed based on the finite-volume approach. A case study on one section of the residual heat removal system of a nuclear power plant is considered as an example for numerical quantification.

Support vector regression (SVR) is a widely used technique for reliability prediction. The key issue for high prediction accuracy is the selection of SVR parameters, which is essentially an optimization problem. As one of the most effective evolutionary optimization methods, particle swarm optimization (PSO) has been successfully applied to tune SVR parameters and is shown to perform well. However, the inherent drawbacks of PSO, including slow convergence and local optima, have hindered its further application in practical reliability prediction problems. To overcome these drawbacks, many improvement strategies are being developed on the mechanisms of PSO, whereas there is little research exploring a priori information about historical data to improve the PSO performance in the SVR parameter selection task. In this paper, a novel method controlling the inertial weight of PSO is proposed to accelerate its convergence and guide the evolution out of local optima, by utilizing the analytical selection (AS) method based on a priori knowledge about SVR parameters. Experimental results show that the proposed ASPSO method is almost as accurate as the traditional PSO and outperforms it in convergence speed and ability in tuning SVR parameters. Therefore, the proposed ASPSO-SVR shows promising results for practical reliability prediction tasks.

Today's fast-pace evolving and digitalizing World is posing new challenges to reliability engineering. On the other hand, the continuous advancement of technical knowledge and the increasing capabilities of monitoring and computing offer opportunities for new developments in reliability engineering. In this paper, I reflect on some of these challenges and opportunities in research and application. The underlying perspective taken stands on the following: The belief that the knowledge, information, and data (KID) available for the modeling, computations, and analyses done in reliability engineering is substantially grown and continue to do so; The belief that the technical capabilities for reliability engineering have been significantly advanced; The recognition of the increased complexity of the systems, nowadays more and more made of heterogeneous, highly interconnected elements. In line with this perspective, opportunities and challenges for reliability engineering are discussed in relation to degradation modeling and integration of multistate and physics-based models therein, accelerated degradation testing, component-, system- and fleet-wide prognostics and health management in evolving environments. The paper is not a review, nor a state of the art work, but rather it offers a vision of reflection on reliability engineering, for consideration and discussion by the interested scientific community. It does not pretend to give the unique view, nor to be complete in the subject discussed and the related literature referenced to.

Components are often subject to multiple competing degradation processes. This paper presents a reliability assessment framework for multicomponent systems whose component degradation processes are modeled by multistate and physics-based models with limited statistical degradation/failure data. The piecewise-deterministic Markov process modeling approach is employed to treat dependencies between the degradation processes within one component or/and among components. A computational method combining binary decision diagrams (BDDs) and Monte Carlo simulation (MCS) is developed to solve the model. A BDD is used to encode the fault tree of the system and obtain all the paths leading to system failure or operation. MCS is used to generate random realizations of the model and compute the system reliability. A case study is presented, with reference to one branch of the residual heat removal system of a nuclear power plant.

Components are often subject to multiple competing degradation processes. For multicomponent systems, the degradation dependence within one component or/and among components need to be considered. Physics-based models and multistate models are often used for component degradation processes, particularly when statistical data are limited. In this paper, we treat dependence between degradation processes within a piecewise-deterministic Markov process (PDMP) modeling framework. Epistemic (subjective) uncertainty can arise due to the incomplete or imprecise knowledge about the degradation processes and the governing parameters, to take this into account, we describe the parameters of the PDMP model as fuzzy numbers. Then, we extend the finite-volume method to quantify the (fuzzy) reliability of the system. The proposed method is tested on one subsystem of the residual heat removal system of a nuclear power plant, and a comparison is offered with a Monte Carlo simulation solution the results show that our method can be most efficient.

We consider the task of performing prediction with neural networks (NNs) on the basis of uncertain input data expressed in the form of intervals. We aim at quantifying the uncertainty in the prediction arising from both the input data and the prediction model. A multilayer perceptron NN is trained to map interval-valued input data onto interval outputs, representing the prediction intervals (PIs) of the real target values. The NN training is performed by nondominated sorting genetic algorithm-II, so that the PIs are optimized both in terms of accuracy (coverage probability) and dimension (width). Demonstration of the proposed method is given in two case studies: 1) a synthetic case study, in which the data have been generated with a 5-min time frequency from an autoregressive moving average model with either Gaussian or Chi-squared innovation distribution and 2) a real case study, in which experimental data consist of wind speed measurements with a time step of 1 h. Comparisons are given with a crisp (single-valued) approach. The results show that the crisp approach is less reliable than the interval-valued input approach in terms of capturing the variability in input.

