This study presents a tri-state modeling-based FTA-BN hybrid diagnostic framework for assessing the reliability of diesel powertrains in heavy-duty railway maintenance machinery. Recognizing the critical role of these power units in project timeline control and construction efficiency, the framework addresses the challenge of multi-phase fault evolution under harsh operating conditions. A tri-state model—comprising fully operational, performance degradation, and functional failure states—is introduced to enable dynamic, quantitative evaluation of system degradation, overcoming the limitations of traditional binary-state models in characterizing performance degradation processes. The methodology employs a modular failure decomposition strategy and constructs a multi-level fault tree model based on the functional topology of the Deutz BF12L513C powertrain. By integrating a Dynamic Bayesian Network (DBN), the framework achieves three objectives: (1) probabilistic fusion of long-term operational data, (2) quantification of expert-based conditional probabilities using the Delphi method, and (3) uncertainty propagation among coupled failure modes. Analysis of 3 years of field data yields a system reliability of 0.9330, with a 6.25% probability of performance degradation and a 0.44% probability of functional failure. Fault path analysis identifies hydraulic circuit integrity (node X67) and the exhaust energy recovery subsystem (node X73) as key reliability bottlenecks. This framework offers a scalable approach for preventive maintenance and life-cycle reliability management in complex engineering systems.

Survival probabilities for systems with protection operating under renewal shock process are derived. This is done using a method of simultaneous integral equations specifically developed for this setting. The solutions are given in terms of the Laplace Transforms that can be inverted numerically. The protection system’s failure under a shock means that it was not neutralized and have reached the main system. On the other hand, the neutralized shock does not affect the main system. Various specific cases are discussed as examples and ‘fast repair’ approximations are provided. Specifically, when the probabilities of failures of the protection and main systems on a single shock are sufficiently small, a simple, speaking for itself fast repair asymptotic result describes the ‘double thinning’ of the original process of shocks.

Ensemble learning methods, particularly stacking, are often expected to enhance the performance of machine learning models. In this study, an investigation was carried out on whether stacking consistently improves classification accuracy in the context of fault diagnosis. Vibration signals collected from a reciprocating air compressor wherein three distinct features such as statistical, histogram and autoregressive moving average (ARMA) features were extracted. The most significant features were selected using the J48 algorithm and a variety of machine learning classifiers were trained on these features. The performances of individual classifiers were recorded and compared against stacking ensembles built from the same models. The results show that while several individual models achieved high classification performance, stacking did not provide consistent improvements. These findings highlight that stacking was ineffective on the considered air compressor dataset and is not always advantageous in fault diagnosis.

In today’s competitive industrial environment, maximizing the Operational Availability (OA) of complex systems while minimizing their Life Cycle Cost (LCC) is a key challenge. This paper proposes an innovative methodology that jointly integrates design and maintenance strategies to enhance the performance of complex, multi-component, and repairable systems. The proposed approach integrates several parameters: reliability, redundancy, maintainability, logistics, and diagnostics. It is based on a dynamic planning of maintenance operations, guided by the concept of Maintenance-Free Operating Period (MFOP). The proposed approach accounts for multiple interrelated parameters, including reliability, redundancy, maintainability, logistics, and diagnostics. It is structured around a dynamic maintenance planning framework, guided by the Maintenance-Free Operating Period (MFOP) concept to ensure interruption-free operation over a specified duration with a defined confidence level. To achieve these objectives, a hybrid computational approach is developed, combining Monte Carlo simulation (MCS), discrete-event simulation (DES), and the multi-objective evolutionary algorithm (NSGA-II). MCS generates stochastic failure and repair durations, while DES models the system’s dynamic behavior to assess operational availability and associated costs. The methodology generates a set of Pareto-optimal solutions, enabling decision-makers to evaluate balanced trade-offs between OA and LCC, optimized via NSGA-II. The results demonstrate the efficacy of joint design-maintenance optimization in addressing the challenges of complex systems.

Human behavior has become a significant concern in almost every economic activity. Human errors are on the top accident list, leading to death tolls, material losses, and environmental damages. The lack of large and good-quality datasets for Human Reliability Assessment (HRA) studies is still a problem in applications of safety sciences, and the present work proposes an approach to collecting HRA data from simulator utilizing Game Engines (GE) 3D-Virtual Environments (VE). As validation, an experiment was conducted for an Oil and Gas refinery evacuation scenario under toxic cloud release, with a detailed description of environmental development. Two variables were analyzed: evacuation time and individual risk exposure. Then, a Bayesian Belief Network (BBN) was created to investigate the tool for HRA, considering variables related to training, visibility, and complexity. The study provides valuable insights into human behavior and the generation of datasets, representing a helpful tool for data collection.

