Anne Barros

A periodic inspection and replacement policy for systems subject to competing failure modes due to degradation and traumatic events

Abstract This paper deals with the condition-based maintenance of single-unit systems which are subject to the competing and dependent failures due deterioration and traumatic shock events. The main aim is to provide a model to assess the value of condition monitoring information for the maintenance decision-making. A condition-based periodic inspection/replacement policy is developed and compared with a benchmark time-based block replacement policy. Numerical results show that it is indeed useful to follow closely the actual evolution of the system to adapt the maintenance decisions to the true system state to improve the performance of maintenance policies. The analysis of the maintenance costs savings can be used to justify or not the choice to implement a policy based on condition monitoring information and to invest in condition monitoring devices.

Remaining useful life estimation based on stochastic deterioration models: A comparative study

Prognostic of system lifetime is a basic requirement for condition-based maintenance in many application domains where safety, reliability, and availability are considered of first importance. This paper presents a probabilistic method for prognostic applied to the 2008 PHM Conference Challenge data. A stochastic process (Wiener process) combined with a data analysis method (Principal Component Analysis) is proposed to model the deterioration of the components and to estimate the RUL on a case study. The advantages of our probabilistic approach are pointed out and a comparison with existing results on the same data is made.

Dynamic grouping maintenance with time limited opportunities

This paper presents firstly a dynamic grouping maintenance strategy for multi-component systems with positive economic dependence, which implies that combining maintenance activities is cheaper than performing maintenance on components separately. Preventive maintenance durations and multiple occurrences of maintenance activities within scheduling horizon are considered. Moreover, in a dynamic context, maintenance opportunities, defined as inactivity periods of the systems at which several maintenance activities could be executed with reduced maintenance costs, may randomly occur with time. The second objective of the paper is to propose a new algorithm to optimally update online the grouped maintenance planning by taking into account the maintenance opportunities. A numerical example of a five components system is finally introduced to illustrate the proposed dynamic grouping maintenance strategy.

Reliability importance analysis of Markovian systems at steady state using perturbation analysis

Sensitivity analysis has been primarily defined for static systems, i.e. systems described by combinatorial reliability models (fault or event trees). Several structural and probabilistic measures have been proposed to assess the components importance. For dynamic systems including inter-component and functional dependencies (cold spare, shared load, shared resources, etc.), and described by Markov models or, more generally, by discrete events dynamic systems models, the problem of sensitivity analysis remains widely open. In this paper, the perturbation method is used to estimate an importance factor, called multi-directional sensitivity measure, in the framework of Markovian systems. Some numerical examples are introduced to show why this method offers a promising tool for steady-state sensitivity analysis of Markov processes in reliability studies.

From differential to difference importance measures for Markov reliability models

This paper presents the development of the differential importance measures (DIM), proposed recently for the use in risk-informed decision-making, in the context of Markov reliability models. The proposed DIM are essentially based on directional derivatives. They can be used to quantify the relative contribution of a component (or a group of components, a state or a group of states) of the system on the total variation of system performance provoked by the changes in system parameters values. The estimation of DIM at steady state using only a single sample path of a Markov process is also investigated. A numerical example of a dynamic system is finally introduced to illustrate the use of DIM, as well as the advantages of proposed evaluation approaches.

Age-based preventive maintenance for passive components submitted to stress corrosion cracking

This paper deals with an age-based preventive maintenance for critical systems or structures subject to a gradual degradation phenomenon such as stress corrosion cracking. We analyze a system subjected to different cracks. A crack can be only detected when its length exceeds a detection threshold. When the length of the crack reaches a given threshold, the system fails. The length of the crack is modeled using a gamma process. Furthermore, when the number of cracks detected in the system attains a fixed value, the system fails. Corrective maintenance actions are performed after a system failure. A preventive maintenance is performed when the age of the system is T. Maintenance actions replace the system by a new one with an associated cost. The problem is to determine an optimal planned replacement time T, minimizing the expected cost rate of the system. The analytical solution to the problem is obtained under some general assumptions. A numerical example is shown to illustrate the problem.

