基于BP神经网络的引信贮存可靠性预计

根据引信贮存可靠性的特点和BP神经网络结构．建立了引信贮存可靠性预计的神经网络模型;结合库存引信可靠性的实测数据,应用BP神经网络的误差反向传播算法．对引信贮存可靠度进行了训练并预计出引信贮存可靠度下限值,并与极大似然估计的可靠度下限值进行了比较,结果相吻合．神经网络在引信贮存可靠性预计中的应用．对处理目前库存引信的决策具有重要意义．     

Reliability and risk models based on independent trials

An analysis has been made of the mathematical relationship between two alternative models for reliability and risk estimation under the assumption of mutual independence. In cases where the reliability formulation is expressible as a compound union event, the resultant reliability expressions are analogous to the Bernoulli and Poisson trials processes. Nonparametric inequality relationships aredeveloped that demonstrate that a Bayesian-Bernoulli model always predicts event probabilities that are less than Bernoulli probabilities, which are always less than or equal to probabilities predicted by the finer grained Poisson trials model. An analysis of the maximum relative prediction error indicates when the individual probabilities are less than 0.1, the relative error between the Bernoulli and Poisson models is always less than 5 percent. The results are demonstrated to have utility in system reliability, engineered design lifetime risk analysis, and simulation applications in which the model is based on independent trials.     

基于ADT的电阻型湿度传感器可靠性评估

传统的寿命试验对电子元件可靠性进行评估需较长时间。如何快速、准确获取电子元件的性能指标是工程实践和试验研究迫切需要解决的问题。本文以电阻型湿度传感器为研究对象,基于加速退化试验(ADT)的可靠性评估方法,建立加速模型和退化轨迹,结合最大似然估计和最小二乘法求解加速应力下伪失效寿命分布参数,从而得出正常应力水平下可靠度函数。结果表明,采用ADT能够准确获取湿度传感器的可靠性信息,缩短试验周期,此评估方法同样适用于其它电子元件的可靠性研究,存在广泛的应用价值。     

A generic method for estimating system reliability using Bayesian networks

This study presents a holistic method for constructing a Bayesian network (BN) model for estimating system reliability. BN is a probabilistic approach that is used to model and predict the behavior of a system based on observed stochastic events. The BN model is a directed acyclic graph (DAG) where the nodes represent system components and arcs represent relationships among them. Although recent studies on using BN for estimating system reliability have been proposed, they are based on the assumption that a pre-built BN has been designed to represent the system. In these studies, the task of building the BN is typically left to a group of specialists who are BN and domain experts. The BN experts should learn about the domain before building the BN, which is generally very time consuming and may lead to incorrect deductions. As there are no existing studies to eliminate the need for a human expert in the process of system reliability estimation, this paper introduces a method that uses historical data about the system to be modeled as a BN and provides efficient techniques for automated construction of the BN model, and hence estimation of the system reliability. In this respect K2, a data mining algorithm, is used for finding associations between system components, and thus building the BN model. This algorithm uses a heuristic to provide efficient and accurate results while searching for associations. Moreover, no human intervention is necessary during the process of BN construction and reliability estimation. The paper provides a step-by-step illustration of the method and evaluation of the approach with literature case examples.     

Error bounds for the reliability index in finite element reliability analysis

This work presents an extension of the goal‐oriented error estimation techniques to the reliability analysis of a linear elastic structure. We use a first‐order reliability method in conjunction with a finite element analysis (FEA) to compute the failure probability of the structure. In such a situation the output of interest that is computed from the FEA is the reliability index β. The accuracy of this output, and thus of the reliability analysis, depends, in particular, on the accuracy of the FEA. In this paper, upper and lower bounds of the reliability index are proposed, as well as simple bounds of the failure probability. An application to linear fracture mechanics is presented. Copyright © 2011 John Wiley & Sons, Ltd.     

