Bayesian methodology for system reliability assessment

Purpose This study aims to assess and enhance the production efficiency and reliability of fertilizer manufacturing plants through the application of Markovian analysis. By integrating stochastic modeling with performance management theory, the research provides a predictive performance measurement framework that links reliability metrics to key business outcomes. The research develops a probabilistic framework to model system state transitions, evaluate system availability and optimize maintenance strategies to reduce downtime and improve overall plant performance. Design/methodology/approach A discrete-time Markov chain model is constructed to represent the operational dynamics of critical units within a fertilizer plant. Historical operational and maintenance data are analyzed to develop transition probability matrices that capture the likelihood of state transitions between operational, idle and failure conditions. Key performance metrics such as steady-state probabilities, mean first passage times and recurrence times are computed to assess long-term system behavior. Statistical hypothesis testing – including paired t-tests, Wilcoxon signed-rank tests and chi-square tests – is employed to validate improvements in reliability and efficiency following the implementation of Markov-based maintenance strategies. Regression analysis is also conducted to examine the relationships between operational parameters (e.g. downtime and failure frequency) and production output. Findings The analysis reveals that the production units remain in the operational (RUN) state approximately 61.01% of the time, compared to 53% prior to optimization. Markovian-based maintenance strategies significantly reduced average weekly downtime from 42 h to 29 h (p < 0.001). Weekly production output increased from an average of 1,250 tons to 1,375 tons (p < 0.001). A chi-square test confirmed statistically significant changes in system state transitions (p < 0.001), favoring increased operational continuity. Confidence intervals constructed for key reliability parameters further strengthened the robustness of the findings. Practical implications This study provides a data-driven methodology for improving maintenance planning and production reliability in fertilizer plants. By modeling system behavior through Markovian analysis and applying statistical validation techniques, maintenance managers can develop predictive strategies that reduce unplanned downtime and enhance production efficiency. The methodology is adaptable to other continuous process industries where uptime and reliability are operational priorities and can be integrated into existing performance management systems to support data-driven decision-making and strategic alignment of maintenance activities with productivity goals. Originality/value This study offers a novel academic contribution by applying discrete-time Markov chain modeling to fertilizer manufacturing using empirical operational data. It advances performance management research by integrating stochastic modeling with statistical validation to quantify production efficiency and system reliability. The linkage between probabilistic reliability metrics (e.g. steady-state probabilities and mean first passage times) and business key performance indicators (e.g. downtime and weekly output) provides a new data-driven framework for industrial performance evaluation. This work bridges theoretical modeling with applied maintenance strategies, offering a transferable methodology relevant to researchers and practitioners seeking to optimize reliability and productivity in continuous-process industries.

As the most attractive realization of component-based software, composite Web service asks a heavy demand for reliability prediction in the early stage for service discovery and service composition. Although it is believed that input should be one of the most important aspects affecting the reliability of a component-based system, research of both giving more details on an operational profile and considering software architecture to get more accurate reliability prediction is scant. This paper describes an algebra-based reliability prediction approach in three aspects: (i) a sub-set of OWL-S that provides reliability estimation information is discussed; (ii) a framework that describes syntax and reliability prediction semantics of the sub-set is explained and implemented in Maude; (iii) an example of how to use the framework to calculate reliability is studied.

Polymer composite and metallic materials have found wide application in various industries such as aviation, rocket, car manufacturing, ship manufacturing, etc. Many design elements need permanent quality control. Ensuring high quality and reliability of products is impossible without effective nondestructive testing methods. One of these methods is penetrant testing using penetrating substances based on liquid penetration into defect cavities. In this paper, we propose a model of liquid flow to determine the rates of filling the defect cavities with various materials and, based on this, to choose optimal control modes.