This work proposes a comparison of three data-driven signal reconstruction methods, which are Auto-Associative Kernel Regression (AAKR), Fuzzy Similarity (FS), and Elman Recurrent Neural Network (RNN), for fault detection based on the difference between the signal observations and the reconstructions of the signal in normal (typical) operating conditions. The aim is to show the capabilities and drawbacks of the methods, and propose a strategy for the aggregation of their outcomes, to overcome their limitations. For this purpose, the performance of each method is evaluated in terms of fault detection capability, considering accuracy, robustness, and resistance to the spillover effect of the obtained signal reconstructions. The comparison is supported by the application to a real industrial case study regarding temperature signals collected during operation of a rotating machine in an energy production plant. An ensemble of the three methods is proposed to overcome the limitations of the three methods.

The aim of this paper is to address the maintenance optimization problem when the maintenance models encode stochastic processes, which rely on parameters that are imprecisely known, and when these parameters are only determined through information elicited from experts. A genetic algorithms (GA)-based technique is proposed to deal with such uncertainty setting; this approach requires addressing three main issues: i) the representation of the uncertainty in the parameters and its propagation onto the fitness values; ii) the development of a ranking method to sort the obtained uncertain fitness values, in case of single-objective optimization; and iii) the definition of Pareto dominance, for multi-objective optimization problems. A known hybrid Monte Carlo-Dempster-Shafer Theory of Evidence method is used to address the first issue, whereas two novel approaches are developed for the second and third issues. For verification, a practical case study is considered concerning the optimization of maintenance for the nozzle system of a turbine in the Oil & Gas industry.

In this paper, a fuzzy classification approach applying a combination of Echo-State Networks (ESNs) and a Restricted Boltzmann Machine (RBM) is proposed for predicting potential railway rolling stock system failures using discrete-event diagnostic data. The approach is demonstrated on a case study of a railway door system with real data. Fuzzy classification enables the use of linguistic variables for the definition of the time intervals in which the failures are predicted to occur. It provides a more intuitive way to handle the predictions by the users, and increases the acceptance of the proposed approach. The research results confirm the suitability of the proposed combination of algorithms for use in predicting railway rolling stock system failures. The proposed combination of algorithms shows good performance in terms of prediction accuracy on the railway door system case study.

In this paper, we propose to define the problem of predicting the remaining useful life of a component as a binary classification task. This approach is particularly useful for problems in which the evolution of the system condition is described by a combination of a large number of discrete-event diagnostic data, and for which alternative approaches are either not applicable, or are only applicable with significant limitations or with a large computational burden. The proposed approach is demonstrated with a case study of real discrete-event data for predicting the occurrence of railway operation disruptions. For the classification task, Extreme Learning Machine (ELM) has been chosen because of its good generalization ability, computational efficiency, and low requirements on parameter tuning.

We extend a multi-state physics model (MSPM) framework for component reliability assessment by including semi-Markov and random shock processes. Two mutually exclusive types of random shocks are considered: extreme, and cumulative. Extreme shocks lead the component to immediate failure, whereas cumulative shocks simply affect the component degradation rates. General dependences between the degradation and the two types of random shocks are considered. A Monte Carlo simulation algorithm is implemented to compute component state probabilities. An illustrative example is presented, and a sensitivity analysis is conducted on the model parameters. The results show that our extended model is able to characterize the influences of different types of random shocks onto the component state probabilities and the reliability estimates.

In this paper, a novel Dynamic-Weighted Probabilistic Support Vector Regression-based Ensemble (DW-PSVR-ensemble) approach is proposed for prognostics of time series data monitored on components of complex power systems. The novelty of the proposed approach consists in i) the introduction of a signal reconstruction and grouping technique suited for time series data, ii) the use of a modified Radial Basis Function (RBF) kernel for multiple time series data sets, iii) a dynamic calculation of sub-models weights for the ensemble, and iv) an aggregation method for uncertainty estimation. The dynamic weighting is introduced in the calculation of the sub-models' weights for each input vector, based on Fuzzy Similarity Analysis (FSA). We consider a real case study involving 20 failure scenarios of a component of the Reactor Coolant Pump (RCP) of a typical nuclear Pressurized Water Reactor (PWR). Prediction results are given with the associated uncertainty quantification, under the assumption of a Gaussian distribution for the predicted value.