Reliability estimation based on operational data of manufacturing systems

Maintenance management has a direct influence on equipment reliability and safety. However, a large portion of traditional maintenance models and reliability analysis methods usually assumes that only perfect maintenance is performed on the system and the system will restore to as good as new regardless of the kind of preventive maintenance work‐order that is performed. This is not practical in reality and may result in an inaccurate parametric estimation. The research objective of this paper is to develop a maximum likelihood estimation method to obtain more accurately estimated parameters based on the operational data of manufacturing systems, taking into consideration the difference between perfect and imperfect maintenance work‐orders. Weibull distribution is specifically studied for this purpose. A practical case study based on industrial operational data from an automotive assembly line is performed to illustrate the implementation and efficiency of the proposed reliability estimation method. Copyright © 2008 John Wiley & Sons, Ltd.     

Application of power law model in reliability evaluation of machine tools by considering working condition difference

The influence of working condition difference has not been considered in the traditional reliability modeling of numerical control (NC) machine tools. To solve the problem, a reliability evaluation method based on mixture variable parameter power law model (MVPPLM) is proposed in this study. First, the scale parameter of the PLM is obtained by multi‐dimensional exponential distribution. Second, a proportional relation of failure rate function between each working condition and reference working condition is established. The proportion coefficient is solved using the partial likelihood function. Working condition factors with a significant influence on reliability levels are selected through the chi‐squared test. Third, reliability evaluation models under different working condition levels are established through mixture distributions. The mixture weight coefficient is calculated by the standard deviation of working condition covariates. The maximum likelihood estimation method is used to estimate parameters. Finally, results of a case analysis based on the data of NC machine tools in the user field tracing test show that the MVPPLM has higher precision than the traditional method. Therefore, reliability evaluation that considers working condition difference is valuable for engineering application.     

Bayesian reliability analysis of a products of probabilities model for parallel systems with dependent components

In a Bayesian reliability analysis of a system with dependent components, an aggregate analysis (i.e. system-level analysis) or a simplified disaggregate analysis with independence assumptions may be preferable if the estimations obtained from employing these two approaches do not deviate substantially from those derived through a disaggregate analysis, which is generally considered the most accurate method. This study was conducted to identify the key factors and their range of values that lead to estimation errors of great magnitude. In particular, a copula-based Bayesian reliability model was developed to formulate the dependence structure for a products of probabilities model of a simple parallel system. Monte Carlo simulation, regionalised sensitivity analysis and classification tree learning were employed to investigate the key factors. The resulting classification tree achieved favourable predictive accuracy. Several decision rules suggesting the optimal approach under different combinations of conditions were also extracted. This study has made a methodological contribution in laying the groundwork for investigating systems with dependent components using copula-based Bayesian reliability models. With regard to practical implications, this study also derived useful guidelines for selecting the most appropriate analysis approach under different scenarios with different magnitude of dependence.     

Reliability analysis of wind turbine subassemblies based on the 3-P Weibull model via an ergodic artificial bee colony algorithm

The Weibull distribution is the most widely used model for the reliability evaluation of wind turbine subassemblies. Considering the important role of the location parameter in the three-parameter (3-P) Weibull model and its rare application in wind turbines, this study conducted a reliability analysis of wind turbine subassemblies based on field data that obeyed the 3-P Weibull distribution model via maximum likelihood estimation (MLE). An improved ergodic artificial bee colony algorithm (ErgoABC) was proposed by introducing the chaos search theory, global best solution, and Lévy flights strategy into the classical artificial bee colony (ABC) algorithm to determine the maximum likelihood estimates of the Weibull distribution parameters. This was validated against simulation calculations and proved to be efficient for high-dimensional function optimization and parameter estimation of the 3-P Weibull distribution. Finally, reliability analyses of the wind turbine subassemblies based on different types of field failure data were conducted using ErgoABC. The results show that the 3-P Weibull model can reasonably evaluate the lifetime distribution of critical wind turbine subassemblies, such as generator slip rings and main shafts, on which the location parameter has a significant effect.     