Due to their unique zero carbon emission feature, electric vehicles (EVs) have garnered increasing interest in recent years. However, concerns persist regarding their reliability, particularly concerning critical components. Efforts have been made to assess the reliability of drive motors in EVs. Yet, given the integration of drive motors and motor controllers as a single system in EVs, a combined assessment is essential for a more accurate reliability prediction. Additionally, considering the multiple components within both drive motors and motor controllers, their structure, type, and characteristics may influence the overall reliability of the motor system, a factor overlooked in previous research. To bridge this knowledge gap, this study investigates the reliability of the entire motor system in pure electric vans, encompassing both drive motors and motor controllers. The research begins with predicting the theoretical failure rates of subassemblies and components within the drive motor and motor controller. Subsequently, based on these predictions, the reliability of the entire motor system is assessed. The findings highlight vulnerable subassemblies and components within the motor system, shedding light on potential shortcomings in existing reliability research. These new insights are deemed crucial for informing future reliability design and maintenance practices for pure electric vans. This study undertakes an exhaustive investigation into the reliability of the motor system in EVs. As EVs gain prominence for their environmental benefits and increasing market share, concerns regarding the reliability of critical components, such as the drive motor and associated control systems, become paramount. While prior research has addressed component-level reliability, this study takes a holistic approach by assessing the entire motor system. Through theoretical modeling and empirical analysis, the study predicts failure rates of subassemblies and components within the drive motor and control systems, culminating in an evaluation of the overall motor system reliability. By uncovering vulnerabilities and potential areas for improvement, this research offers valuable insights to advance the design, development, and maintenance of reliable electric vehicles.

Predictive models are used in an attempt to anticipate future transitions, mitigate losses, and maximize economic gains. In today’s market, companies look for high reliability and quality of products due to great market competition. Hence warranty data is of considerable interest to companies. Warranty shows the ability of a system or component to perform its functions within a given customer usage. Many statistical and data mining methods are available to predict the warranty data. This study focuses on analyzing the predictive efficiency of the T-method and Cobb-Douglas production function on warranty data by comparing their prediction capability. The T-method, developed by Genichi Taguchi, is founded upon the fundamentals of the Taguchi System of Quality Engineering which is used to calculate an overall prediction based on signal-to-noise ratio. Using this method, the required parameters are calculated to obtain an overall estimate of the true value of the output for each signal member. The Cobb-Douglas production function is then applied on the same dataset. In economics, the Cobb-Douglas functional form of production function is widely used to represent the relationship of the output to inputs. The strength of the relationship is then assessed using the R-squared and adjusted R-squared values.

To increase energy demand, reliability, and increasing efficiency, thin-film solar cells get the main focus. Various types of solar cell like (Cu(In,Ga)Se2 (CIGS), Cadmium telluride (CdTe) and copper zinc tin sulfide (CZTS) based absorber compound now eliminated by CTS (Cu2SnS3) for different reason like toxicity, shortage and structural complexity. Focus on CTS is increasing because it is nontoxic and it is environmentally friendly. CTS-based solar cell's power conversation efficiency increases and 30% theoretical efficiency indicates it's developing sign. For analyzing the synthesis property, X-ray diffraction (XRD), Raman, Energy Dispersive X-ray (EDX), X-ray fluorescence (XRF) techniques are used. For the synthesis of thin-film solar cells like CTS, different physical methods are used among all sputtering methods. The electron beam evaporation vacuum evaporation method has exposed better efficiency and high product quality and reliability. Additionally, the composition, variation of thickness, structural defects, and elemental composition make the absorber layer's quality affect the performance. This paper has discussed the annealing treatment of CTS-based solar cell synthesis, and it's essential to improve the thin film properties of thin films. This paper describes the reason for efficiency reduction also the scope for future research.

A robust reliability prediction method using Weighted Least Square Support Vector Machine equipped with Chaos Modified Particle Swarm Optimization and Online Correcting Strategy

Traditional reliability assessment methods are generally on the basis of failure lifetime data. The failure lifetime data of high reliability and long lifetime product are obtained difficultly by life test and accelerated life test. Degradation data can provide useful information about the reliability assessment for these products. Aiming at these problems, a method of reliability prediction is presented based on failure data's reliability assessment, accelerated performance degradation and information fusion theory. Finally, this paper gives a case of accelerated performance degradation testing of 24V/2A power supply board to steady voltage to apply the method, whose results demonstrate that this method is correct and valid for reliability assessment of high reliability and long life electronic equipments.