On the Use of Mean Residual Life as a Condition Index for Condition-Based Maintenance Decision-Making

This paper provides a methodology to analyze the efficiency of mean residual life in condition-based maintenance decision-making. A degradation-threshold-dependent-shock model is used to describe the evolution of a system subject to the dependent and competing failure modes due to degradation and shock. Based on this model, we compute the mean residual life of system and analyze its monotonicity. This property of mean residual life function allows introducing a new condition-based maintenance strategy whose preventive maintenance decision is based on the mean residual life. The proposed strategy is then compared to a maintenance strategy based on the degradation level only. Analyzing the equivalence, the performance and the flexibility of both strategies allow us to give some conclusions on the interest of the mean residual life as a condition index for maintenance decision-making.

Reliability analysis for new technology-based transmitters

The reliability analysis of new technology-based transmitters has to deal with specific issues: various interactions between both material elements and functions, undefined behaviours under faulty conditions, several transmitted data, and little reliability feedback. To handle these particularities, a “3-step” model is proposed, based on goal tree–success tree (GTST) approaches to represent both the functional and material aspects, and includes the faults and failures as a third part for supporting reliability analyses. The behavioural aspects are provided by relationship matrices, also denoted master logic diagrams (MLD), with stochastic values which represent direct relationships between system elements. Relationship analyses are then proposed to assess the effect of any fault or failure on any material element or function. Taking these relationships into account, the probabilities of malfunction and failure modes are evaluated according to time. Furthermore, uncertainty analyses tend to show that even if the input data and system behaviour are not well known, these previous results can be obtained in a relatively precise way. An illustration is provided by a case study on an infrared gas transmitter. These properties make the proposed model and corresponding reliability analyses especially suitable for intelligent transmitters (or “smart sensors”).

Adaptive condition-based maintenance decision framework for deteriorating systems operating under variable environment and uncertain condition monitoring

The present article deals with the efficient use of different types of monitoring information in optimizing condition-based maintenance decision making for a deteriorating system operating under variable environment. The degradation phenomenon of a system is the fatigue crack growth that is modeled by a physics-based stochastic process. The environment process is assumed to be modeled by a time-homogenous Markov chain with finite state space. We suppose that the environmental condition is observed perfectly, while the crack depth can be assessed imperfectly through a non-destructive ultrasonic technique. As such, two kinds of indirect information are available on the system at each inspection time: environmental covariate and diagnostic covariate. Based on this set of information, two condition-based maintenance strategies adaptive to environmental conditions are developed. In the first one, the adaptation scheme is time-based, while in the second, it is condition-based. These maintenance strategies are compared one with another and to a classical non-adaptive one to point out the performances of each adaptation scheme and hence the appreciation of using different information sources in maintenance decision making.

Remaining useful life estimation on the non-homogenous gamma with noise deterioration based on Gibbs filtering: A case study

Prognostic of system lifetime is a basic requirement for condition-based maintenance in many application domains where safety, reliability, and availability are considered of first importance. Assessment of residual lifetime of component is always taken as one of important tasks of prognostic. In the framework of prognostic, the non-probabilistic approaches are mostly considered because of their connection to the scientific community that first developed the research area corresponding to the prognostic problem and started it from a very operational point of view. However, more and more probabilistic approaches such as hidden Markov model, life cycle data analysis, proportional hazards models, etc. have been applied to prognostic. In this paper, a probabilistic approach is considered where a lifetime distribution or a stochastic process is associated to the sys tem or component under consideration. This study considers the simulated noisy observations set corresponding to a Gamma process with additive Gaussian noise which is associated to the deterioration phenomenon. The presence of the Gaussian noise is due to the noisy and irregularly sampled observations data. In order to propose a remaining useful lifetime estimation, first by a stochastic filtering with Gibbs sampler the hidden degradation state is estimated. Since this latter evolves according to a gamma process, based on the gamma process properties the remaining useful life distribution is calculated. The interest of our probabilistic method is pointed out.