Integration of computation and testing for reliability estimation

This paper develops a methodology to integrate reliability testing and computational reliability analysis for product development. The presence of information uncertainty such as statistical uncertainty and modeling error is incorporated. The integration of testing and computation leads to a more cost-efficient estimation of failure probability and life distribution than the tests-only approach currently followed by the industry. A Bayesian procedure is proposed to quantify the modeling uncertainty using random parameters, including the uncertainty in mechanical and statistical model selection and the uncertainty in distribution parameters. An adaptive method is developed to determine the number of tests needed to achieve a desired confidence level in the reliability estimates, by combining prior computational prediction and test data. Two kinds of tests — failure probability estimation and life estimation — are considered. The prior distribution and confidence interval of failure probability in both cases are estimated using computational reliability methods, and are updated using the results of tests performed during the product development phase.     

基于改进统计矩点估计法和最大熵原理的结构整体可靠度分析

结构整体可靠度评估一直以来是结构可靠度领域研究的热点与难点。该文将结构整体可靠度分类,并给出其对应功能函数的统一描述;结合提出的有效维度两步分析法和共轭无迹变换法,发展了改进统计矩点估计法;结合最大熵原理和改进统计矩点估计法,提出了适用于两类结构整体可靠度的统一分析方法;通过2个数值算例对该文方法进行了验证。算例分析结果表明:同一精度水平下,该文方法的计算效率较传统的三变量降维近似统计矩点估计法高2.3倍~2.6倍;该文方法具有高的精度水平,其最大相对误差低于2%,适用于结构整体可靠度评估。     

Asymptotic sampling for high-dimensional reliability analysis

Computational procedures for reliability analysis in many cases suffer from substantially increased effort with increasing dimensionality. This means that methods which are well-suited for cases with a small or moderately large number of random variables may not be tractable for situations involving a large number of random variables. Such situations typically occur when random processes or random fields are discretized in terms of spectral representations. The present paper introduces a novel asymptotic sampling strategy which allows a reasonably accurate estimation of the generalized reliability index using a small number of random or quasi-random samples. This strategy utilizes well-established asymptotic results from reliability theory together with a simple regression technique. Several numerical examples demonstrate the applicability, versatility, and accuracy of the approach.     

Automotive reliability inference based on past data and technical knowledge

The constantly increasing market requirements of high quality vehicles ask for the automotive manufacturers to carry out—before starting mass production—reliability demonstration tests on new products. However, due to cost and time limitation, a small number of copies of the new product are available for testing, so that, when the classical approach is used, a very low level of confidence in reliability estimation results in. In this paper, a Bayes procedure is proposed for making inference on the reliability of a new upgraded version of a mechanical component, by using both failure data relative to a previous version of the component and prior information on the effectiveness of design modifications introduced in the new version. The proposed procedure is then applied to a case study and its feasibility in supporting reliability estimation is illustrated.     

Reliability modeling for competing failure processes considering degradation rate variation under cumulative shock

Most systems experience both random shocks (hard failure) and performance degradation (soft failure) during service span, and the dependence of the two competing failure processes has become a key issue. In this study, a novel dependent competing failure processes (DCFPs) model with a varying degradation rate is proposed. The comprehensive impact of random shocks, especially the effect of cumulative shock, is reasonably considered. Specifically, a shock will cause an abrupt degradation damage, and when the cumulative shock reaches a predefined threshold, the degradation rate will change. An analytical reliability solution is derived under the concept of first hitting time (FHT). Besides, a one-step maximum likelihood estimation method is established by constructing a comprehensive likelihood function. Finally, the reasonability of the closed form reliability solution and the feasibility and effectiveness of the proposed DCFPs modeling methodology are demonstrated by a comparative simulation study.     

Sensibility of Bayesian inference methods for reliability prediction of ageing systems,case of Diesel locomotives

This paper addresses the problem of reliability analysis of in-service identical systems when a limited number of lifetime data is available compared to censored ones. Lifetime (resp. censored) data characterise the life of failed (resp. non-failed) systems in the sample. Because, censored data induce biassed estimators of reliability model parameters, a methodology approach is proposed to overcome this inconvenience and improve the accuracy of the parameter estimation based on Bayesian inference methods. These methods combine, in an effective way, system’s life data and expert opinions learned from failure diagnosis of similar systems. Three Bayesian inference methods are considered: Classical Bayesian, Extended Bayesian and Bayesian Restoration Maximisation methods. Given a sample of lifetime data, simulated according to prior opinions of maintenance expert, a sensibility analysis of each Bayesian method is performed. Reliability analysis of critical subsystems of Diesel locomotives is established under the proposed methodology approach. The relevance of each Bayesian inference methods with respect to collected reliability data of critical subsystems and expert opinions is discussed.     