We propose a visual interactive method for the identification of the Prime Implicants (PIs) of dynamic non-coherent systems. Visual interactive methods integrate mathematical and symbolic models with runtime interaction and real-time graphic display, which allow visualizing the underlying physical relationships among process parameters. The proposed method is based on a parallel coordinates data mining tool that relies on an innovative pruning procedure which, on the basis of a proper selection of characteristic features of the accident sequences, retrieves the PIs among the whole set of implicants in terms of process parameter values or component failure states or both. The method is exemplified on an artificial case study, and then applied to the dynamic reliability analysis of the Airlock System of a Canadian Deuterium Uranium reactor.

Proper detection of unknown patterns plays an important role in diagnosing new classes of faults. This can be done by incremental learning of novel information and updating the diagnostic system by appending newly trained fault classifiers in an ensemble design. We consider a new-class fault detector previously developed by the authors and based on thresholding the normalized weighted average of the outputs (NWAO) of the base classifiers in a multi-classifier diagnostic system. A proper tuning of the thresholds in the NWAO detector is necessary to achieve a satisfactory performance. This is done in this paper by specifically introducing a performance function and optimizing it within the necessary trade-off between new class false alarm and new class missed alarm rates, by means of an Invasive Weed Optimization (IWO) algorithm. The optimal NWAO detector is tested with respect to a set of simulated sensor faults in the doubly-fed induction generator (DFIG) of a wind turbine.

Predictive Maintenance (PrM) exploits the estimation of the equipment Residual Useful Life (RUL) to identify the optimal time for carrying out the next maintenance action. Particle Filtering (PF) is widely used as a prognostic tool in support of PrM, by reason of its capability of robustly estimating the equipment RUL without requiring strict modeling hypotheses. However, a precise PF estimate of the RUL requires tracing a large number of particles, and thus large computational times, often incompatible with the need of rapidly processing information for making maintenance decisions in due time. This work considers two different Risk Sensitive Particle Filtering (RSPF) schemes proposed in the literature, and investigates their potential for PrM. The computational burden problem of PF is addressed. The effectiveness of the two algorithms is analyzed on a case study concerning a mechanical component affected by fatigue degradation.

Many engineering systems can perform their intended tasks with various levels of performance, which are modeled as multi-state systems (MSS) for system availability and reliability assessment problems. Uncertainty is an unavoidable factor in MSS modeling, and it must be effectively handled. In this work, we extend the traditional universal generating function (UGF) approach for multi-state system (MSS) availability and reliability assessment to account for both aleatory and epistemic uncertainties. First, a theoretical extension, named hybrid UGF (HUGF), is made to introduce the use of random fuzzy variables (RFVs) in the approach. Second, the composition operator of HUGF is defined by considering simultaneously the probabilistic convolution and the fuzzy extension principle. Finally, an efficient algorithm is designed to extract probability boxes ($p$-boxes) from the system HUGF, which allow quantifying different levels of imprecision in system availability and reliability estimation. The HUGF approach is demonstrated with a numerical example, and applied to study a distributed generation system, with a comparison to the widely used Monte Carlo simulation method.

Birnbaum Importance Measure (IM) allows ranking the components of a system with respect to the impact that their failures have on the system's performance, e.g., its reliability or availability. Such ranking is done in industry to efficiently manage Operation and Maintenance (O&M) activities, and to optimize plant design. In the computation of the Birnbaum IM of the components, uncertainty in the parameters of the system model is often neglected. This neglect may lead to erroneous, possibly non-conservative ranking. In this work, we develop a method based on Possibility Theory (PT) for giving due account to epistemic uncertainties in Birnbaum IMs. An example is given with reference to the components of the Auxiliary FeedWater System (AFWS) of a Nuclear Power Plant.

