ABSTRACT
In Chapter 13, we presented analysis techniques for repairable systems, focusing primarily on renewal processes. The times between failures were independent and identically distributed (i.i.d.) observations from a single population. If the i.i.d. assumptions do not hold, the renewal process is not a suitable model. Consequently, we need other analysis methods that deal with more general patterns of sequential repair times. For example, if a trend is present, such information would be useful for both the user and the manufacturer in determining if reliability objectives are being met. One may also ask what design or operation factors inµuence repair frequency, whether maintenance schedules are appropriate, what warranty costs are anticipated, whether the provisions for spare parts are adequate, and so on. In many real-life instances, the time to a subsequent repair is generally a function
of many variables including the basic system design, the operating conditions, the environment, and the quality of the repairs (the materials used, the competency of the technician, etc.). Thus, there is a genuine possibility of nonrenewal processes in which interrepair times are not independent or identically distributed. For multicomponent systems and different types of repair actions, the renewal model becomes even less plausible. This chapter continues the development of key concepts for analyzing data from sys-
tems subject to repair. We present both graphical and analytical procedures. We consider speci—c models and goodness-of-—t tests. Methods for detecting trend are presented. Simulation methods are covered. The topics of reliability growth and associated models are also discussed. This chapter deals with the topic of nonrenewal repairable reliability, an area that has many important and practical applications.