A model of storage reliability with possible initial failures

Recently, storage reliability has attracted attention because of the increasing demand for high reliability of products in storage in both military and commercial industries. In this paper we study a general storage reliability model for the analysis of storage failure data. It is indicated that the initial failures, which are usually neglected, should be incorporated in the estimation of storage failure probability. Data from the reliability testing before and during the storage should be combined to give more accurate estimates of both initial failure probability and the probability of storage failures. The results are also useful for decision-making concerning the amount of testing to be carried out before storage. A numerical example is also given to illustrate the idea.     

Binary State Reliability Computation for a Complex System Based on Extended Bernoulli Trials: Multiple Autonomous Robots

Availability of a system is a crucial factor for planning and optimization. The concept is more challenging for modern systems such as robots and autonomous systems consisting of a complex configuration of components. In this paper, a reliability evaluation framework is developed for a system of binary state autonomous robots in an automated manufacturing environment. In this framework, the concepts in functional block diagram, table of truth, and sum of state are employed simultaneously to develop a binary state reliability model. Due to inefficacy of the method for larger number of components involved in complex systems, an extension of the Bernoulli trials is proposed. In an implementation study, the effectiveness and computational efficiency of the proposed method are illustrated. In addition, an analysis on the failure rate using the maximum likelihood estimation and confidence interval is reported. Copyright © 2017 John Wiley & Sons, Ltd.     

双参数Rayleigh分布环境因子的经验Bayes估计及应用

本文针对Rayleigh分布位置参数已知的情形,给出了Rayleigh分布环境因子的极大似然估计和经验Bayes估计,并将环境因子的估计结果应用于Rayleigh部件的可靠性评估,给出了该部件可靠度函数与失效率的估计。最后的随机模拟例子表明,经验Bayes估计优于极大似然估计,并且在考虑环境因子的情形下,Rayleigh部件可靠性指标的估计优于未考虑环境因子时的估计。     

Method for calculating the lifetime of unit metal packages

The technical characteristics of packages, determined by corresponding standards and technical requirements, are verified by package reliability. The reliability of a package is defined as an ability to protect the quality of the product in a package to a suitable level, in a definite condition, under the influence of various factors and for a definite time interval. The best index of the reliability of food packages is the maximal time from packing to the consumption of a product when the conditions of preservation and storage as well as the preparation of food for consumption are taken into account. Any change in the parameters characterizing a product/package/environment system can affect this time; thus, results of the investigation on randomly chosen samples and the analysis of these results can be taken as a basis for the estimation of the technical characteristics of packages. Copyright © 1999 John Wiley & Sons, Ltd.     

Reliability assessment of complex electromechanical systems: A network perspective

The network theory is widely applied to improve the reliability of a complex electromechanical system. In this application, system reliability assessment with network theory has been paid a great deal of attention. Because of instrument malfunctions, staff omissions, imperfect inspection strategies, and complex structures, field failure data are often subject to interval censoring, making the holistic reliability assessment becomes a difficult task. Most traditional methods assume reliability of critical components or partial reliability as system reliability, which may cause a large bias in system reliability estimation. This paper proposes a novel method to evaluate and predict the system reliability of a complex electromechanical system subject to the insufficient fault data problem from a network perspective. First, the system modeling based on network theory is developed to describe the topology of a holistic system. Second, interval‐valued intuitionistic hesitant fuzzy number is proposed in order to solve insufficient data for single component. Then, a new measure—comprehensive reliability—that can reflect the reliability of nodes in combination with functional properties and topological properties, which are formulated by failure data and network model, respectively, is constructed for system reliability assessment. Subsequently, an improved system reliability model based on percolation theory is given in terms of comprehensive reliability of nodes. Finally, to verify the effectiveness of the proposed method, a simulation and a real case study for traction system are implemented.