An age-adjusted comparison of field failure data for repairable systems

Abstract

The conventional calendar-based mean cumulative function (MCF) plot is useful in monitoring the field reliability of a population of repairable systems deployed in a large quantity. It is simple and easily understood by management. However, it is age-confounded when the system population is heterogeneous with age. Assuming that the systems follow the well-known "Bathtub" behavior, the population which consists of a higher proportion of aged systems will perform badly on the MCF plot compared to another population made up of mainly newer systems. Hence, direct comparisons between the two populations may not be fair. This paper illustrates a few simple steps that aid in mitigating such age heterogeneity issue prior to plotting the MCF. The ageadjusted MCF allows a fairer comparison of maintenance performance between the populations. The applicability of the proposed approach is demonstrated using actual field failure data. The case study shows that if differences in system age compositions are not accounted for, different conclusions could be drawn which could be detrimental to the maintenance personnel morale. Worse still, precious maintenance resources might be channeled to the wrong location. Other than age, the proposed approach can be easily extended to adjust for other system attributes. © 2008 IEEE.