In this paper, we investigate the feasibility of a strategy of fault detection capable of controlling misclassification probabilities, i.e., balancing false and missed alarms. The novelty of the proposed strategy consists of i) a signal grouping technique and signal reconstruction modeling technique (one model for each subgroup), and ii) a statistical method for defining the fault alarm level. We consider a real case study concerning 46 signals of the Reactor Coolant Pump (RCP) of a typical Pressurized Water Reactor (PWR). In the application, the reconstructions are provided by a set of Auto-Associative Kernel Regression (AAKR) models, whose input signals have been selected by a hybrid approach based on Correlation Analysis (CA) and Genetic Algorithm (GA) for the identification of the groups. Sequential Probability Ratio Test (SPRT) is used to define the alarm level for a given expected classification performance. A practical guideline is provided for optimally setting the SPRT parameters' values.

This paper extends previous works related to the assessment of component degradation, through Markov multi-state physic models. The extension includes the evaluation of the effects of uncertain parameters in the model, and the definition of their importance with respect to their influence on the output of the model. Global Sensitivity Analysis (GSA) is selected as the technique because it enables us to 1) consider the simultaneous effects of parameters variations, and 2) to define importance indexes that allow a ranking of the components. GSA requires a large number of evaluations for specific points, identified by an appropriate design of experiment. To avoid the many costly evaluations, a meta-model is built based on polynomial chaos expansion (PCE). A PCE is a multi-dimensional polynomial approximation of the model with coefficients determined by evaluating the model in a reduced set of predetermined points. Importance index values are then derived directly from the PCE. Because, in the problem considered, the model provides the time-dependent behavior of the state probabilities, the importance indexes are also functions of time. An application is presented, related to the cracking process in an Alloy 82/182 dissimilar metal weld in the primary coolant system of a nuclear power plant.

This paper looks into the vulnerabilities of the electric power grid and associated communication network, in the face of intermittent power generation and uncertain demand within a complex network framework of analysis of smart grids. The perspective is typical for the system of systems analysis of interdependencies in a critical infrastructure (CI), i.e., the smart grid for electricity distribution. We assess how the integration of the two systems copes with requests to increase power generation due to enhanced power consumption at a load bus. We define adequate measures of vulnerability to identify the most limiting communication time delays. We quantify the probability that a reduction in the functionality of the communication system yields a faulty condition in the electric power grid, and find that a factual indicator to quantify the coupling strength between the two networks is the frequency of load-shedding actions due to excessive communication time delay. We evaluate safety margins with respect to communication specifications, i.e., the data rate of the network, to comply with the safety requirements in the electric power grid. Finally, we find a catastrophic phase transition with respect to this parameter, which affects the safe operation of the CI.

Risk analysis models describing aleatory (i.e., random) events contain parameters (e.g., probabilities, failure rates, …) that are epistemically-uncertain, i.e., known with poor precision. Whereas aleatory uncertainty is always described by probability distributions, epistemic uncertainty may be represented in different ways (e.g., probabilistic or possibilistic), depending on the information and data available. The work presented in this paper addresses the issue of accounting for (in)dependence relationships between epistemically-uncertain parameters. When a probabilistic representation of epistemic uncertainty is considered, uncertainty propagation is carried out by a two-dimensional (or double) Monte Carlo (MC) simulation approach; instead, when possibility distributions are used, two approaches are undertaken: the hybrid MC and Fuzzy Interval Analysis (FIA) method and the MC-based Dempster-Shafer (DS) approach employing Independent Random Sets (IRSs). The objectives are: i) studying the effects of (in)dependence between the epistemically-uncertain parameters of the aleatory probability distributions (when a probabilistic/possibilistic representation of epistemic uncertainty is adopted) and ii) studying the effect of the probabilistic/possibilistic representation of epistemic uncertainty (when the state of dependence between the epistemic parameters is defined). The Dependency Bound Convolution (DBC) approach is then undertaken within a hierarchical setting of hybrid (probabilistic and possibilistic) uncertainty propagation, in order to account for all kinds of (possibly unknown) dependences between the random variables. The analyses are carried out with reference to two toy examples, built in such a way to allow performing a fair quantitative comparison between the methods, and evaluating their rationale and appropriateness in relation to risk analysis.

