ABSTRACT
In all previous chapters, we focused on the reliability of nonrepairable components such as lightbulbs or integrated circuits. The working assumption was that times to failure were a truly random sample of independent and identically distributed (i.i.d.) observations from a single population. Consequently, individual failure times could be combined for analysis, neglecting any order of occurrence of the original data. For example, suppose limited oven space requires us to divide a sample of units for stress into two groups and run each group in separate weeks. When both groups have completed the stress (under identical conditions), we reasonably assume that we can combine the data for analysis irrespective of the —rst group’s failure times occurring before the failure times of the second group. However, there are many common situations in which the occurrence order of failure
times has signi—cance. Consider an action that restores a failed system or process to operation, for example, replacing a circuit board —xes a computer, changing a clogged gas line —lter repairs an automobile, resuming electrical power to return a factory to production following a utility failure, or calling a server reconnects an interrupted video-teleconfere ncing session. In all these instances, the failures occur sequentially in time, and assumptions of a single population distribution and independence for the times between failures may be invalid. If the i.i.d. properties hold for the times between failures, the repair rate is stable, and
the renewal process brieµy mentioned in Chapter 2 is an appropriate model. This chapter focuses mainly on the characteristics and features of renewal processes. Conversely, if there is evidence of a trend such as varying (e.g., improving or worsening) repair rates, the renewal model no longer applies. Then, assessing and modeling the system behavior requires consideration of the patterns of the sequential repair times. Chapter 13 covers the subject of nonrenewal processes in detail. Chapters 12 and 13 present key concepts on analyzing data from systems under going
repair. The topics include useful methods, both nonparametric and parametric, for describing and modeling such data. Veri—cation of assumptions for analysis is stressed. To enhance comprehension, simulation techniques are provided. We cover many important and practical issues involving the reliability of repairable systems in this chapter.
