This paper presents some developed and implemented theory which can be applied to optimization of condition-based maintenance decisions within the context of physical asset management. We examine a replacement problem for a system subject to stochastic deterioration. In particular, the analysis of a preventive replacement policy of the control-limit type for a system subject to inspections at discrete points of time is presented. Cox's Proportional Hazards Model with Weibull baseline hazard function and time dependent stochastic covariates, reflecting the item's condition, is used to describe the hazard rate of the system. Statistical results are then blended with economic or performance considerations to establish long-run optimal maintenance strategies. The structure of the decision-making software EXAKT? is presented, and recently added optimization and prognostic capabilities are described.
Received on January 11, 2007
References: 11

A variety of different management approaches have been devised over time for the repair and eventual replacement of equipment within industry. Common models such as Reliability Centre Maintenance and Operational Research Optimal strategies have existed for sometime. They can be successful in achieving management goal, but often they lack insight to the underlying process which is often non-monotonic. A data driven risk management approach is described in this paper. It provides insight into the performance of the equipment when subject to maintenance, both repair and refurbishment. It can provide economic replacement times based on the cost of continued maintenance. This is useful for single system, but can become more problematic over a series of similar pieces of equipment. The decision to take action in such circumstances will reflect the impact of the action as well as lack of action. A risk management approach based on a 5x5 risk matrix is described. The zones for action can be defined and so the overall strategy defined. The paper will be usefult for the water industry to illustrate application of the approach undertaken.
Received on December 15, 2006
References: 17

Equipments used in production need to be maintained effectively with the objectives of realizing multiple goals related to: availability, maintenance cost and earning rate. But it is difficult to evolve commonly accepted maintenance policies to satisfy all these goals. To this effect, a collaborative system is developed, enhancing the basic model proposed by Kay and Weibull analysis, to arrive at suitable maintenance policy. This system evaluates the preferability of preventive maintenance over breakdown maintenance on collaborative basis by using different measures. In this system, various trade-offs are resolved using numerical techniques and graphical plots. The details of the system with application on four different cases are reported in this paper, along with appropriate results and discussion.
Received on February 06, 2007
References: 09

Accelerated life testing provides timely information about the life distribution of products, components and materials. Accelerated testing is attained by subjecting the products or parts to testing conditions much higher than these products or parts are likely to experience under normal use conditions. The three-parameter Weibull model is commonly used to represent the underlying failure-time distribution for electronic products, in which the minimum lives of these products are different from zero. The standard maximum likelihood method for estimating the parameters of the three parameter Weibull model can have problems since the regularity conditions are not met (see Murthy et al. [1]; Blischke [2]; Zanakis and Kyparisis [3]). In this study, we will develop an accelerated life testing model in which the underlying sampling distribution is the three-parameter Weibull model. We will be assuming a linear acceleration condition. To estimate the shape, scale and minimum life of the three-parameter Weibull model we will use a maximum likelihood approach for censored failure data. To overcome the "no-regularity" problem resulting from the above mentioned, we will apply a modification proposed by Cohen et al. [4]. An example will illustrate the application of the proposed accelerated life testing model.
Received on November 28, 2006
References: 08

Time-limited dispatch (TLD) allows the dispatch of aircraft with faults present in their control systems for limited time periods. In order for TLD to be applied to an aircraft system it is first necessary to demonstrate that the relevant safety and certification requirements are being met by modelling the system in question. To do this existing modelling techniques use variations of fault tree analysis and Markov analysis with various simplifying assumptions, made to assist in the analytical process. Monte Carlo simulation is presented here as an alternative method of analysis, which can deal well with the potential difficulties that may present themselves when modelling TLD, such as the complex architectures of aircraft systems and dependencies that are introduced when applying TLD. In this paper a simple example system is introduced and the application of TLD to it is modelled using the existing variation of Markov analysis and a Monte Carlo simulation technique. The results obtained using the different techniques are seen to differ and a number of reasons are suggested for this difference.
Received on December 07, 2006
References: 05

A prognosis of the machinery condition is an important potential tool to avoid unpredictable and expensive disturbances, failure and shut down in industrial production and transportation. A diagnosis of machinery condition can be carried out based on online condition monitoring. The prediction of future performance needs, in addition to knowledge of the present state, also a good understanding of trends and physical laws influencing the progress of the deterioration processes in components. Tribology offers an understanding of the wear and friction related phenomena in machinery and their mechanisms. After a short transient period of running-in wear, the wear process is linear for a long period at steady state running but it is close to exponential at catastrophic wear. Monitoring based regression signal analysis offers a new possibility to predict wear progress and wear failure. The VTT Diagnostic Circle is a method for integrated tribological and vibration analysis of machinery condition.
Received on December 22, 2006
References: 30

The Reliability Centred Maintenance strategy aims at deciding predictive maintenance tasks on process equipments. The approaches such as FMEA, HAZOP and FTA are most widely used analytical tools for identifying the predominant failure modes, failure causes and possible remedial measures thereof. The success of predictive maintenance program lies in raising an appropriate 'Early Warning' with the help of monitoring the condition of parameters of interest. Randomness of degradation, dependency of the failure causes, uncertain and insufficient data further complicate the estimation of failure probabilities. Process plant engineers always look for a practical and approximate solutions rather than an exact one that uses rigorous analytical models.
In the present paper, to estimate the top event failure probability, the 'Interval of Confidence' of failure probabilities using fault tree analysis and simulation is proposed. The approach uses fuzzy set theory, hence the failure probability may be called as Fuzzy Top Event Probability. Ease and robustness of the approach is demonstrated with condition monitoring data collected on large Electric Motors in an Integrated Steel Plant. This paper also highlights use of the results from the proposed approach to review the maintenance requirements of the equipment to strengthen the implementation of the RCM.
Received on August 07, 2007
References: 24

Comparative effect of temperature and relative humidity data at five places along Indian coasts on corrosion initiation time of steel reinforcement in concrete members subjected to chloride ingress is investigated. Corrosion initiation time at all the places is predicted by considering respective temperature and relative humidity data and assuming same values of all other parameters. A large variation is found in corrosion initiation time which showed necessity to consider the temperature and relative humidity data in a region. Corrosion initiation time is useful for owner, designer, or to an organization to take decision in time of priority of repairs, repair strategy, corrosion protection. Cover required for 50 years time to initiate corrosion is also determined and presented. Similar approach can be applied to design a new structure for expected time to initiate corrosion.
Received on May 11, 2007
References: 07

Most existing reliability growth models ignore the reliability test/improvement process, where engineers identify distinct failure modes through tests and redesign the product/system to remove these modes. In this paper, we present a discrete reliability growth model, which is similar to the Crow's projection and extended reliability growth models but without the doubtful assumptions implied by these models. We demonstrate the use of our model in the project selection of reliability improvement projects during system redesign.
Received on March 13, 2008
References: 02