The safety of nuclear power plants can be enhanced, and the costs of operation and maintenance reduced, by means of prognostic and health management systems which enable detecting, diagnosing, predicting, and proactively managing the equipment degradation toward failure. We propose a prognostic method which predicts the Remaining Useful Life (RUL) of a degrading system by means of an ensemble of empirical models. The RUL predictions of the individual models are aggregated through a Kalman Filter (KF)-based algorithm. The method is applied to the prediction of the RUL of turbine blades affected by a developing creep.

The behavior of multi-component engineered systems is typically characterized by transitions among discrete modes of operation and failure, each one giving rise to a specific continuous dynamics of evolution. The detection of the system's mode change time represents a particularly challenging task because it requires keeping track of the transitions among the multiple system dynamics corresponding to the different modes of operation and failure. To this purpose, we implement a novel particle filtering method within a log-likelihood ratio approach here, specifically tailored to handle hybrid dynamic systems. The proposed method relies on the generation of multiple particle swarms for each discrete mode, each originating from the nominal particle swarm at different time instants. The hybrid system considered consists of a hold up tank filled with liquid, whose level is autonomously maintained between two thresholds; the system behavior is controlled by discrete mode actuators whose states are estimated by a Monte Carlo-based particle filter on the basis of noise level, and temperature measurements.

Multistate physics modeling (MSPM) of degradation processes is an approach proposed for estimating the failure probability of components and systems. This approach integrates multistate modeling, which describes the degradation process through transitions among discrete states (e.g., initial, microcrack, rupture, etc.), and physics modeling by (physics) equations that describe the degradation process within the states. In reality, the degradation process is non-Markovian, its transition rates are time-dependent, and the degradation is possibly influenced by uncertain external factors such as temperature and stress. Under these conditions, it is in general difficult to derive the state probabilities analytically. In this paper, we overcome this difficulty by building a simulation model supported by a stochastic Petri net representing the multistate degradation process. The proposed modeling approach is applied to the problem of a nuclear component undergoing stress corrosion cracking. The results are compared with those derived from the state-space enrichment Markov chain approximation method applied in a previous work of literature.

This paper addresses the issue of the classification of accident scenarios generated in a dynamic safety and reliability analyses of a Nuclear Power Plant (NPP) equipped with a Digital Instrumentation and Control system (I&C). More specifically, the classification of the final state reached by the system at the end of an accident scenario is performed by Fuzzy C-Means clustering the Functional Principal Components (FPCs) of selected relevant process variables. The approach allows capturing the characteristics of the process evolution determined by the occurrence, timing, and magnitudes of the fault events. An illustrative case study is considered, regarding the fault scenarios of the digital I&C system of the Lead Bismuth Eutectic eXperimental Accelerator Driven System (LBE-XADS). The results obtained are compared with those of the Kth Nearest Neighbor (KNN), and Classification and Regression Tree (CART) classifiers.

Infrastructure interdependency stems from the functional and logical relations among individual components in different distributed systems. To characterize the extent to which a contingency affecting an infrastructure is going to weaken, and possibly disrupt, the safe operation of an interconnected system, it is necessary to model the relations established through the connections linking the multiple components of the involved infrastructures. In this work, the modeling of interdependencies among network systems and of their effects on failure propagation is carried out within the simulation framework of a failure cascade process. The sensitivity of the critical loading value (the lower bound of the cascading failure region) and of the average cascade size with respect to the coupling parameters defining the interdependency strength is investigated as a means to arrive at the definition and prescription of cascade-safe operating margins.

The Annual Technical Report this year is focused on infrastructure reliability. Infrastructure constitutes those things that are apparent only in their absence. We take the infrastructure for granted, assuming it will always be there. We turn on our water facet, and drinkable water has always flowed out, for most of us, most of the time. Our infrastructure is subject to environment breakages (e.g., earthquakes), accidents (e.g., dig ups of cables), sabotage, intrusion, and compromise. Also everyday component, software or system failures can bring our infrastructure down. Our global connectivity and communications, as well as our world wide distributed development and maintenance systems, increase our productivity and efficiency, but can also increase our vulnerabilities. Our critical infrastructures can be found in many places.

