Track 2 Session 5
8:50 to 10:00 a.m. Thursday April 12, 2007

Repairable Systems: Data Analysis and Modeling

Data obtained from repairable systems must be analyzed differently than data from non-repairable systems. For non-repairable systems, the failure data is collected and a representative distribution is fitted in order to perform reliability estimations. For repairable systems, the analysis is not as straightforward since a stochastic process is in place. Different approaches are available depending on the data and the objective of the analysis. For example, the engineer can analyze the data "directly" by fitting the NHPP-Power Law model or the Generalized Renewal Process model. Another approach is to break down the system into its Lowest Replaceable Units and failure modes, obtain the failure distributions of these, and then create and analyze the Reliability Block Diagram of the system. Each method and model has its underlying assumptions and it is imperative for the analyst to be aware of them. In this presentation, the different approaches of treating repairable system data are reviewed, and the available models for each approach are explained and demonstrated using a case study. In addition, a variety of additional considerations and results pertaining to repairable systems are presented.