Quality and Reliability Engineering International: Special Issue on Selected Papers from the 7th IMA International Conference on Modelling in Industrial Maintenance and Reliability (MIMAR)


  • Date: 26 Sep 2012

The recent special issue of Quality and Reliability Engineering International, guest edited by Wenbin Wang (University of Science and Technology, Beijing), Philip Scarf (University of Salford) and Shaomin Wu (University of Cranfield) features selected papers presented at the 7th IMA International Conference on Modelling in Industrial Maintenance and Reliability (MIMAR), held at Sidney Sussex College,  University of Cambridge, UK, on 18–19 April 2011. The 10 papers published in this special issue all centre around modelling issues in maintenance and reliability to some extent.

thumbnail image: Quality and Reliability Engineering International: Special Issue on Selected Papers from the 7th IMA International Conference on Modelling in Industrial Maintenance and Reliability (MIMAR)

Maintenance and reliability modelling has attracted the attention of the researchers and practitioners in the maintenance and reliability field for a long time. Better reliability is a key objective of all plant managers, and preventive maintenance is a way to improve system reliability. One of the main difficulties in modelling maintenance and reliability problems is the uncertainty associated with plant failures; therefore, much effort has been put in the modelling to quantify such uncertainty.

The paper by Alkali presented generalized proportional intensities models and fitted the models to gas turbine generator data. The model parameters were estimated by a closed form maximum likelihood method and the goodness of fit was discussed. The modelling objective is to determine the optimal preventive maintenance schedule. Studies using real data should be much encouraged and Alkali has made a good attempt.

Most of papers in maintenance and reliability use a single objective such as cost or downtime or reliability. However, in practice more than one objective may be of interest. Almeida presented a multicriteria decision model to support decision makers in choosing the optimal maintenance interval based on the combination of conflicting criteria. He proposed a procedure based on the well-known multi-attribute utility theory (MAUT). As the paper of Alkali, Almeida also used real data to show the applicability of the proposed approach.

The paper by Bazrafshan and Hajjari is somehow unlike the rest of papers in this special issue in that they focused on organizational issues of maintenance. They developed a framework based on a systematic perspective on the design of a maintenance system. They described a case study regarding the application of the organizational perspective to a practical maintenance problem. It is known that a well-organized maintenance function can lead to much improvement without putting much effort on modelling other issues such as the maintenance interval. However, not much has been reported in this regard.

The paper by Bell and Percy focused on a traditional topic in maintenance modelling, that is, to determine the optimal maintenance interval. They used, however, a Bayesian approach in dealing with the problem of estimating the model parameters that are uncertain in nature. The main model was a proportional intensity model similar to that of Alkali, and also they fitted their model to gas turbine data, which were appended to the paper.

Bayesian approaches seemed popular compared with the classical statistical methods. Bovey and Senalp also used a Bayesian method to update the maintenance model they study using in-service troubleshooting and component testing records. One of the challenges is how to justify each change to the system experts. They therefore presented a novel approach for the justification of updates to a Bayesian network model. The result showed the improvement by fitting a model to the new records and a good justification can be obtained from quantities already computed from the learning process.

Condition-based maintenance models have gained much attention recently due to the advance in monitoring techniques and signal processing. Do Van and Berenguer presented a condition-based maintenance policy for a single unit production system. They assumed that preventive maintenance is not perfect in that it can only partially restore the system to a better than old status. They investigated different cost functions due to different types of imperfect preventive maintenance and demonstrated their model by a numerical example of a multistate production system.

Jiang in his paper proposed a unified representation of the Cox model and its extensions and called it the general proportional model. A multistep procedure was proposed to sequentially determine the three parts of the model, that is, the baseline, covariate and the stochastic parts. The main advantage of the proposed model is that it makes modelling flexible and transparent. The appropriateness and usefulness of the model was illustrated through a detailed analysis of two real-world examples published in literature.

Jiang et al. also presented a condition-based maintenance model under partial observed information. They used a continuous time homogeneous semi-Markov process to describe the state transition process. Because the state is not completely observable, their modelling objective is to determine the maximum likelihood estimates of the model parameters using the EM algorithm, which is a proven approach for dealing with hidden information. They derived explicit formulae for both the pseudo-likelihood function and the parameter updates in each iteration of the EM algorithm.

Lu et al. dealt with a one-out-of-two system with a cold standby component. The failure process of the components was described by the delay time concept. They analyzed five scenarios of the system state at an inspection epoch and derived their analytical expressions. One feature of the study was to consider the impact of the repair time on the system performance and this complicated the modelling development. The stationary distributions of the elapsed time for the normal, delay and repair times exist and they were calculated by an iterative approach.

Akbarov and Wu presented something that is different from the rest of the papers in that they looked at a problem of warranty claims. They used a weighted maximum likelihood estimation method to estimate the model parameters because they argued that recent claims may carry more up-to-date information. The model objective was to forecast the future claims and they compared the forecasting performance of both parametric and nonparametric models. They showed that the mixed nonhomogeneous Poisson process models are better models among the class they studied, and the model built using weighted maximum likelihood estimation yielded a smaller error than that using standard maximum likelihood.

Related Topics

Related Publications

Related Content

Site Footer


This website is provided by John Wiley & Sons Limited, The Atrium, Southern Gate, Chichester, West Sussex PO19 8SQ (Company No: 00641132, VAT No: 376766987)

Published features on StatisticsViews.com are checked for statistical accuracy by a panel from the European Network for Business and Industrial Statistics (ENBIS)   to whom Wiley and StatisticsViews.com express their gratitude. This panel are: Ron Kenett, David Steinberg, Shirley Coleman, Irena Ograjenšek, Fabrizio Ruggeri, Rainer Göb, Philippe Castagliola, Xavier Tort-Martorell, Bart De Ketelaere, Antonio Pievatolo, Martina Vandebroek, Lance Mitchell, Gilbert Saporta, Helmut Waldl and Stelios Psarakis.