Predictive maintenance for machines typically involves model-based condition monitoring systems. However, many state-of-the-art methods are not designed to work under varying production plans or machine operating conditions. Contextual data, such as machine parameters, input material details, and environmental factors, can provide valuable insights into the effectiveness of existing models in changing environments. However, research in this area is limited. This article takes a broad view on adaptive ensemble learning for machine health prognosis from context information-based fitness estimation. Prognostic models are leveraged to increase accuracy under varying operating conditions by using spatial proximity computations between new contextual data features and data from their training phases. Several experiments aim to answer open research questions from previous work. In particular, this study extends the base algorithm from a single one to a set of five common regression algorithms and two classification algorithms and the demonstration domain from turbofans to a simplified synthetic and a milling scenario. In addition, the effect of the available time horizon for online prediction of the context information is evaluated. The results show that the adaptive ensemble method consistently improves degradation percentage estimation accuracy over both single models and conventional averaging ensembles. For classification tasks with highly imbalanced data, such as the milling scenario, the method offers marginal gains, indicating limited benefit from cross-validation weighting in such cases.

This article explores reliability assessment from a systemic and hierarchical perspective. It highlights that while failures occur at the component level, they are often driven by stresses resulting from system use and environmental conditions. To address this, the U-Cycle approach is introduced, providing a structured method to link system decomposition with potential failure mechanisms. The proposed methodology, supported by five complementary views, captures both the organization of the system and its behaviour under operational stress. It integrates a top-down functional analysis, which translates mission profiles into stress variables acting on components, with a bottom-up analysis that examines how failures propagate through subsystems to affect the system-level functionality. Together, these views provide a comprehensive framework for assessing system reliability. A case study on the electrical flight-control function of an aircraft demonstrates the practical implementation of the U-Cycle approach. It shows how stress propagation and degradation mechanisms influence the function’s availability, thereby illustrating the method’s applicability for both predictive reliability modelling during design and for diagnostic reasoning during operation.

We present a model for the corrective and preventive maintenance of a system. The latter is based on a bivariate policy that replaces the system either at age T or after the M failure, whichever comes first. A repair follows each of the M − 1 first failures, restoring the system operational state, but with a lower reliability than before failing. We present two scenarios with constant and time-dependent repair costs. The results reveal that systems with low initial reliability can greatly benefit from the bivariate policy. The advantage decreases for poor quality repairs. We also obtain conditions to obtain the optimum number M when T is given. This result is helpful to assess whether a system should be replaced sooner than originally planned.

This paper examines a finite-capacity machine repair system with M parallel operating machines, W warm standby machines, and C cold standby machines. It incorporates retrial and Bernoulli feedback queueing characteristics to investigate reliability and optimization issues. The system is equipped with an unreliable repairman responsible for repairing failed machines. During operation, the repairman may experience working breakdown, and provides repair services at a lower rate during the failure. Firstly, based on Markov process theory, matrix analysis method, and Cramer’s rule, transient and steady-state analyses are conducted on the system. By solving the balance equations in matrix form, the queueing and reliability indicators of the system are obtained. Furthermore, the closed form solution for the system’s transient probabilities is derived using the Laplace transform and the eigenvalue method. Subsequently, numerical experiments and numerical simulations are presented to illustrate the effects of system parameters on performance indicators, followed by a sensitivity analysis of reliability measures. Finally, from the decision-maker’s perspective, a bi-objective optimization model is formulated to maximize steady-state availability and minimize total cost, using the Non-dominated Sorting Genetic Algorithm II (NSGA-II) and Multi-Objective Particle Swarm Optimization (MOPSO) algorithms to seek the Pareto fronts. By using Bootstrap method combined with similarity measurement, the stability of the obtained Pareto solution sets are evaluated to ensure the robustness and feasibility of optimization results. The results indicate that both algorithms can achieve effective optimization outcomes, but NSGA-II outperforms MOPSO in terms of optimization performance and solution set stability, exhibiting superior robustness and reliability, thereby providing a theoretical reference for decision-making in practical maintenance systems.

The proliferation of unmanned aerial vehicle (UAV) swarms presents critical challenges to system-level reliability and safety. Traditional maintenance strategies, designed for single assets, are fundamentally inadequate for the systemic complexities and multifaceted risks of swarm operations. This study addresses this gap by developing a multi-objective optimization framework to derive optimal maintenance policies for heterogeneous UAV swarms. We formulate the problem to simultaneously minimize maintenance cost, maximize mission reliability, and minimize a composite operational risk encompassing both crash and hazardous material release. The framework distinguishes between nodal and non-nodal UAV roles and employs the Non-dominated Sorting Genetic Algorithm II (NSGA-II) to explore the complex trade-off space. The framework is validated through numerical experiments against a heuristic benchmark, yielding significant results. Optimized policies reduce injury risk by more than tenfold compared to traditional methods, while simultaneously doubling mission reliability. Furthermore, the analysis reveals a distinct “efficiency frontier” for safety investment, providing a novel, data-driven tool for managerial decision-making. Ultimately, this research delivers a holistic, risk-informed framework that bridges the gap between theoretical optimization and the practical challenges of safe, reliable, and cost-effective swarm operation, offering actionable guidance for tailoring maintenance strategies to specific risk tolerance levels.