Feature selection for transient classification is the problem of choosing among several monitored parameters (i.e., the features) to be used for efficiently recognizing the developing transient patterns. It is a critical issue for the application of “on condition” diagnostic techniques in complex systems, such as the nuclear power plants, where hundreds of parameters are measured. Indeed, irrelevant and noisy features have been shown to unnecessarily increase the complexity of the classification problem and degrade the diagnostic performance. In this paper, the problem of selecting the features to be used for efficient transient classification is tackled by means of multiobjective genetic algorithms. The approach leads to the identification of a family of equivalently optimal subsets of features, in the Pareto sense. However, difficulties in the convergence of the standard Pareto-based multiobjective genetic algorithm search in large feature spaces may arise in terms of representativeness of the identified Pareto front whose elements may turn out to be unevenly distributed in the objective functions space, thus not providing a full picture of the potential Pareto-optimal solutions. To overcome this problem, a niched Pareto genetic algorithm is embraced in this work. The performance of the feature subsets examined during the search is evaluated in terms of two optimization objectives: the classification accuracy of a Fuzzy K-Nearest Neighbors classifier and the number of features in the subsets. During the genetic search, the algorithm applies a controlled “niching pressure” to spread out the population in the search space so that convergence is shared on different niches of the Pareto front, which is thus evenly covered. The method is tested on a diagnostic problem characterized by a very large number of process features available for the classification of simulated transients in the feedwater system of a boiling water reactor. The dynamics of the transient signals is captured by wavelet decomposition, which actually increases the complexity of the search for the optimal feature subsets by triplicating the number of features to be considered. © 2008 Wiley Periodicals, Inc.

In this work, the reversible-jump Markov chain Monte Carlo technique is applied for identifying the parameters governing stochastic processes of component degradation. Two case studies are examined concerning the evolution of deteriorating systems whose parameters undergo step changes in time. The method turns out to be capable of identifying the instances of change in behavior, and of estimating the parameter values. A Bayesian updating strategy is proposed to refine the parameter estimates as new data are made available.

In practice, network designs can be based on multiple choices of redundant configurations, and different available components which can be used to form links. More specifically, the reliability of a network system can be improved through redundancy allocation, or for a fixed network topology, by selection of highly reliable links between node pairs, yet with limited overall budgets, and other constraints as well. The choice of a preferred network system design requires the estimation of its reliability. However, the uncertainty associated with such estimates must also be considered in the decision process. Indeed, network system reliability is generally estimated from estimates of the reliability of lower-level components (nodes & links) affected by uncertainties. The propagation of the estimation uncertainty from the components degrades the accuracy of the system reliability estimation. This paper formulates a multiple-objective optimization approach aimed at maximizing the network reliability estimate, and minimizing its associated variance when component types, with uncertain reliability, and redundancy levels are the decision variables. In the proposed approach, Genetic Algorithms (GA) and Monte Carlo (MC) simulation are effectively combined to identify optimal network designs with respect to the stated objectives. A set of Pareto optimal solutions are obtained so that the decision-makers have the flexibility to choose the compromised solution which best satisfies their risk profiles. Sample networks are solved in the paper using the proposed approach. The results indicate that significantly different designs are obtained when the formulation incorporates estimation uncertainty into the optimal design problem objectives.

The determination of optimal Surveillance Test Intervals (STI) is a matter of great importance in risk-informed applications, due both to the implications that maintenance actions have on the safety of the risky plants such as the nuclear ones, and to the important amount of resources invested in maintenance operations by the industrial organizations. The common approach to determining the optimal STI uses a simplified system-availability model relying on a set of parameters at the component level (failure rate, repair rate, frequency of failure on demand, human error rate, inspection duration, etc.), whose values are typically estimated on the basis of few, sparse data, and can suffer from appreciable uncertainties. Thus, the prediction of the system behavior on the basis of the parameters' best estimates is scarcely significant, if not accompanied by some measure of the associated uncertainty, such as the variance. This paper proposes a multi-objective optimization approach, based on genetic algorithms, which transparently and explicitly account for the uncertainties in the parameters. The objectives considered (the inverse of the s-expected system failure probability and the inverse of its variance), are such as to drive the genetic search toward solutions which are guaranteed to give optimal performance with high assurance. For validation purposes, a simple case study regarding the optimization of the layout of a pipeline is firstly presented. The procedure is then applied to a more complex system taken from literature, the Residual Heat Removal safety system of a Boiling Water Reactor, for determining the optimal STI of the system components. The approach provides the decision maker with a useful tool for determining those solutions which, besides being optimal with respect to the s-expected safety behavior, allow a high degree of assurance in the actual system performance.