In the majority of works connected with research of reliability of production and other technical systems the Kolmogorov axiomatics is applied. The corresponding analytical apparatus allows to solve problems of maximization of probability of performance of production tasks (for example, daily schedule of steel smelting or monthly production plan). Evaluation of reliability of both production and information systems is a mandatory procedure in their design and is based on the construction of a special structural scheme, called the system model for calculating its reliability. It should be noted that Kolmogorov’s axioms are formulated for a system consisting of two elements. The corresponding graphical explanation of axiom equity is based on an imaginary experiment on a random drop of a certain point on a unit square, inside of which there are two intersectingcircles whose planes are proportional to the probability of reliable operation of the corresponding elements. But there is no mention that the centers of circles should be at a certain distance from each other. In this paper, for the known Kolmogorov axioms underlying the classical probability theory, their analytical conclusion is given.

Purpose: The purpose of this study is to explore end-to-end Linux administration practices, focusing on installation, configuration, performance tuning, and production reliability in heterogeneous computing environments. With the increasing complexity of distributed systems, Linux administrators must manage diverse infrastructures efficiently while ensuring system stability and scalability. Design/methodology/approach: This research adopts a qualitative and analytical approach by reviewing existing literature, system administration frameworks, and real-world case studies. It evaluates Linux system lifecycle stages, including deployment, configuration management, performance optimization, and reliability engineering across heterogeneous platforms such as cloud, edge, and enterprise systems. Findings: The findings reveal that effective Linux administration requires integrated tools, automation, and proactive monitoring. Performance tuning and configuration optimization significantly enhance system efficiency, while reliability frameworks improve fault tolerance. Heterogeneous environments demand adaptive strategies due to varied hardware and software configurations. Practical implications: This study provides system administrators and IT professionals with structured methodologies to improve operational efficiency. It highlights best practices for automation, monitoring, and system tuning to achieve high availability and performance in production environments. Originality/value: This paper contributes a unified perspective on Linux administration by integrating installation, configuration, tuning, and reliability into a single framework tailored for heterogeneous systems, bridging gaps in prior fragmented studies.

Philips has developed a new Ultrasound system, ClearVue, targeted for emerging market countries such as India and China. The product has been developed at different global locations in the world, including North America, China and India. The authors are part of the India team located at Bangalore. The authors, in their role as system architect, system reliability lead, and software architect, have led several activities focused on improving the overall product reliability from software and hardware aspects. This paper is to share their experiences on usage of a variation of FMEA (Failure Mode Effect Analysis) techniques to improve the software reliability of the system. The variation that the team followed was to include criticality analysis of the identified failure modes. This variation, called as FMECA (Failure mode effect and criticality analysis), was specifically chosen since the software under consideration is for a medical device where patient safety and correct diagnosis are of critical importance to the reliable software operation.

Reliability prediction is a critical process of conceptual design for one evolutionary system when it has not been physically built. Due to the limitation of operation data, during the early stage of design, the reliability of a new product is difficult to predict, especially for complex mechanical systems. This article introduces a new method to predict reliability in the early stage of design phase, which uses fuzzy reasoning Petri net to generate the three values, representing complexity, importance and quality of subsystems for the product count reliability prediction. This approach solves the problems of data deficiency and high complexity in the existing methods. The effectiveness and advantages of the proposed method are validated by a case study of a real mechanical system of the spacecraft solar array. The analysis results show that this new method is able to reach an accurate predicted failure rate compared with the reported lifetime of the solar arrays.

Offshore system safety and reliability considering microbial influenced multiple failure modes and their interdependencies

In competitive world markets most manufacturing companies have learned that survival is based on a commitment towards continual process, and that product improvement competes with high quality of product and low price. In today′s rapidly changing marketplace a company must deliver low cost together with high quality and reliability of product to ensure retaining adequate market share. The necessary improvements can be achieved as have already been determined first by a large number of Japanese companies and later by US and European companies with implementation of suitable JIT systems. A number of articles have been written, which attempt to identify and discuss the critical variable essential for successful JIT implementation. To date little have reported on quantitative tangible and intangible benefits of JIT. Discusses tangible and intangible benefits which European companies have derived from JIT.

A brand choice model for TV advertising management using single-source data is proposed. The model replaces household-specific advertising exposure, which is often used as a covariate in a brand choice model, with gross rating points (GRP), a managerial control variable for advertising. In particular, given daily GRP, a probabilistic model of advertising exposure for heterogeneous customers is integrated into a brand choice model with advertising threshold under a Bayesian framework. Through hierarchical modeling, demographic information on panels provides managerial insights into advertising planning.